GSOC 2013 idea

classic Classic list List threaded Threaded
9 messages Options
Reply | Threaded
Open this post in threaded view
|

GSOC 2013 idea

mohit kumar
Hi all,
I participated in last year's Summer of Code as a students from Opticks and made a plug-in for object based image analysis. That work included basic object creation, object attribution and change detection techniques.
This year I want to take part in GSOC again and as I am familiar with opticks development, I would love to apply again as an opticks student.
I have a basic idea on extending my last year's work and making a few more tools using object based techniques.
The three tools I have in mind are as follows.
  1. Image Fusion:
    • The Fusion software is a tool fuses a PAN image and a correspondingly registered MSS image. The fused image has salient information from both the images. The fused image can have complementary spatial and spectral resolution characteristics.
    • GLSI fusion[1]. Will do the fusion on each object instead of the 3X3 window as mentioned in the paper.
    • Improved classification. The classification we get on the fused image will be better as compared to MSS or panchromatic classification done separately. 

  2. Missing Data Filling
    • The missing data filling tool detects the occlusions in a given image.
    • Satellite images can sometimes have occlusions which may lead to loss of information.
    • So, in order to overcome this loss, the missing information is filled in by removing these occlusions with the help of similar supporting images of the same area.
    • In general, the occlusions may be clouds, shadows, trees etc. Initially I want to focus on cloud detection and removal using this algorithm[2]. .

  3. Improved Change Detection
    • In addition to the earlier area based approach, this time I will use the texture and shape features of the objects in the image to classify the change .
    • Querying the change image. e.g. the user will be able to visualize "appeared objects which are circular in shape" etc.
    • Classification of the objects prior to change detection to ease the queries. The query can now be " disappeared buildings"
    • The UI will enable the user to modify the change class(appeared, merged, split etc) in the change image if the user is not happy with the change detection output.

References

  1. A. Khandelwal, K. S. Rajan. An E cient Algorithm for Generating Pan Sharpened Quick Look Images.
  2. Q. Meng, B. E. Borders, C. J. Cieszewski and M. Madden. Closest Spectral Fit for Removing Clouds and Cloud Shadows Photogrammetric Engineering and Remote Sensing, 75(5):569 - 576,may 2009.


Looking forward for feedback and suggestions.


Thanks and Regards,

--
Mohit Kumar
+91-970-3840-175
Lab For Spatial Informatics
International Institute of Information Technology
Hyderabad, India

------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
_______________________________________________
Opticks-devs mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/opticks-devs
Reply | Threaded
Open this post in threaded view
|

Re: GSOC 2013 idea

tclarke
Administrator

Comments in red..overall, I think it’s a good start and could be an appropriate level of effort for a summer provided the work as proper bounds.

 

From: mohit kumar [mailto:[hidden email]]
Sent: Thursday, April 11, 2013 6:50 AM
To: [hidden email]
Subject: [Opticks-devs] GSOC 2013 idea

 

Hi all,

I participated in last year's Summer of Code as a students from Opticks and made a plug-in for object based image analysis. That work included basic object creation, object attribution and change detection techniques.

This year I want to take part in GSOC again and as I am familiar with opticks development, I would love to apply again as an opticks student.

I have a basic idea on extending my last year's work and making a few more tools using object based techniques.

The three tools I have in mind are as follows.

  1. Image Fusion:
    • The Fusion software is a tool fuses a PAN image and a correspondingly registered MSS image. The fused image has salient information from both the images. The fused image can have complementary spatial and spectral resolution characteristics.
    • GLSI fusion[1]. Will do the fusion on each object instead of the 3X3 window as mentioned in the paper.
    • Improved classification. The classification we get on the fused image will be better as compared to MSS or panchromatic classification done separately. 

This seems like it might be a workable change to the G-LSI algorithm but I was unable to find the original paper (just a poster) so I can’t say for sure. I’m guessing the publication is only available to journal members (or for purchase) so I’ll try and track down a copy. Could you send a summary of the G-LSI process with a little more detail than the poster? (or a link to the paper if it’s available on the web)

 

  1. Missing Data Filling
    • The missing data
      filling tool detects the occlusions in a given image.
    • Satellite images can sometimes have occlusions which may lead to loss of information.
    • So, in order to overcome this loss, the missing information is
      filled in by removing these occlusions with the help of similar supporting images of the same area.
    • In general, the occlusions may be clouds, shadows, trees etc. Initially I want to focus on cloud detection and removal using this algorithm[2]. .

This algorithm suggest the use reflectance data with atmospheric affects removed. What tool or algorithm do you intend to use for this? The usual algorithm used commercially is FLAASH (licensing can be expensive..it’s usually included in commercial tools such as ENVI). While not necessary, this step is increasingly important when you have stronger atmospheric effects such as in images with lots of aerosol content (cloudy images, etc.). If you skip atmospheric correction, you’ll need to carefully choose your test data to minimize atmospheric effects.

I assume you’ll be using Landsat 5 data for this? (the paper provides threshold values for TM images…using another source will require and algorithm for identifying cloud and shadow pixels)

  1. Improved Change Detection
    • In addition to the earlier area based approach, this time I will use the texture and shape features of the objects in the image to classify the change .
    • Querying the change image. e.g. the user will be able to visualize "appeared objects which are circular in shape" etc.
    • Classification of the objects prior to change detection to ease the queries. The query can now be " disappeared buildings"
    • The UI will enable the user to modify the change class(appeared, merged, split etc) in the change image if the user is not happy with the change detection output.

References

  1. A. Khandelwal, K. S. Rajan. An E cient Algorithm for Generating Pan Sharpened Quick Look Images.
  2. Q. Meng, B. E. Borders, C. J. Cieszewski and M. Madden. Closest Spectral Fit for Removing Clouds and Cloud Shadows Photogrammetric Engineering and Remote Sensing, 75(5):569 - 576,may 2009.

 

Looking forward for feedback and suggestions.

 

Thanks and Regards,

--

Mohit Kumar

+91-970-3840-175

Lab For Spatial Informatics
International Institute of Information Technology
Hyderabad, India


This message and any enclosures are intended only for the addressee.  Please 

notify the sender by email if you are not the intended recipient.  If you are 

not the intended recipient, you may not use, copy, disclose, or distribute this 

message or its contents or enclosures to any other person and any such actions 

may be unlawful.  Ball reserves the right to monitor and review all messages 

and enclosures sent to or from this email address.

------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
_______________________________________________
Opticks-devs mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/opticks-devs
Reply | Threaded
Open this post in threaded view
|

Re: GSOC 2013 idea

dadkins
Administrator

This is a pretty good start.

As tclarke mentioned, you will need to scope your project appropriately, especially the second algorithm.

And we will definitely need to be able to get at the papers you mentioned to mentor this project.

 

--Dustan

 

From: Clarke, Trevor [mailto:[hidden email]]
Sent: Thursday, April 11, 2013 10:02 AM
To: [hidden email]
Subject: Re: [Opticks-devs] GSOC 2013 idea

 

Comments in red..overall, I think it’s a good start and could be an appropriate level of effort for a summer provided the work as proper bounds.

 

From: mohit kumar [[hidden email]]
Sent: Thursday, April 11, 2013 6:50 AM
To: [hidden email]
Subject: [Opticks-devs] GSOC 2013 idea

 

Hi all,

I participated in last year's Summer of Code as a students from Opticks and made a plug-in for object based image analysis. That work included basic object creation, object attribution and change detection techniques.

This year I want to take part in GSOC again and as I am familiar with opticks development, I would love to apply again as an opticks student.

I have a basic idea on extending my last year's work and making a few more tools using object based techniques.

The three tools I have in mind are as follows.

  1. Image Fusion:
    • The Fusion software is a tool fuses a PAN image and a correspondingly registered MSS image. The fused image has salient information from both the images. The fused image can have complementary spatial and spectral resolution characteristics.
    • GLSI fusion[1]. Will do the fusion on each object instead of the 3X3 window as mentioned in the paper.
    • Improved classification. The classification we get on the fused image will be better as compared to MSS or panchromatic classification done separately. 

This seems like it might be a workable change to the G-LSI algorithm but I was unable to find the original paper (just a poster) so I can’t say for sure. I’m guessing the publication is only available to journal members (or for purchase) so I’ll try and track down a copy. Could you send a summary of the G-LSI process with a little more detail than the poster? (or a link to the paper if it’s available on the web)

 

  1. Missing Data Filling
    • The missing data
      filling tool detects the occlusions in a given image.
    • Satellite images can sometimes have occlusions which may lead to loss of information.
    • So, in order to overcome this loss, the missing information is
      filled in by removing these occlusions with the help of similar supporting images of the same area.
    • In general, the occlusions may be clouds, shadows, trees etc. Initially I want to focus on cloud detection and removal using this algorithm[2]. .

This algorithm suggest the use reflectance data with atmospheric affects removed. What tool or algorithm do you intend to use for this? The usual algorithm used commercially is FLAASH (licensing can be expensive..it’s usually included in commercial tools such as ENVI). While not necessary, this step is increasingly important when you have stronger atmospheric effects such as in images with lots of aerosol content (cloudy images, etc.). If you skip atmospheric correction, you’ll need to carefully choose your test data to minimize atmospheric effects.

I assume you’ll be using Landsat 5 data for this? (the paper provides threshold values for TM images…using another source will require and algorithm for identifying cloud and shadow pixels)

  1. Improved Change Detection
    • In addition to the earlier area based approach, this time I will use the texture and shape features of the objects in the image to classify the change .
    • Querying the change image. e.g. the user will be able to visualize "appeared objects which are circular in shape" etc.
    • Classification of the objects prior to change detection to ease the queries. The query can now be " disappeared buildings"
    • The UI will enable the user to modify the change class(appeared, merged, split etc) in the change image if the user is not happy with the change detection output.

References

  1. A. Khandelwal, K. S. Rajan. An E cient Algorithm for Generating Pan Sharpened Quick Look Images.
  2. Q. Meng, B. E. Borders, C. J. Cieszewski and M. Madden. Closest Spectral Fit for Removing Clouds and Cloud Shadows Photogrammetric Engineering and Remote Sensing, 75(5):569 - 576,may 2009.

 

Looking forward for feedback and suggestions.

 

Thanks and Regards,

--

Mohit Kumar

+91-970-3840-175

Lab For Spatial Informatics
International Institute of Information Technology
Hyderabad, India

 
 
This message and any enclosures are intended only for the addressee.  Please 
 
notify the sender by email if you are not the intended recipient.  If you are 
 
not the intended recipient, you may not use, copy, disclose, or distribute this 
 
message or its contents or enclosures to any other person and any such actions 
 
may be unlawful.  Ball reserves the right to monitor and review all messages 
 
and enclosures sent to or from this email address.

This message and any enclosures are intended only for the addressee.  Please 

notify the sender by email if you are not the intended recipient.  If you are 

not the intended recipient, you may not use, copy, disclose, or distribute this 

message or its contents or enclosures to any other person and any such actions 

may be unlawful.  Ball reserves the right to monitor and review all messages 

and enclosures sent to or from this email address.

------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
_______________________________________________
Opticks-devs mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/opticks-devs
Reply | Threaded
Open this post in threaded view
|

Re: GSOC 2013 idea

mohit kumar
Thanks Tclarke and Dadkins for your feedback and interest.
I have attached the the paper for the fusion algorithm to this mail.

Regarding the missing data filling algorithm, I just had a look at the paper and did not thought much about it. Initially I want to implement the basic cloud detection without any atmospheric effects. So the scope of this tool stays limited to images without a lot of atmospheric disturbances.
The threshold values for the cloud pixels either could vary from image source or can be user defined or can be set a fixed value for every image.

Thanks and regards,
Mohit Kumar

On Fri, Apr 12, 2013 at 2:15 AM, Adkins, Dustan <[hidden email]> wrote:

This is a pretty good start.

As tclarke mentioned, you will need to scope your project appropriately, especially the second algorithm.

And we will definitely need to be able to get at the papers you mentioned to mentor this project.

 

--Dustan

 

From: Clarke, Trevor [mailto:[hidden email]]
Sent: Thursday, April 11, 2013 10:02 AM
To: [hidden email]
Subject: Re: [Opticks-devs] GSOC 2013 idea

 

Comments in red..overall, I think it’s a good start and could be an appropriate level of effort for a summer provided the work as proper bounds.

 

From: mohit kumar [[hidden email]]
Sent: Thursday, April 11, 2013 6:50 AM
To: [hidden email]
Subject: [Opticks-devs] GSOC 2013 idea

 

Hi all,

I participated in last year's Summer of Code as a students from Opticks and made a plug-in for object based image analysis. That work included basic object creation, object attribution and change detection techniques.

This year I want to take part in GSOC again and as I am familiar with opticks development, I would love to apply again as an opticks student.

I have a basic idea on extending my last year's work and making a few more tools using object based techniques.

The three tools I have in mind are as follows.

  1. Image Fusion:
    • The Fusion software is a tool fuses a PAN image and a correspondingly registered MSS image. The fused image has salient information from both the images. The fused image can have complementary spatial and spectral resolution characteristics.
    • GLSI fusion[1]. Will do the fusion on each object instead of the 3X3 window as mentioned in the paper.
    • Improved classification. The classification we get on the fused image will be better as compared to MSS or panchromatic classification done separately. 

This seems like it might be a workable change to the G-LSI algorithm but I was unable to find the original paper (just a poster) so I can’t say for sure. I’m guessing the publication is only available to journal members (or for purchase) so I’ll try and track down a copy. Could you send a summary of the G-LSI process with a little more detail than the poster? (or a link to the paper if it’s available on the web)

 

  1. Missing Data Filling
    • The missing data
      filling tool detects the occlusions in a given image.
    • Satellite images can sometimes have occlusions which may lead to loss of information.
    • So, in order to overcome this loss, the missing information is
      filled in by removing these occlusions with the help of similar supporting images of the same area.
    • In general, the occlusions may be clouds, shadows, trees etc. Initially I want to focus on cloud detection and removal using this algorithm[2]. .

This algorithm suggest the use reflectance data with atmospheric affects removed. What tool or algorithm do you intend to use for this? The usual algorithm used commercially is FLAASH (licensing can be expensive..it’s usually included in commercial tools such as ENVI). While not necessary, this step is increasingly important when you have stronger atmospheric effects such as in images with lots of aerosol content (cloudy images, etc.). If you skip atmospheric correction, you’ll need to carefully choose your test data to minimize atmospheric effects.

I assume you’ll be using Landsat 5 data for this? (the paper provides threshold values for TM images…using another source will require and algorithm for identifying cloud and shadow pixels)

  1. Improved Change Detection
    • In addition to the earlier area based approach, this time I will use the texture and shape features of the objects in the image to classify the change .
    • Querying the change image. e.g. the user will be able to visualize "appeared objects which are circular in shape" etc.
    • Classification of the objects prior to change detection to ease the queries. The query can now be " disappeared buildings"
    • The UI will enable the user to modify the change class(appeared, merged, split etc) in the change image if the user is not happy with the change detection output.

References

  1. A. Khandelwal, K. S. Rajan. An E cient Algorithm for Generating Pan Sharpened Quick Look Images.
  2. Q. Meng, B. E. Borders, C. J. Cieszewski and M. Madden. Closest Spectral Fit for Removing Clouds and Cloud Shadows Photogrammetric Engineering and Remote Sensing, 75(5):569 - 576,may 2009.

 

Looking forward for feedback and suggestions.

 

Thanks and Regards,

--

Mohit Kumar

+91-970-3840-175

Lab For Spatial Informatics
International Institute of Information Technology
Hyderabad, India

 
 
This message and any enclosures are intended only for the addressee.  Please 
 
notify the sender by email if you are not the intended recipient.  If you are 
 
not the intended recipient, you may not use, copy, disclose, or distribute this 
 
message or its contents or enclosures to any other person and any such actions 
 
may be unlawful.  Ball reserves the right to monitor and review all messages 
 
and enclosures sent to or from this email address.
This message and any enclosures are intended only for the addressee.  Please 

notify the sender by email if you are not the intended recipient.  If you are 

not the intended recipient, you may not use, copy, disclose, or distribute this 

message or its contents or enclosures to any other person and any such actions 

may be unlawful.  Ball reserves the right to monitor and review all messages 

and enclosures sent to or from this email address.

------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
_______________________________________________
Opticks-devs mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/opticks-devs




--
Mohit Kumar
+91-970-3840-175
Lab For Spatial Informatics
International Institute of Information Technology
Hyderabad, India

------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
_______________________________________________
Opticks-devs mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/opticks-devs

G-LSI_Fusion.pdf (513K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: GSOC 2013 idea

tclarke
Administrator

That sounds good  to me…limiting your data and test cases such that the atmospheric effects don’t need to be addressed is fine.

 

From: mohit kumar [mailto:[hidden email]]
Sent: Friday, April 12, 2013 6:15 AM
To: [hidden email]
Subject: Re: [Opticks-devs] GSOC 2013 idea

 

Thanks Tclarke and Dadkins for your feedback and interest.

I have attached the the paper for the fusion algorithm to this mail.

Regarding the missing data filling algorithm, I just had a look at the paper and did not thought much about it. Initially I want to implement the basic cloud detection without any atmospheric effects. So the scope of this tool stays limited to images without a lot of atmospheric disturbances.

The threshold values for the cloud pixels either could vary from image source or can be user defined or can be set a fixed value for every image.

 

Thanks and regards,

Mohit Kumar

 

On Fri, Apr 12, 2013 at 2:15 AM, Adkins, Dustan <[hidden email]> wrote:

This is a pretty good start.

As tclarke mentioned, you will need to scope your project appropriately, especially the second algorithm.

And we will definitely need to be able to get at the papers you mentioned to mentor this project.

 

--Dustan

 

From: Clarke, Trevor [mailto:[hidden email]]
Sent: Thursday, April 11, 2013 10:02 AM
To: [hidden email]
Subject: Re: [Opticks-devs] GSOC 2013 idea

 

Comments in red..overall, I think it’s a good start and could be an appropriate level of effort for a summer provided the work as proper bounds.

 

From: mohit kumar [[hidden email]]
Sent: Thursday, April 11, 2013 6:50 AM
To: [hidden email]
Subject: [Opticks-devs] GSOC 2013 idea

 

Hi all,

I participated in last year's Summer of Code as a students from Opticks and made a plug-in for object based image analysis. That work included basic object creation, object attribution and change detection techniques.

This year I want to take part in GSOC again and as I am familiar with opticks development, I would love to apply again as an opticks student.

I have a basic idea on extending my last year's work and making a few more tools using object based techniques.

The three tools I have in mind are as follows.

  1. Image Fusion:
    • The Fusion software is a tool fuses a PAN image and a correspondingly registered MSS image. The fused image has salient information from both the images. The fused image can have complementary spatial and spectral resolution characteristics.
    • GLSI fusion[1]. Will do the fusion on each object instead of the 3X3 window as mentioned in the paper.
    • Improved classification. The classification we get on the fused image will be better as compared to MSS or panchromatic classification done separately. 

This seems like it might be a workable change to the G-LSI algorithm but I was unable to find the original paper (just a poster) so I can’t say for sure. I’m guessing the publication is only available to journal members (or for purchase) so I’ll try and track down a copy. Could you send a summary of the G-LSI process with a little more detail than the poster? (or a link to the paper if it’s available on the web)

 

  1. Missing Data Filling
    • The missing data
      filling tool detects the occlusions in a given image.
    • Satellite images can sometimes have occlusions which may lead to loss of information.
    • So, in order to overcome this loss, the missing information is
      filled in by removing these occlusions with the help of similar supporting images of the same area.
    • In general, the occlusions may be clouds, shadows, trees etc. Initially I want to focus on cloud detection and removal using this algorithm[2]. .

This algorithm suggest the use reflectance data with atmospheric affects removed. What tool or algorithm do you intend to use for this? The usual algorithm used commercially is FLAASH (licensing can be expensive..it’s usually included in commercial tools such as ENVI). While not necessary, this step is increasingly important when you have stronger atmospheric effects such as in images with lots of aerosol content (cloudy images, etc.). If you skip atmospheric correction, you’ll need to carefully choose your test data to minimize atmospheric effects.

I assume you’ll be using Landsat 5 data for this? (the paper provides threshold values for TM images…using another source will require and algorithm for identifying cloud and shadow pixels)

  1. Improved Change Detection
    • In addition to the earlier area based approach, this time I will use the texture and shape features of the objects in the image to classify the change .
    • Querying the change image. e.g. the user will be able to visualize "appeared objects which are circular in shape" etc.
    • Classification of the objects prior to change detection to ease the queries. The query can now be " disappeared buildings"
    • The UI will enable the user to modify the change class(appeared, merged, split etc) in the change image if the user is not happy with the change detection output.

References

  1. A. Khandelwal, K. S. Rajan. An E cient Algorithm for Generating Pan Sharpened Quick Look Images.
  2. Q. Meng, B. E. Borders, C. J. Cieszewski and M. Madden. Closest Spectral Fit for Removing Clouds and Cloud Shadows Photogrammetric Engineering and Remote Sensing, 75(5):569 - 576,may 2009.

 

Looking forward for feedback and suggestions.

 

Thanks and Regards,

--

Mohit Kumar

+91-970-3840-175

Lab For Spatial Informatics
International Institute of Information Technology
Hyderabad, India

 
 
This message and any enclosures are intended only for the addressee.  Please 
 
 
notify the sender by email if you are not the intended recipient.  If you are 
 
not the intended recipient, you may not use, copy, disclose, or distribute this 
 
message or its contents or enclosures to any other person and any such actions 
 
may be unlawful.  Ball reserves the right to monitor and review all messages 
 
and enclosures sent to or from this email address.
This message and any enclosures are intended only for the addressee.  Please 
 
notify the sender by email if you are not the intended recipient.  If you are 
 
not the intended recipient, you may not use, copy, disclose, or distribute this 
 
message or its contents or enclosures to any other person and any such actions 
 
may be unlawful.  Ball reserves the right to monitor and review all messages 
 
and enclosures sent to or from this email address.


------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
_______________________________________________
Opticks-devs mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/opticks-devs




--

Mohit Kumar

+91-970-3840-175

Lab For Spatial Informatics
International Institute of Information Technology
Hyderabad, India


This message and any enclosures are intended only for the addressee.  Please 

notify the sender by email if you are not the intended recipient.  If you are 

not the intended recipient, you may not use, copy, disclose, or distribute this 

message or its contents or enclosures to any other person and any such actions 

may be unlawful.  Ball reserves the right to monitor and review all messages 

and enclosures sent to or from this email address.

------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
_______________________________________________
Opticks-devs mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/opticks-devs
Reply | Threaded
Open this post in threaded view
|

Re: GSOC 2013 idea

mohit kumar
Thanks tclarke for the feedback, so what should be the next step?
Or any more addition to the work?

Regards,
Mohit


On Mon, Apr 15, 2013 at 7:40 PM, Clarke, Trevor <[hidden email]> wrote:

That sounds good  to me…limiting your data and test cases such that the atmospheric effects don’t need to be addressed is fine.

 

From: mohit kumar [mailto:[hidden email]]
Sent: Friday, April 12, 2013 6:15 AM


To: [hidden email]
Subject: Re: [Opticks-devs] GSOC 2013 idea

 

Thanks Tclarke and Dadkins for your feedback and interest.

I have attached the the paper for the fusion algorithm to this mail.

Regarding the missing data filling algorithm, I just had a look at the paper and did not thought much about it. Initially I want to implement the basic cloud detection without any atmospheric effects. So the scope of this tool stays limited to images without a lot of atmospheric disturbances.

The threshold values for the cloud pixels either could vary from image source or can be user defined or can be set a fixed value for every image.

 

Thanks and regards,

Mohit Kumar

 

On Fri, Apr 12, 2013 at 2:15 AM, Adkins, Dustan <[hidden email]> wrote:

This is a pretty good start.

As tclarke mentioned, you will need to scope your project appropriately, especially the second algorithm.

And we will definitely need to be able to get at the papers you mentioned to mentor this project.

 

--Dustan

 

From: Clarke, Trevor [mailto:[hidden email]]
Sent: Thursday, April 11, 2013 10:02 AM
To: [hidden email]
Subject: Re: [Opticks-devs] GSOC 2013 idea

 

Comments in red..overall, I think it’s a good start and could be an appropriate level of effort for a summer provided the work as proper bounds.

 

From: mohit kumar [[hidden email]]
Sent: Thursday, April 11, 2013 6:50 AM
To: [hidden email]
Subject: [Opticks-devs] GSOC 2013 idea

 

Hi all,

I participated in last year's Summer of Code as a students from Opticks and made a plug-in for object based image analysis. That work included basic object creation, object attribution and change detection techniques.

This year I want to take part in GSOC again and as I am familiar with opticks development, I would love to apply again as an opticks student.

I have a basic idea on extending my last year's work and making a few more tools using object based techniques.

The three tools I have in mind are as follows.

  1. Image Fusion:
    • The Fusion software is a tool fuses a PAN image and a correspondingly registered MSS image. The fused image has salient information from both the images. The fused image can have complementary spatial and spectral resolution characteristics.
    • GLSI fusion[1]. Will do the fusion on each object instead of the 3X3 window as mentioned in the paper.
    • Improved classification. The classification we get on the fused image will be better as compared to MSS or panchromatic classification done separately. 

This seems like it might be a workable change to the G-LSI algorithm but I was unable to find the original paper (just a poster) so I can’t say for sure. I’m guessing the publication is only available to journal members (or for purchase) so I’ll try and track down a copy. Could you send a summary of the G-LSI process with a little more detail than the poster? (or a link to the paper if it’s available on the web)

 

  1. Missing Data Filling
    • The missing data
      filling tool detects the occlusions in a given image.
    • Satellite images can sometimes have occlusions which may lead to loss of information.
    • So, in order to overcome this loss, the missing information is
      filled in by removing these occlusions with the help of similar supporting images of the same area.
    • In general, the occlusions may be clouds, shadows, trees etc. Initially I want to focus on cloud detection and removal using this algorithm[2]. .

This algorithm suggest the use reflectance data with atmospheric affects removed. What tool or algorithm do you intend to use for this? The usual algorithm used commercially is FLAASH (licensing can be expensive..it’s usually included in commercial tools such as ENVI). While not necessary, this step is increasingly important when you have stronger atmospheric effects such as in images with lots of aerosol content (cloudy images, etc.). If you skip atmospheric correction, you’ll need to carefully choose your test data to minimize atmospheric effects.

I assume you’ll be using Landsat 5 data for this? (the paper provides threshold values for TM images…using another source will require and algorithm for identifying cloud and shadow pixels)

  1. Improved Change Detection
    • In addition to the earlier area based approach, this time I will use the texture and shape features of the objects in the image to classify the change .
    • Querying the change image. e.g. the user will be able to visualize "appeared objects which are circular in shape" etc.
    • Classification of the objects prior to change detection to ease the queries. The query can now be " disappeared buildings"
    • The UI will enable the user to modify the change class(appeared, merged, split etc) in the change image if the user is not happy with the change detection output.

References

  1. A. Khandelwal, K. S. Rajan. An E cient Algorithm for Generating Pan Sharpened Quick Look Images.
  2. Q. Meng, B. E. Borders, C. J. Cieszewski and M. Madden. Closest Spectral Fit for Removing Clouds and Cloud Shadows Photogrammetric Engineering and Remote Sensing, 75(5):569 - 576,may 2009.

 

Looking forward for feedback and suggestions.

 

Thanks and Regards,

--

Mohit Kumar

+91-970-3840-175

Lab For Spatial Informatics
International Institute of Information Technology
Hyderabad, India

 
 
This message and any enclosures are intended only for the addressee.  Please 
 
 
notify the sender by email if you are not the intended recipient.  If you are 
 
not the intended recipient, you may not use, copy, disclose, or distribute this 
 
message or its contents or enclosures to any other person and any such actions 
 
may be unlawful.  Ball reserves the right to monitor and review all messages 
 
and enclosures sent to or from this email address.
This message and any enclosures are intended only for the addressee.  Please 
 
notify the sender by email if you are not the intended recipient.  If you are 
 
not the intended recipient, you may not use, copy, disclose, or distribute this 
 
message or its contents or enclosures to any other person and any such actions 
 
may be unlawful.  Ball reserves the right to monitor and review all messages 
 
and enclosures sent to or from this email address.


------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
_______________________________________________
Opticks-devs mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/opticks-devs




--

Mohit Kumar

+91-970-3840-175

Lab For Spatial Informatics
International Institute of Information Technology
Hyderabad, India

This message and any enclosures are intended only for the addressee.  Please 

notify the sender by email if you are not the intended recipient.  If you are 

not the intended recipient, you may not use, copy, disclose, or distribute this 

message or its contents or enclosures to any other person and any such actions 

may be unlawful.  Ball reserves the right to monitor and review all messages 

and enclosures sent to or from this email address.

------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
_______________________________________________
Opticks-devs mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/opticks-devs




--
Mohit Kumar
+91-970-3840-175
Lab For Spatial Informatics
International Institute of Information Technology
Hyderabad, India

------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
_______________________________________________
Opticks-devs mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/opticks-devs
Reply | Threaded
Open this post in threaded view
|

Re: GSOC 2013 idea

mohit kumar
Hi all,
Sorry for running late.
This is my first draft of the proposal http://researchweb.iiit.ac.in/~mohit.kumar/Opticksgsoc2013.html

Feedback and suggestions are welcome.

Thanks and Regards,
Mohit


On Mon, Apr 15, 2013 at 8:24 PM, mohit kumar <[hidden email]> wrote:
Thanks tclarke for the feedback, so what should be the next step?
Or any more addition to the work?

Regards,
Mohit


On Mon, Apr 15, 2013 at 7:40 PM, Clarke, Trevor <[hidden email]> wrote:

That sounds good  to me…limiting your data and test cases such that the atmospheric effects don’t need to be addressed is fine.

 

From: mohit kumar [mailto:[hidden email]]
Sent: Friday, April 12, 2013 6:15 AM


To: [hidden email]
Subject: Re: [Opticks-devs] GSOC 2013 idea

 

Thanks Tclarke and Dadkins for your feedback and interest.

I have attached the the paper for the fusion algorithm to this mail.

Regarding the missing data filling algorithm, I just had a look at the paper and did not thought much about it. Initially I want to implement the basic cloud detection without any atmospheric effects. So the scope of this tool stays limited to images without a lot of atmospheric disturbances.

The threshold values for the cloud pixels either could vary from image source or can be user defined or can be set a fixed value for every image.

 

Thanks and regards,

Mohit Kumar

 

On Fri, Apr 12, 2013 at 2:15 AM, Adkins, Dustan <[hidden email]> wrote:

This is a pretty good start.

As tclarke mentioned, you will need to scope your project appropriately, especially the second algorithm.

And we will definitely need to be able to get at the papers you mentioned to mentor this project.

 

--Dustan

 

From: Clarke, Trevor [mailto:[hidden email]]
Sent: Thursday, April 11, 2013 10:02 AM
To: [hidden email]
Subject: Re: [Opticks-devs] GSOC 2013 idea

 

Comments in red..overall, I think it’s a good start and could be an appropriate level of effort for a summer provided the work as proper bounds.

 

From: mohit kumar [[hidden email]]
Sent: Thursday, April 11, 2013 6:50 AM
To: [hidden email]
Subject: [Opticks-devs] GSOC 2013 idea

 

Hi all,

I participated in last year's Summer of Code as a students from Opticks and made a plug-in for object based image analysis. That work included basic object creation, object attribution and change detection techniques.

This year I want to take part in GSOC again and as I am familiar with opticks development, I would love to apply again as an opticks student.

I have a basic idea on extending my last year's work and making a few more tools using object based techniques.

The three tools I have in mind are as follows.

  1. Image Fusion:
    • The Fusion software is a tool fuses a PAN image and a correspondingly registered MSS image. The fused image has salient information from both the images. The fused image can have complementary spatial and spectral resolution characteristics.
    • GLSI fusion[1]. Will do the fusion on each object instead of the 3X3 window as mentioned in the paper.
    • Improved classification. The classification we get on the fused image will be better as compared to MSS or panchromatic classification done separately. 

This seems like it might be a workable change to the G-LSI algorithm but I was unable to find the original paper (just a poster) so I can’t say for sure. I’m guessing the publication is only available to journal members (or for purchase) so I’ll try and track down a copy. Could you send a summary of the G-LSI process with a little more detail than the poster? (or a link to the paper if it’s available on the web)

 

  1. Missing Data Filling
    • The missing data
      filling tool detects the occlusions in a given image.
    • Satellite images can sometimes have occlusions which may lead to loss of information.
    • So, in order to overcome this loss, the missing information is
      filled in by removing these occlusions with the help of similar supporting images of the same area.
    • In general, the occlusions may be clouds, shadows, trees etc. Initially I want to focus on cloud detection and removal using this algorithm[2]. .

This algorithm suggest the use reflectance data with atmospheric affects removed. What tool or algorithm do you intend to use for this? The usual algorithm used commercially is FLAASH (licensing can be expensive..it’s usually included in commercial tools such as ENVI). While not necessary, this step is increasingly important when you have stronger atmospheric effects such as in images with lots of aerosol content (cloudy images, etc.). If you skip atmospheric correction, you’ll need to carefully choose your test data to minimize atmospheric effects.

I assume you’ll be using Landsat 5 data for this? (the paper provides threshold values for TM images…using another source will require and algorithm for identifying cloud and shadow pixels)

  1. Improved Change Detection
    • In addition to the earlier area based approach, this time I will use the texture and shape features of the objects in the image to classify the change .
    • Querying the change image. e.g. the user will be able to visualize "appeared objects which are circular in shape" etc.
    • Classification of the objects prior to change detection to ease the queries. The query can now be " disappeared buildings"
    • The UI will enable the user to modify the change class(appeared, merged, split etc) in the change image if the user is not happy with the change detection output.

References

  1. A. Khandelwal, K. S. Rajan. An E cient Algorithm for Generating Pan Sharpened Quick Look Images.
  2. Q. Meng, B. E. Borders, C. J. Cieszewski and M. Madden. Closest Spectral Fit for Removing Clouds and Cloud Shadows Photogrammetric Engineering and Remote Sensing, 75(5):569 - 576,may 2009.

 

Looking forward for feedback and suggestions.

 

Thanks and Regards,

--

Mohit Kumar

+91-970-3840-175

Lab For Spatial Informatics
International Institute of Information Technology
Hyderabad, India

 
 
This message and any enclosures are intended only for the addressee.  Please 
 
 
notify the sender by email if you are not the intended recipient.  If you are 
 
not the intended recipient, you may not use, copy, disclose, or distribute this 
 
message or its contents or enclosures to any other person and any such actions 
 
may be unlawful.  Ball reserves the right to monitor and review all messages 
 
and enclosures sent to or from this email address.
This message and any enclosures are intended only for the addressee.  Please 
 
notify the sender by email if you are not the intended recipient.  If you are 
 
not the intended recipient, you may not use, copy, disclose, or distribute this 
 
message or its contents or enclosures to any other person and any such actions 
 
may be unlawful.  Ball reserves the right to monitor and review all messages 
 
and enclosures sent to or from this email address.


------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
_______________________________________________
Opticks-devs mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/opticks-devs




--

Mohit Kumar

+91-970-3840-175

Lab For Spatial Informatics
International Institute of Information Technology
Hyderabad, India

This message and any enclosures are intended only for the addressee.  Please 

notify the sender by email if you are not the intended recipient.  If you are 

not the intended recipient, you may not use, copy, disclose, or distribute this 

message or its contents or enclosures to any other person and any such actions 

may be unlawful.  Ball reserves the right to monitor and review all messages 

and enclosures sent to or from this email address.

------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
_______________________________________________
Opticks-devs mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/opticks-devs




--
Mohit Kumar
+91-970-3840-175
Lab For Spatial Informatics
International Institute of Information Technology
Hyderabad, India



--
Mohit Kumar
+91-970-3840-175
Lab For Spatial Informatics
International Institute of Information Technology
Hyderabad, India

------------------------------------------------------------------------------
Try New Relic Now & We'll Send You this Cool Shirt
New Relic is the only SaaS-based application performance monitoring service
that delivers powerful full stack analytics. Optimize and monitor your
browser, app, & servers with just a few lines of code. Try New Relic
and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_apr
_______________________________________________
Opticks-devs mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/opticks-devs
Reply | Threaded
Open this post in threaded view
|

Re: GSOC 2013 idea

tclarke
Administrator

It looks pretty good. I have a couple of comments.

 

The pre-coding period mentions creating a github repo. You can use your existing repo for the project and just add to it.

 

I like that you have buffer time built into your schedule. I’d also like to see a short paragraph discussing what happens if you still are behind. Your project has a couple of components which don’t strongly depend on each other so I suspect your mitigation strategy will be to deliver fewer, fully functional components.

 

From: mohit kumar [mailto:[hidden email]]
Sent: Friday, April 26, 2013 11:23 AM
To: [hidden email]
Subject: Re: [Opticks-devs] GSOC 2013 idea

 

Hi all,
Sorry for running late.
This is my first draft of the proposal http://researchweb.iiit.ac.in/~mohit.kumar/Opticksgsoc2013.html

Feedback and suggestions are welcome.

Thanks and Regards,

Mohit

 

On Mon, Apr 15, 2013 at 8:24 PM, mohit kumar <[hidden email]> wrote:

Thanks tclarke for the feedback, so what should be the next step?

Or any more addition to the work?

Regards,

Mohit

 

On Mon, Apr 15, 2013 at 7:40 PM, Clarke, Trevor <[hidden email]> wrote:

That sounds good  to me…limiting your data and test cases such that the atmospheric effects don’t need to be addressed is fine.

 

From: mohit kumar [mailto:[hidden email]]
Sent: Friday, April 12, 2013 6:15 AM


To: [hidden email]
Subject: Re: [Opticks-devs] GSOC 2013 idea

 

Thanks Tclarke and Dadkins for your feedback and interest.

I have attached the the paper for the fusion algorithm to this mail.

Regarding the missing data filling algorithm, I just had a look at the paper and did not thought much about it. Initially I want to implement the basic cloud detection without any atmospheric effects. So the scope of this tool stays limited to images without a lot of atmospheric disturbances.

The threshold values for the cloud pixels either could vary from image source or can be user defined or can be set a fixed value for every image.

 

Thanks and regards,

Mohit Kumar

 

On Fri, Apr 12, 2013 at 2:15 AM, Adkins, Dustan <[hidden email]> wrote:

This is a pretty good start.

As tclarke mentioned, you will need to scope your project appropriately, especially the second algorithm.

And we will definitely need to be able to get at the papers you mentioned to mentor this project.

 

--Dustan

 

From: Clarke, Trevor [mailto:[hidden email]]
Sent: Thursday, April 11, 2013 10:02 AM
To: [hidden email]
Subject: Re: [Opticks-devs] GSOC 2013 idea

 

Comments in red..overall, I think it’s a good start and could be an appropriate level of effort for a summer provided the work as proper bounds.

 

From: mohit kumar [[hidden email]]
Sent: Thursday, April 11, 2013 6:50 AM
To: [hidden email]
Subject: [Opticks-devs] GSOC 2013 idea

 

Hi all,

I participated in last year's Summer of Code as a students from Opticks and made a plug-in for object based image analysis. That work included basic object creation, object attribution and change detection techniques.

This year I want to take part in GSOC again and as I am familiar with opticks development, I would love to apply again as an opticks student.

I have a basic idea on extending my last year's work and making a few more tools using object based techniques.

The three tools I have in mind are as follows.

  1. Image Fusion:
    • The Fusion software is a tool fuses a PAN image and a correspondingly registered MSS image. The fused image has salient information from both the images. The fused image can have complementary spatial and spectral resolution characteristics.
    • GLSI fusion[1]. Will do the fusion on each object instead of the 3X3 window as mentioned in the paper.
    • Improved classification. The classification we get on the fused image will be better as compared to MSS or panchromatic classification done separately. 

This seems like it might be a workable change to the G-LSI algorithm but I was unable to find the original paper (just a poster) so I can’t say for sure. I’m guessing the publication is only available to journal members (or for purchase) so I’ll try and track down a copy. Could you send a summary of the G-LSI process with a little more detail than the poster? (or a link to the paper if it’s available on the web)

 

  1. Missing Data Filling
    • The missing data
      filling tool detects the occlusions in a given image.
    • Satellite images can sometimes have occlusions which may lead to loss of information.
    • So, in order to overcome this loss, the missing information is
      filled in by removing these occlusions with the help of similar supporting images of the same area.
    • In general, the occlusions may be clouds, shadows, trees etc. Initially I want to focus on cloud detection and removal using this algorithm[2]. .

This algorithm suggest the use reflectance data with atmospheric affects removed. What tool or algorithm do you intend to use for this? The usual algorithm used commercially is FLAASH (licensing can be expensive..it’s usually included in commercial tools such as ENVI). While not necessary, this step is increasingly important when you have stronger atmospheric effects such as in images with lots of aerosol content (cloudy images, etc.). If you skip atmospheric correction, you’ll need to carefully choose your test data to minimize atmospheric effects.

I assume you’ll be using Landsat 5 data for this? (the paper provides threshold values for TM images…using another source will require and algorithm for identifying cloud and shadow pixels)

  1. Improved Change Detection
    • In addition to the earlier area based approach, this time I will use the texture and shape features of the objects in the image to classify the change .
    • Querying the change image. e.g. the user will be able to visualize "appeared objects which are circular in shape" etc.
    • Classification of the objects prior to change detection to ease the queries. The query can now be " disappeared buildings"
    • The UI will enable the user to modify the change class(appeared, merged, split etc) in the change image if the user is not happy with the change detection output.

References

  1. A. Khandelwal, K. S. Rajan. An E cient Algorithm for Generating Pan Sharpened Quick Look Images.
  2. Q. Meng, B. E. Borders, C. J. Cieszewski and M. Madden. Closest Spectral Fit for Removing Clouds and Cloud Shadows Photogrammetric Engineering and Remote Sensing, 75(5):569 - 576,may 2009.

 

Looking forward for feedback and suggestions.

 

Thanks and Regards,

--

Mohit Kumar

+91-970-3840-175

Lab For Spatial Informatics
International Institute of Information Technology
Hyderabad, India

 
 
This message and any enclosures are intended only for the addressee.  Please 
 
 
notify the sender by email if you are not the intended recipient.  If you are 
 
not the intended recipient, you may not use, copy, disclose, or distribute this 
 
message or its contents or enclosures to any other person and any such actions 
 
may be unlawful.  Ball reserves the right to monitor and review all messages 
 
and enclosures sent to or from this email address.
This message and any enclosures are intended only for the addressee.  Please 
 
notify the sender by email if you are not the intended recipient.  If you are 
 
not the intended recipient, you may not use, copy, disclose, or distribute this 
 
message or its contents or enclosures to any other person and any such actions 
 
may be unlawful.  Ball reserves the right to monitor and review all messages 
 
and enclosures sent to or from this email address.


------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
_______________________________________________
Opticks-devs mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/opticks-devs




--

Mohit Kumar

+91-970-3840-175

Lab For Spatial Informatics
International Institute of Information Technology
Hyderabad, India

This message and any enclosures are intended only for the addressee.  Please 
 
notify the sender by email if you are not the intended recipient.  If you are 
 
not the intended recipient, you may not use, copy, disclose, or distribute this 
 
message or its contents or enclosures to any other person and any such actions 
 
may be unlawful.  Ball reserves the right to monitor and review all messages 
 
and enclosures sent to or from this email address.


------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
_______________________________________________
Opticks-devs mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/opticks-devs




--

Mohit Kumar

+91-970-3840-175

Lab For Spatial Informatics
International Institute of Information Technology
Hyderabad, India




--

Mohit Kumar

+91-970-3840-175

Lab For Spatial Informatics
International Institute of Information Technology
Hyderabad, India


This message and any enclosures are intended only for the addressee.  Please 

notify the sender by email if you are not the intended recipient.  If you are 

not the intended recipient, you may not use, copy, disclose, or distribute this 

message or its contents or enclosures to any other person and any such actions 

may be unlawful.  Ball reserves the right to monitor and review all messages 

and enclosures sent to or from this email address.

------------------------------------------------------------------------------
Try New Relic Now & We'll Send You this Cool Shirt
New Relic is the only SaaS-based application performance monitoring service
that delivers powerful full stack analytics. Optimize and monitor your
browser, app, & servers with just a few lines of code. Try New Relic
and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_apr
_______________________________________________
Opticks-devs mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/opticks-devs
Reply | Threaded
Open this post in threaded view
|

Re: GSOC 2013 idea

mohit kumar
Thanks tclarke for the quick response.
The repository will be the same, I will just add to my existing repository.
My mitigation strategy will be as you said "deliver fewer but fully functional components".

I will include this in my proposal.

Thanks and Regards,
Mohit


On Fri, Apr 26, 2013 at 9:49 PM, Clarke, Trevor <[hidden email]> wrote:

It looks pretty good. I have a couple of comments.

 

The pre-coding period mentions creating a github repo. You can use your existing repo for the project and just add to it.

 

I like that you have buffer time built into your schedule. I’d also like to see a short paragraph discussing what happens if you still are behind. Your project has a couple of components which don’t strongly depend on each other so I suspect your mitigation strategy will be to deliver fewer, fully functional components.

 

From: mohit kumar [mailto:[hidden email]]
Sent: Friday, April 26, 2013 11:23 AM


To: [hidden email]
Subject: Re: [Opticks-devs] GSOC 2013 idea

 

Hi all,
Sorry for running late.
This is my first draft of the proposal http://researchweb.iiit.ac.in/~mohit.kumar/Opticksgsoc2013.html

Feedback and suggestions are welcome.

Thanks and Regards,

Mohit

 

On Mon, Apr 15, 2013 at 8:24 PM, mohit kumar <[hidden email]> wrote:

Thanks tclarke for the feedback, so what should be the next step?

Or any more addition to the work?

Regards,

Mohit

 

On Mon, Apr 15, 2013 at 7:40 PM, Clarke, Trevor <[hidden email]> wrote:

That sounds good  to me…limiting your data and test cases such that the atmospheric effects don’t need to be addressed is fine.

 

From: mohit kumar [mailto:[hidden email]]
Sent: Friday, April 12, 2013 6:15 AM


To: [hidden email]
Subject: Re: [Opticks-devs] GSOC 2013 idea

 

Thanks Tclarke and Dadkins for your feedback and interest.

I have attached the the paper for the fusion algorithm to this mail.

Regarding the missing data filling algorithm, I just had a look at the paper and did not thought much about it. Initially I want to implement the basic cloud detection without any atmospheric effects. So the scope of this tool stays limited to images without a lot of atmospheric disturbances.

The threshold values for the cloud pixels either could vary from image source or can be user defined or can be set a fixed value for every image.

 

Thanks and regards,

Mohit Kumar

 

On Fri, Apr 12, 2013 at 2:15 AM, Adkins, Dustan <[hidden email]> wrote:

This is a pretty good start.

As tclarke mentioned, you will need to scope your project appropriately, especially the second algorithm.

And we will definitely need to be able to get at the papers you mentioned to mentor this project.

 

--Dustan

 

From: Clarke, Trevor [mailto:[hidden email]]
Sent: Thursday, April 11, 2013 10:02 AM
To: [hidden email]
Subject: Re: [Opticks-devs] GSOC 2013 idea

 

Comments in red..overall, I think it’s a good start and could be an appropriate level of effort for a summer provided the work as proper bounds.

 

From: mohit kumar [[hidden email]]
Sent: Thursday, April 11, 2013 6:50 AM
To: [hidden email]
Subject: [Opticks-devs] GSOC 2013 idea

 

Hi all,

I participated in last year's Summer of Code as a students from Opticks and made a plug-in for object based image analysis. That work included basic object creation, object attribution and change detection techniques.

This year I want to take part in GSOC again and as I am familiar with opticks development, I would love to apply again as an opticks student.

I have a basic idea on extending my last year's work and making a few more tools using object based techniques.

The three tools I have in mind are as follows.

  1. Image Fusion:
    • The Fusion software is a tool fuses a PAN image and a correspondingly registered MSS image. The fused image has salient information from both the images. The fused image can have complementary spatial and spectral resolution characteristics.
    • GLSI fusion[1]. Will do the fusion on each object instead of the 3X3 window as mentioned in the paper.
    • Improved classification. The classification we get on the fused image will be better as compared to MSS or panchromatic classification done separately. 

This seems like it might be a workable change to the G-LSI algorithm but I was unable to find the original paper (just a poster) so I can’t say for sure. I’m guessing the publication is only available to journal members (or for purchase) so I’ll try and track down a copy. Could you send a summary of the G-LSI process with a little more detail than the poster? (or a link to the paper if it’s available on the web)

 

  1. Missing Data Filling
    • The missing data
      filling tool detects the occlusions in a given image.
    • Satellite images can sometimes have occlusions which may lead to loss of information.
    • So, in order to overcome this loss, the missing information is
      filled in by removing these occlusions with the help of similar supporting images of the same area.
    • In general, the occlusions may be clouds, shadows, trees etc. Initially I want to focus on cloud detection and removal using this algorithm[2]. .

This algorithm suggest the use reflectance data with atmospheric affects removed. What tool or algorithm do you intend to use for this? The usual algorithm used commercially is FLAASH (licensing can be expensive..it’s usually included in commercial tools such as ENVI). While not necessary, this step is increasingly important when you have stronger atmospheric effects such as in images with lots of aerosol content (cloudy images, etc.). If you skip atmospheric correction, you’ll need to carefully choose your test data to minimize atmospheric effects.

I assume you’ll be using Landsat 5 data for this? (the paper provides threshold values for TM images…using another source will require and algorithm for identifying cloud and shadow pixels)

  1. Improved Change Detection
    • In addition to the earlier area based approach, this time I will use the texture and shape features of the objects in the image to classify the change .
    • Querying the change image. e.g. the user will be able to visualize "appeared objects which are circular in shape" etc.
    • Classification of the objects prior to change detection to ease the queries. The query can now be " disappeared buildings"
    • The UI will enable the user to modify the change class(appeared, merged, split etc) in the change image if the user is not happy with the change detection output.

References

  1. A. Khandelwal, K. S. Rajan. An E cient Algorithm for Generating Pan Sharpened Quick Look Images.
  2. Q. Meng, B. E. Borders, C. J. Cieszewski and M. Madden. Closest Spectral Fit for Removing Clouds and Cloud Shadows Photogrammetric Engineering and Remote Sensing, 75(5):569 - 576,may 2009.

 

Looking forward for feedback and suggestions.

 

Thanks and Regards,

--

Mohit Kumar

+91-970-3840-175

Lab For Spatial Informatics
International Institute of Information Technology
Hyderabad, India

 
 
This message and any enclosures are intended only for the addressee.  Please 
 
 
notify the sender by email if you are not the intended recipient.  If you are 
 
not the intended recipient, you may not use, copy, disclose, or distribute this 
 
message or its contents or enclosures to any other person and any such actions 
 
may be unlawful.  Ball reserves the right to monitor and review all messages 
 
and enclosures sent to or from this email address.
This message and any enclosures are intended only for the addressee.  Please 
 
notify the sender by email if you are not the intended recipient.  If you are 
 
not the intended recipient, you may not use, copy, disclose, or distribute this 
 
message or its contents or enclosures to any other person and any such actions 
 
may be unlawful.  Ball reserves the right to monitor and review all messages 
 
and enclosures sent to or from this email address.


------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
_______________________________________________
Opticks-devs mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/opticks-devs




--

Mohit Kumar

+91-970-3840-175

Lab For Spatial Informatics
International Institute of Information Technology
Hyderabad, India

This message and any enclosures are intended only for the addressee.  Please 
 
notify the sender by email if you are not the intended recipient.  If you are 
 
not the intended recipient, you may not use, copy, disclose, or distribute this 
 
message or its contents or enclosures to any other person and any such actions 
 
may be unlawful.  Ball reserves the right to monitor and review all messages 
 
and enclosures sent to or from this email address.


------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
_______________________________________________
Opticks-devs mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/opticks-devs




--

Mohit Kumar

+91-970-3840-175

Lab For Spatial Informatics
International Institute of Information Technology
Hyderabad, India




--

Mohit Kumar

+91-970-3840-175

Lab For Spatial Informatics
International Institute of Information Technology
Hyderabad, India

This message and any enclosures are intended only for the addressee.  Please 

notify the sender by email if you are not the intended recipient.  If you are 

not the intended recipient, you may not use, copy, disclose, or distribute this 

message or its contents or enclosures to any other person and any such actions 

may be unlawful.  Ball reserves the right to monitor and review all messages 

and enclosures sent to or from this email address.

------------------------------------------------------------------------------
Try New Relic Now & We'll Send You this Cool Shirt
New Relic is the only SaaS-based application performance monitoring service
that delivers powerful full stack analytics. Optimize and monitor your
browser, app, & servers with just a few lines of code. Try New Relic
and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_apr
_______________________________________________
Opticks-devs mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/opticks-devs




--
Mohit Kumar
+91-970-3840-175
Lab For Spatial Informatics
International Institute of Information Technology
Hyderabad, India

------------------------------------------------------------------------------
Try New Relic Now & We'll Send You this Cool Shirt
New Relic is the only SaaS-based application performance monitoring service
that delivers powerful full stack analytics. Optimize and monitor your
browser, app, & servers with just a few lines of code. Try New Relic
and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_apr
_______________________________________________
Opticks-devs mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/opticks-devs