Public Lab Research note


Question: doubts about processing NIR images.

by ahcalori | July 09, 2016 16:43 | 94 views | 8 comments | #13274 | 94 views | 8 comments | #13274 09 Jul 16:43

Hi,

I have a modified camera to capture NIR images. In the last events, I have been experienced a systematic problem. All the pics taken appear not proper to use in NDVI analysis - the top right corner always return low values of NDVI (fig. 2).

Follow some of my examples: The first image is a NIR-raw (fig. 1). The second, is a NDVI image processed using ArcMaps software (fig. 2).

Ps: This problem happens independent of the hour of day, excluding the possibility of shading.

Does anyone has a suggestion for this issue?

Thank you very much!

NIR_image.jpg Fig 1. NIR-raw image.

NDVI_image.jpg Fig 1. NDVI image.


8 Comments

The color gradient in the NIR photo from lower left to upper right is conspicuous.This could be caused by the ambient lighting or by the camera. Without detailed information about the camera, how it was converted to capture NIR, and how it is being used, it is hard to speculate about the cause. A photo of the same scene with a normal camera could rule out ambient lighting as a cause.

Chris

Reply to this comment...


@cfastie, Thanks for reply. The camera that has been used is one bought in the 'Public Lab Store'. The curious fact is that some pics does not has this effect. Follow early images, NIR and NDVI, respectively:

1.jpg

1_NDVI.jpg

Reply to this comment...


Although the diagonal color gradient is not conspicuous in the NDVI image, it is present in the NIR photo (pinkish in the lower left and bluish in the upper right). There is still no evidence favoring either ambient light or the camera as the cause. I don't know which camera this is (Mobius?) or how it was modified to become an NIR camera.

Chris

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


@cfastie, Is not the Mobius camera. Is just one with modified filters. I will post the specification here. Thanks!

Reply to this comment...


Hi, dear @cfastie, the webcam that I have been used is described in the follow link: https://www.adafruit.com/product/1722. The filter used allow take an infrared photo in the "red" channel of the camera, and a visible image in the "blue" channel.

Do you know a better software to use? I'm suspecting of the software used (ArcMaps).

Besides the NDVI image, I need of numeric values, as mean, minimum and maximum.

Thank you very much!

1722-00.jpg

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


The color gradient in the photos taken by your modified camera seems to be the cause of the artifacts in your NDVI image. To determine whether this is due to the camera or to the ambient lighting, take a new photo that shows the gradient and then rotate the camera 90° and take another photo of the same scene.

I suspect the artifact is due to the camera. Did you buy Rosco #2007 filter from Public Lab, cut a small piece and install it in the web cam? In front of the lens or behind the lens? If so, I suspect the piece of filter is not perfectly flat and perpendicular to the optical path. If you adjust the filter the problem could be fixed.

Did you also buy some red filter? Red filter sometimes works better with the little CMOS sensors in web cams.

Chris

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


@cfastie. I think you are right. The problem must be the filters. I will open the cam and see what happens. I'm just using the product the way I bought - no modification was performed. Can you recommend a software besides ArcMaps, that calculate values as mean, maximum, etc? Thanks again!

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


ImageJ and Fiji have very powerful tools for extracting, summarizing, and displaying the NDVI values in the floating point images produced by Ned's Photo Monitoring plugin: https://publiclab.org/notes/nedhorning/01-13-2016/packaged-photo-monitoring-plugins-available-on-the-github-repositoy

Reply to this comment...


Login to comment.

Public Lab is open for anyone and will always be free. By signing up you'll join a diverse group of community researchers and tap into a lot of grassroots expertise.

Sign up