EDGE detection of thermal images? n then how to find the local maxima from edge image?

Hi, I need help for thermal images.

Answer 1

[Image Analyst]

To find local maxima, use imregionalmax().

To find local minima, use imregionalmin().

Answer 2

[Malini]

Hi, I need to extract plant image from the background which is not maintaining a constant temperature. It is a thermal image. How to apply imregionalmax() in that case

[Malini]

How to apply imregionalmax() for the above two images where the background temperature are different

[Image Analyst]

maxImage1 = imregionalmax(image1);
maxImage2 = imregionalmax(image2);

[Malini]

Hi Image Analyst,

I am working with NIR images of plants with different water levels. Tell me how to segment only plant area from the NIR image and get the mean intensity values.

I=imread('R1.png');

g= rgb2gray(I);
I_eq=adapthisteq(g);

Sobl= edge(I_eq,'Sobel');
figure, imshow(Sobl);title('Sobel of Hist equalization');

bw=im2bw(Sobl,graythresh(I_eq));
figure,imshow(bw), title(' adaptive histogram equalization');

bw2= imfill(bw, 'holes');
figure, imshow(bw2), title('Imfill');

bm=boundarymask(bw2);
figure, imshow(bm),title('Boundary mask');

[Image Analyst]

First you need to correct your horrible illumination problem. It would be SO much easier if there wasn't a gigantic noisy hot spot in your image. Otherwise to first estimate a background to "undo" it would be a lot of complicated code (more than I can do in 5 minutes).

Once you have a uniform illumination, you can find it simply by global thresholding.

[Malini]

Thanks. You mean to say that, I should have an image without the plant and work on it like a background correction?

[Image Analyst]

Yes. First of all, take a photo without the plant and make sure that none of the image is clipped to 255. Take a histogram to make sure. Then take an image, with the same exposure, with the plant in there and divide it by the background image. See my background correction demo, attached.

[Image Analyst]

Are you able to get an image without the plant in there? Snap one like that, without the hotspot in the middle, and upload it.

[Malini]

Since the camera is fixed and surrounded by yellow bulbs, the hotspot is the part of the image in the middle.

[Image Analyst]

Rather than keep bugging VPS, why don't you start your own new thread?