Finding blobs(wide lines in an image) and cropping them out when Background is not clear

Hello,

I have adapted from the code BlobDemo.m by the expert (Image Analyst) here in mathworks forum. His answers to questions have been very helpful to me. Thanks for that. I am a learner and trying to work on image processing of Lateral flow strips. (Test lines and Control lines in the strip).

I have two images:

(a) a nicely cropped image('scanned_f10.jpg') of a strip (with white background and visible Strip lines)

(b)a camera shot image ('cam_e8.jpg')( Background is non-white and so filtering is to done as well, I think so.... I am not sure. Please guide me )

When I run the program with "baseFileName = scanned_f10.jpg" in the code attahced(T_C_lines_Blob_analysis.m), the program gives me perfectly what I want ( fig attached as 'f10_blobs.jpg') and the blob measurements.

(I am bit confused with the Image processing steps I performed. Please help)

How to perform thresholding for such images?

My issues in which I need help. Thank you in advance

Now, I use "baseFileName = cam_e8.jpg" in the code. I need to make changes to the code to perform proper thresholding and remove the blue background, crop the lateral strip and then find bounding boxes and other blob measurements.

Please help me in getting the Blobs from the camera shot image and also correct me in my first code as well where I have done wrong.

Thanks a lot.

[Image Analyst]

What do you want? Do you want the bounding boxes like the one image has? Do you want two output images, each cropped out from the image with imcrop()? Have you tried to flatten the image a bit using adapthisteq()?

[symmer cul]

Yes, I intend to get the bounding boxes for "2 wide strip lines" in the camera shot image ( like the one in the image "f10_blobs.jpg"(obtained from scanned image given). While doing so, I further want to crop them out for further analysis.

I did use the imcrop to crop out the required blobs(in scanned_image). The extended code is attached for reference. Cropped T and L lines are also shown in the attached figure.

I haven't tried the adapthisteq(). Please guide me on that.

Thresholding the camera shot image, and subsequently PICKING only the lateral flow strip first , and then detecting blobs within the lateral flow next, was in mind. You may correct me on that.

Uneven illumination, spread of colour throughout the lateral flow strip and the blue background is causing me trouble.

(Black background, proper positioning(rather same positioning of lateral flow strips during image capture), proper illumination, PNG, rather than JPG files were all advised by you at other places in this forum . I made a note of these.)

[symmer cul]

The other thing I am unable to solve is the mapping or finding a relation between those "changed" pixel values (in the bounding boxes) to light absorption values. Appreciate it, if I can be helped.

Certain chemical is added to the lateral flow strips, and with an increase in addition of certain concentration of chemicals, the intensities of the Test-line becomes darker (lower pixel values) or remains brighhter(higher pixel values) with less chemical addition. I need to plot a graph between absorption vs wavlength. What I have now, is pixel intensity values.

[symmer cul]

@Image Analyst

Could you please help me with my queries?