I was testing the lin and log polar again. My apologies to @kwagyeman, I did not properly understand it, which is why I did not think it would work. However, using it with the reverse argument set to False returns an image with a seam since I am using the conic omnidirectional mirror, and since the objects I am looking for may be at the mid-top or mid-bottom, blob search might have trouble with a blob split into one half on one side of the image and one on the other. Hence, is there a way to allow find_blobs to wrap?
Thanks
Find blobs looks at the image as a rectangular array. It does not care about the geometry of the image.
Can you provide a sample image?
Real setup, camera view and logpolar view are attached.
As you can see in the annotations on the log polar image, the blue is split in two. Is there any way to avoid this? I can only see two ways: moving the seam(perhaps by rotating the image before applying the filter) or by allowing blob detection to wrap over from one edge of the image to the other side.
Oh, I see. Nope.
You have to detect when you see two blue blobs on the left and right side of the image and then combine them into one. You can do this in software. This is just what you have to deal with to use this filter.
Um, regarding image quality. You should adjust the contrast to make the image pop a little more. Or just turn off auto white balance immediately after 300 ms versus 2 seconds.