Hello everyone!
I was working on light pollution and nighttime images lately and someone suggested me to use Otsu's thresholding method. I understand the basic idea and calculation of Otsu's thresholding and that it throws a value that we use to classify our data into two parts. What I dont get is how do we determine the very initial value of the pixel that classifies the image into background and foreground. Is it purely researcher's call to use any value?
Please help!