Indices like NDVI, NDBI are widely adopted by the scientific community to identify distinct land cover types. Prominently, Otsu's threshold estimation is adopted. But how do we select the lower and upper limits of the threshold for a given index (manually) and are there other better techniques that can be adopted?
Dharmendra Singh
kindly refer to the publication
https://www.sciencedirect.com/science/article/abs/pii/S2352938521000227
0 votes
0 thanks
Gamal Seedahmed
Otsu and k-means can be seen as hard clustering techniques. You need to try the expectation maximization since it considers the membership to clusters in a probabilistic since.
0 votes
0 thanks
More Ashish Mishra's questions See All
Similar questions and discussions