Comparing the Gambini's Algorithm and a CFAR Method to Edge Detection for Polsar Images
Anderson A. De Borba, Maurício Marengoni, Alejandro C. Frery · IGARSS 2022 - 2022 IEEE International Geoscience and Remote Sensing Symposium · 2022
We provide a comparison between competitive approaches, the Gambini's algorithm (GA) and a CFAR method (The Co-efficient of Variation Detector), to highlight their strengths and weaknesses. In doing so, we help researchers and practitioners making informed decisions towards the best technique to be applied in any situation of interest. Both methods were able to detect edges in the simulated images used in this work. The GA has the advantage of finding a single pixel as an edge while the CFAR method detects multiple edges, mainly when the edge is a ramp edge. In the other hand, the CFAR method can be directly applied to the images while the GA needs to have as input a region of interest and lines where the transition point between distributions (edges) are detected.