A Super Pixel Segmentation Algorithm for SAR Images Based on APKSLIC

To improve the accuracy of Simple Linear Iterative Clustering (SLIC) algorithm in image segmentation, a superpixel segmentation algorithm based on adaptive parameters and Kuan filter mean estimation is proposed. This algorithm adaptively generates feature equalization parameter g based on the specific situation of the image, and then combines it with the average grayscale difference and spatial fluctuation difference constructed by Kuan filter estimation to improve the superpixel segmentation effect of SAR images. The experimental results show that compared with the other three superpixel segmentation algorithms, the proposed algorithm has a recall improvement of at least 0.7
^%
, an undersegmentation error rate reduction of at least 5%, and an average running time reduction of at least 15%.