Segmentation of Underwater Objects using CLAHE Enhancement and Thresholding with 3-class Fuzzy C-Means Clustering
Shivendra Singh, Manish Kumar Soni, Ravi Shankar Mishra · 2014
Abstract — Underwater images suffers from low illumination and poor contrast due to refractions of light rays and poor visibility. Therefore, underwater image segmentation and object extraction is a difficult task. This paper proposed an efficient and fast underwater image segmentation method using thresholding with class 3 fuzzy C-means clustering and CLAHE enhancement method. CLAHE enhancement method is used before image segmentation to improve the contrast and illumination of underwater image this in turn improves the segmentation performance significantly. The proposed method uses normally distributed pseudorandom numbers generator to initialize the fuzzy membership function. This modification improves the convergence rate of the standard FCM method. Results of the proposed object segmentation method are tested open the different kind of underwater images. It is found that entropy of segmented object is improved with the proposed method. Paper also compares the performance of FCM with different distance masers.