Determination of Injury Rate on Fish Surface Based on Fuzzy C-means Clustering Algorithm and L*a*b* Color Space Using ZED Stereo Camera

Determination of injury rate on fish surface is one of the major cause for increasing quality of fish on the markets. Detecting injury fishes manually is hard work with low efficiently. An image processing method is usually considered to do this work as an online selecting method of injury area. The injury area segmentation of fish image is based on color features with Fuzzy C-means clustering and L*a*b* color space. Although there are many different color spaces. CIE L*a*b* color space of them is most used for detecting fish injury due to its uniform color distribution and is closest to the one human eye. Generally, a fish injury image has overlapped data with fish body image. The K-means clustering algorithm cannot give the good solution to segment whether pixels of the overlapped data belong to fish body or injury area. To solve this problem, this paper is to present the good solution for segmentation of fish injury image with the overlapped data from the fish body image and determine injury rate on fish surface based on Fuzzy C-means clustering and L*a*b* color space using ZED stereo camera. The proposed image processing method is tested on fishes. The experimental results show that the injury rate measured using the proposed image processing method is close to the real injury rate.