Egg crack image detection method based on improved grasshopper optimization algorithm and canny operator

Abstract

Objective: In order to improve the detection effect of egg linear and reticular cracks. Methods: As the low convergence efficiency of Grasshopper optimization algorithm （GOA） in solving high-dimensional complex optimization problems, an improved fuzzy c-means algorithm （FCM） was designed to classify locust population. The adaptive extreme value reverse learning and coding mutation update mechanism were designed to expand the depth search space and global optimization ability of the algorithm. The improved GOA was used to optimize the parameters, and the improved canny operator was used for egg crack detection. Results: The results showed that the missed detection rates of egg linear crack and mesh crack were improved by about 21.4%~31.2% and 63.2%~69.7% respectively, which was better than other algorithms. Conclusion: This method can effectively improve the accuracy of egg crack detection.

Keywords

• egg
• crack
• grasshopper optimization algorithm
• fuzzy c-means
• missed detection rate