Research and design on the train fixed hook detection system based on the improved FCM

摘要

In line with the characteristics of using color to symbol train carriages in Qinhuangdao port, this paper adopts the improved FCM algorithm, which is improved on CIE Lab space and ant colony algorithm and combined with wavelet denoising to realize color clustering. This method uses global and robustness advantages of ant colony algorithm, combines and sets clustering center through A and B component of CIE Lab space, which is to be initial center of FCM and can overcome FCM local optimal faults. Finally, the ARM9 experimental platform based on Linux system would verify the algorithm, which shows that this system can achieve higher identification accuracy and can meet the application requirements on the detection system.