Abstract:

The cold rolling strip surface defect detection system based on machine vision was designed with multiclassification and modularization idea. For realizing realtime characteristics of system, highspeed linear CCD camera and relational highspeed fiber image collection card, highspeed FPGA image processing card designed by ourselves were adopted in hardware; RTAILinux operation system and parallel processing method were adopted in software. Image segmentation parallelization experiment result indicates that the processing speed of quadcore processor is 3.2 times faster than that of single core; and the system’s performance is improved effectively.

Key words: surface defect, image process, modularization, parallel processing