Real-time Image Processing System Based on Multi-core Processor

This paper proposed a real-time image processing system based on multi-core processor, which is used in the cold rolling surface defect inspection system for enhancing system´s real-time performance. Based on analyzing the application demand of inspection system image processing module, the image processing system is designed that is the combination of RT-Linux real-time operation system and multi-core processor hardware platform. Taking advantage of RT-Linux multi-task scheduling mechanism based on priority, the PCI device driver is developmented to ensure receiving and saving image data in real-time. At the same time, multi-core processor makes multi-task parallel processing to be possible. Two parallelism types are analyzed, including task-level parallelism and data-level parallelism. Finally the combination of thread-level and data-level parallelism of image segmentation based on Ostu method is implemented, verifying performance of real-time image processing system based on multi-core processor. Segmentation algorithm can efficiently utilize the multi-core processor, 3.2 times faster than a single core.