Reconfigurable vision based sensory system for extraction of component properties

Integrating computer vision for industrial automation involves providing computers with the functions typical of human vision. The focus of this research is to develop the capability to effectively compare PC performance versus FPGA for image processing tasks. The algorithms are first tested within a PC environment and their performance measured on various PC platforms. The algorithms with the most promise for parallel execution or speedup are selected and transported into an FPGA environment. The significance of comparing the performance of PC versus FPGA environments for vision applications is not merely for research purposes, but also to transport and apply vision projects to this increasingly popular and versatile platform. The vision algorithms for creation of a vision application capable of selected image processing and analysis functions have been developed. They include storing an image into memory, the application of noise filters, segmenting the image (with particular attention paid to thresholding) and performing connected component analysis