The study of illumination compensation correction algorithm

摘要

Machine Vision applicated in agriculture can be divided into quality inspection and grading of agricultural products, fruit and vegetable picking robots. But there are illumination changes in machine vision systems, making the latter part of the image segmentation difficult. Therefore, how to eliminate the illumination effect is particularly important in the image processing of agricultural products. This paper mainly studies the method of illumination invariant image based on the blackbody radiation theory, does the light compensation experiments on the image of actual fruit, and compares this method with several common methods of illumination compensation or correction. The result of experiment shows that the method of illumination invariant image has the better effect on illumination compensation, helping to improve the accuracy of subsequent image segmentation.