In this paper, an approach, which is used to detect such rice quality parameters as chalkiness automatically based on computer vision technology in place of human visual observation, is presented. The approach meets the needs of automation in the agricultural engineering. To verify the feasibility of the approach, the research develops an algorithm using MATLAB software. Experiments are conducted on different rice images to measure the chalkiness parameters, i.e. degree of chalkiness and rate of chalky grains. The results show that the approach is effective and reliable.