Abstract: At present people is higher and higher requirement for rice quality, testing quality identification technology is also used in agriculture in many aspects, for detecting precision and quality of rice quality of single equipment, the problems such as too much manual intervention, to that end, this paper designed a by visual recognition technology of rice quality analysis device, respectively introduces the work principle and important parts, MatLab software was used to establish the corresponding function and Canny edge detection algorithm model to accurately detect the number of rice and the identification of waist burst grains after drying and realize online weighing, improve the automatic detection level and work efficiency of rice quality device, which has important research significance to improve the detection of waist burst rate of rice and efficient identification number.