Method Of Apple Automatic Grading Based On Machine Vision

Grading is very most important after fruit is adopted. When machine vision is applied in grading,there are many advantages such as no damage to fruit,high-speed. The machine vision applied in the fruit sorting was studied in the paper.In fact, it is difficult to distinguish stem or calyx from default on the face of apple by computer and it is also time-consuming. The idea of combining orientation device with machine vision was supported. The mechanical device instead of computer completed partial tasks, which could improve grading speed. The basic requests of the equipment design were put forward and the orientation experimental device was designed.In order to improve the quality of image acquisition and speed image processing, the background color was studied. The results showed that the black was contributed to the separation of target and background. Super median filter method was employed in the lower-layer image processing to improve image quality. The threshold was applied to segment apple image and contour was followed. And clear and smooth contour could be got.Based on analyzing the traditional method of the extraction of fruit features, author used equivalent diameter instead of apple radius in the aspect of apple size and circular degree to describe in the aspect of the apple shape. The HSI color model was used to describe the color feature of apple when the fruit was graded based on the color. The four chrome average values were substituted for the chrome of apple according to the characteristic of apple chrome histogram, on which BP neural network system was built. The experimental results showed that the system could grade apple precisely and satisfy grading requests.