Design Of Online Automatic Grading System For Apple Picking Robot

At present, the majority of fruits is graded manually in the world, only a small part is done with machines, all of which work in offline (post-harvest grading). This grading approach not only increases the cost and intermediate links of fruit production, but also counts against maintaining the quality of fresh fruits. In view of the above, one kind of autonomous online grading system for fruit picking robot was developed, the system not only can achieve the fruit autonomous grading in the process of harvest, but also can improve productivity, reduce labor intensity and ensure the quality of fresh fruits, etc.Firstly, an online platform for autonomous apple grading was built, with which the characteristic parameters of weight and color of harvested apples can be extracted and graded apple can be shipped to corresponding boxes.Secondly, focusing on the two significant issues of autonomous grading of apples, i.e., apple weight and coloring rate detection, a detection methods to extract the full area of apple was proposed, including image acquisition, preprocessing and image feature extraction, and apple weight detection system was designed, including the detection circuit, the communication module, the motor module, and the PCB process integration.Finally, the mechanism for grading apples was developed, apples were graded according to extracted features based on multi-information fusion analysis. With weight analysis method, weighting factors were selected according to optimal curve.In the laboratory environment, experiments were carried to virify the performance of designed online autonomous grading system.The results show that the classification system is able to collect apples information in real-time and grade apples by weight and color characteristics. The classification accuracy rate is above 89%.