Research And Design On Embedded Fruit Automatic Classification System

Due to the backwardness of fruit detection and classification, fruits are mostly primary products in our country. In recent years, the high-tech fruit industry aboard has brought a great charge to the domestic traditional low-tech fruit industry with the reduction of the fruit tariff of our country. The detection and classification of fruit in China mainly rely on manual labor, leading a low production capacity and high cost.Nowadays, the large domestic fruit manufacturers mainly rely on imported equipment for automatic classification of fruit, the medium and small size fruit manufacturers cannot afford those expensive equipment. Domestic research on the automatic fruit classification system is still in the theoretical research stage, a few of them used for industrial production because of high demanding for platform and low real-time classifying for fruits.This paper designs a low cost embedded fruit automatic classification system aiming at the small and medium size fruit manufactures. This paper researches the real-time fruit classification algorithm and applies it in the high cost-effective embedded platform. In this paper, the fruit video is collected in real time,achieves the fruit real time classification in the video and designs the real time fruit classification algorithm based on decision tree. The test results show that the advantages of this system include low cost, high accuracy and good real-time performance. The system can achieve real-time and efficient classification of fruits, meeting the needs of practical applications.This paper mainly completes the following works:(1) Transplanting Raspbian operation system. Building cross compiler environment. Selecting the CMOS image acquisition equipment according to the application needs. Transplanting OpenCV3.1 image library.(2) Getting the fruit video image in real time by camera. Handing the frame image by gray level conversion, binaryzation, filtering and contour detection preprocessing.Extracting the fruit contour and detect the fruit image in the video frame.(3) Splitting the fruit image in video frame. Extracting the key feature of the fruit image and stored in the Raspberry Pi. Designing and implementing a real-time classification algorithm for fruits that match this system. Completing the design of automatic fruit classification system.(4) Transplanting a new algorithm for automatic classification of fruits in raspberry pie. Achieving a real time classification of fruits in embedded system.