Machine Vision Applications in Agricultural Food Logistics

Agricultural food's logistics needs to be efficient and to provide assurance on the safety and quality of its products which consumers could trust. This paper designs a machine vision system by which fruits or vegetables can be detected for defects and damages during transportation and storage. The color histogram extracted in local image patch is used as image feature and the Linear SVM (Support vector machine) is used for model learning, which provides good robustness, higher accuracy and modest calculation costs. In a case of apple inspection, our system realizes a recall rate of 96.8% and a false detection rate of 1.6%. By the output of this inspection, agri-food producers are able to prevent the products with deformity and blemishes from reaching the end customers, thereby the safety and quality of the agri-food markets can be guaranteed.