Fruit Traceability and Quality Inspection System Based on Blockchain and Computer Vision

Abstract

In traditional fruit traceability systems, the opacity of the process and the ease of data tampering have always been major challenges. To address this, we propose a fruit traceability system based on blockchain and computer vision in this paper. In terms of data storage, we chose the Ethereum blockchain to ensure the authenticity of the data. As the fruit passes through each node, we upload the fruit image and conduct quality checks. We trained the YOLOv5 model to identify fruit types, and on this basis, we trained an improved Unet model for saliency detection to extract the fruit image area, both of which performed very well. Based on the fruit image, we evaluate from four aspects: saturation, hue, defects, and shape. This system can be used for traceability throughout the fruit supply chain.