Support vector machine and YOLO for a mobile food grading system

Food quality and safety are of great concern to society since it is an essential guarantee not only for human health but also social development, stability. Ensuring food a complex process. All processing stages should be considered, from cultivating, harvesting storage preparation consumption. Grading one the processes control quality. This paper proposed mobile visual-based system evaluate grading. Specifically, acquires images bananas when they on moving conveyors. A two-layer image based machine learning used grade bananas, these two layers allocated edge devices cloud servers, respectively. Support Vector Machine (SVM) first layer classify extracted feature vector composed color texture features. Then, You Only Look Once (YOLO) v3 model further locating peel's defected area determining if inputs belong mid-ripened or well-ripened class. According experimental results, layer's performance achieved accuracy 98.5% while second 85.7%, overall 96.4%.