YOLOv3-Darknet with Adaptive Clustering Anchor Box for Intelligent Dry and Wet Garbage Identification and Classification

The implementation of dry and wet garbage classification is conducive to reducing waste land occupation, reducing pollution and improving the utilization rate of resources, which has social, economic and ecological benefits. At present, there are many shortcomings in manual classification, such as: high work intensity, poor environment, high cost. In the traditional detection method, the robustness is poor, the accuracy is low, can't achieve rapid detection in a complex background. We propose an anchor box and yolov3-darknet model based on adaptive data clustering to identify, classify and detect dry and wet garbage. The experimental results show that our method can adapt to the classification of dry and wet garbage in complex environment, and can accurately and quickly identify the dry and wet garbage, ensure the purity of wet garbage above 90%, and meet the requirements of secondary treatment of garbage.