Garbage sorting system based on improved single shot multibox detector

Garbage pollution is a very difficult problem in environmental governance. Due to the many sources of garbage pollution and a wide range of impacts, this problem is only slow to solve by human means. In order to improve the automation of garbage disposal, on the one hand, this paper proposes a garbage detection method based on CNN (convolutional neural network) using multi-layer feature processing. On the other hand, the detection algorithm is combined with an industrial robot to form a complete garbage sorting system. This paper uses the one-stage idea to first optimize the backbone structure to improve the extraction effect of shallow features. Then the attention module is introduced to make the network pay more attention to information that plays a key role in garbage detection. Finally, a multi-layer feature fusion method is used to combine the features of the shallow network with the features of the deep network to generate a fused feature map for use in target detection tasks. The experimental results show that the detection speed of the method proposed in this paper is 13.75% higher than that of SSD, and the garbage detection accuracy reaches 99.5%, which is better than the SSD detection algorithm. The garbage detection method proposed in this paper can quickly realize the precise positioning of garbage and complete automatic robot sorting.