Rearch Of Garbage Classification And Recognition Based On Faster R-CNN Algorithm

In order to protect environment,how to garbage waste sorting and recycling is an important issue.However,in our daily lives,the work of garbage sorting is mainly done manually.This kind of operation has the disadvantages of large workload,low efficiency and inaccurate classification.In order to solve the above problems,this paper proposes an automatic classification and identification method,so as to improve the efficiency of garbage classification.The garbage classification algorithm proposed in this paper is a Faster R-CNN method based on SMC filter.The method iterates the target image into a set of samples,thereby enabling the Faster R-CNN detector to refine the image.The implementation of this paper is mainly based on the following two strategies:(1)Bounding box strategy: By adjusting the architecture of the Faster R-CNN deep network,the bounding box of the candidate of the domestic garbage object is proposed.This strategy gives a set of recommendations consisting of garbage objects predicted by the Faster R-CNN output layer.(2)Verification strategy: Select unlabeled samples from the target scene.The strategy utilizes the association between the confidence score returned by the Faster R-CNN output layer and the information extracted from the target image to facilitate selection of positive samples from the target scene and to reduce the introduction of error markers in the training data set.risks of.In the experiments,this paper uses a network image dataset and a laboratory image dataset,and conducted experiments separately.The experimental results show that the method of this paper has higher recall and precision,so the method of this paper can identify domestic garbage more effectively.