Low-overhead household garbage classification method based on dense convolutional network

Abstract

The invention discloses a low-overhead household garbage classification method based on a dense convolutional network, and the method comprises the steps: (1), preprocessing data, (2), building the dense convolutional network, (3), dividing a data set into a training set, a verification set and a test set, and (4), selecting a proper optimizer and a loss function in a training process, setting hyper-parameters, and setting an evaluation index as the accuracy rate. According to the method, optimization is carried out in preprocessing of input data, matrix fusion is carried out on a three-channel color image and an edge detection image to serve as input of a model, and feature information is enhanced; a dense convolutional network structure is constructed; a Dropout layer is additionally arranged; a learning rate self-adjusting and hyper-parameter adjusting method is used.According to the method, the model has enough feature extraction capability; the feature mapping of the model is usedas the input of a subsequent layer; the problem of gradient disappearance caused by a deep network is relieved; good balance is realized on the aspects of low overhead and high precision; and 90.8% of precision and 5.08 M of file size are realized.

Description

technical field [0001] The invention belongs to the technical field of computer image recognition, and relates to a low-overhead domestic waste classification method based on a dense convolution network. Background technique [0002] Most of the current garbage classification and recognition methods are to transfer weights and network structure models and then fine-tune them. However, the existing transfer learning garbage classification methods focus more on performance improvement, often ignoring the portability of the model (low overhead characteristics). Although the deep and complex network improves the performance, it also brings the disadvantage of large model size. For example, the VGG network model file size is 500M, and the Densenet-121 model file size is 32M. On the current embedded platform: most of the STM32 series With only 1M storage space, although the model can be compressed and embedded, the resources are still very limited, so the model cannot be embedded...

AI Technical Summary

Problems solved by technology

[0003] The purpose of the present invention is to provide a low-overhead domestic waste classification method based on a dense convolutional network, which solves the problem that the network structure model resources of the domestic waste classification method in the prior art are limited, and the model cannot be implanted into an embedded platform, which is difficult for promotion. It has brought difficulties; and the technical problems of simple shallow network generalization ability and low recognition accuracy

Images