Vehicle Identification Based On Deep Learning And Its Application In Tax Monitoring

In order to prevent or avoid the occurrence of tax evasion due to non-invoicing in mineral resources such as coal and sandstone mining,install cameras at the entrances and exits of mineral resource enterprises to capture vehicle shipments,and ensure that product sales confirmations are issued for each vehicle and one ticket.The department is an important means to supervise mineral resources enterprises.With the continuous development of deep learning,it has become a trend to use deep learning-related theories and technologies to automatically identify whether a carrying vehicle is empty or heavy.Based on the image characteristics and scene characteristics of the actual mineral resource carrier vehicles,this paper proposes a method for identifying carrier vehicles based on convolutional neural networks,and implements a carrier vehicle identification system based on this method.The main work of the paper is as follows:(1)Image acquisition and preprocessing of the vehicle.In this paper,after the two mining companies in Dafang County,Guizhou Province collected the pictures of the vehicles on the spot,the images were preprocessed by changing the brightness,contrast,and graying the image,which not only expanded the data volume of the original image.Moreover,the robustness of the model is guaranteed.(2)A vehicle identification model based on convolutional neural network is designed.Based on the Alex Net model,this paper uses the Tensor Flow framework to build a convolutional neural network model.According to the characteristics of the vehicle recognition scene,a convolutional neural network model that is fused with the attention mechanism is designed.Experimental results show that the model has higher accuracy and robustness than traditional convolutional neural network models.(3)The vehicle identification system based on convolutional neural network is realized.This paper designs a complete set of vehicle identification system,which combines vehicle identification model with tax monitoring system to realize a series of processes from snapshot to output detection.In addition,the system is highly robust and can successfully identify a large number of vehicle images from different angles and scenes.