Traffic contraband recognition method based on self-attenuation weight and multiple local constraints

Abstract

The invention discloses a traffic contraband recognition method based on self-attenuation weight and multiple local constraints. The method comprises the following steps: preparing original data; datapreprocessing. Data set creation. A structure of a classical convolutional neural network LeNet-5 network is used as a structure of a contraband classification network; for different layers of convolution features output by the LeNet-5 network, transverse connection is carried out according to the direction from a deep layer to a shallow layer in combination with the self-attenuation weight to construct fusion features; a self-attenuation weight coefficient is adjusted in a self-adaptive manner during model training; and a final category is calculated by designing a local multi-constraint strategy for the traffic contraband. According to the invention, automatic recognition and classification of traffic prohibited articles are realized based on deep learning and the convolutional neural network; and automatically learning high separability characteristics of the traffic forbidden objects by extracting a design characteristic learning network, training a traffic forbidden object automatic classification model, and completing the traffic forbidden object automatic recognition method for the common RGB image.

Description

technical field [0001] The present invention relates to the field of automatic classification of a deep convolutional neural network, and more specifically relates to a method for automatic identification of traffic contraband items for ordinary RGB images. Background technique [0002] As people's awareness of safety protection continues to increase, the control of flammable and explosive hazardous materials has been strengthened year by year. For vehicles with relatively dense population and relatively small space for activities (such as trains, cars, airplanes, etc.), relevant laws and regulations clearly stipulate the types of items that are prohibited to be carried when riding. For these contraband items, the current main inspection method is through the X-ray security inspection machine. For the items carried in the luggage, when passing through the X-ray security inspection machine, the items in the luggage will show a rough shape on the monitor, and different colors...

AI Technical Summary

Problems solved by technology

Relatively high human cost input

Images