Driver action recognition method and device based on three-dimensional convolutional neural network

Abstract

The embodiment of the invention provides a driver action recognition method and device based on a three-dimensional convolutional neural network. Feature data is extracted from a video photographed bya driver through preset feature engineering, and a target model recognizes the behavior of the driver according to the feature data , wherein the target model is obtained by training a constructed three-dimensional convolutional neural network, the three-dimensional convolutional neural network comprises a plurality of combined layer structures connected in sequence, and each combined layer structure comprises a convolutional layer and a pooling layer. By improving the structure of the three-dimensional convolutional neural network, the trained target model has a more accurate recognition result on the action of the driver. And on the other hand, compared with facial feature acquisition, driver action acquisition is not easily disturbed by the environment, the feature data contains optical flow features reflecting time-dependent changes of driver actions, and the accuracy of an identification result is further improved through the action continuity data.

Description

technical field [0001] The present invention relates to the technical field of rail transit, in particular to a driver action recognition method and device based on a three-dimensional convolutional neural network. Background technique [0002] In ensuring the driving safety of urban rail transit, drivers shoulder important responsibilities, and their accurate actions and clear awareness often determine the safety of passenger transportation. Fewer driver configurations, monotonous driving actions, and high automation of train driving are important reasons for driver fatigue. At the same time, the driver's personal living habits, workload, and working hours will also affect whether the driver is fatigued. Some traditional methods are to alleviate the fatigue of train drivers by improving the management system and work plan. The train "anti-sleep" equipment also reduces the driver's fatigue to a certain extent. Sexual movements are not sensitive to the "anti-sleeping death" ...

AI Technical Summary

Problems solved by technology

[0005] The embodiment of the present invention provides a driver action recognition method and device based on a three-dimensional convolutional neural network, which is used to solve the problem that the methods in the prior art mostly judge based on the images collected from the driver's face, and are easily blocked or affected by the environment. Influence, the problem of inaccurate judgment results

Images