Driver behavior recognition method based on deep hybrid encoding and decoding neural network

Abstract

The invention provides a driver behavior recognition method based on a deep hybrid encoding and decoding neural network. The method comprises the steps: constructing a driver behavior recognition dataset; constructing a coding and decoding space-time convolution network; constructing a convolutional long short-term memory network; constructing a classification network; training three networks inthe driver behavior recognition model; adopting the trained driver behavior recognition model to recognize videos in the data set; and sending the video sample into a trained encoding and decoding space-time convolutional network to obtain a short-term behavior space-time feature representation, sending the short-term behavior space-time feature representation into a trained convolutional long-term and short-term memory network to obtain a long-term behavior space-time feature representation, and outputting a final driver behavior classification result by a trained classification network. According to the method, the implicit motion information can be effectively extracted from the short-term video clip, the driver behavior characteristic coding in the long video is realized through space-time fusion, the identification precision is high, and the driver behavior identification in the monitoring video can be realized.

Description

technical field [0001] The invention belongs to the field of image processing and pattern recognition, and relates to a driver behavior recognition method based on a deep hybrid codec neural network. Background technique [0002] Driver behavior recognition aims to distinguish between normal driving behavior and some dangerous driving behaviors, such as driving with hands off the steering wheel, driving on the phone, smoking while driving, etc. Dangerous driving behavior seriously affects the driver's attention, and has always been the main factor causing traffic accidents. According to a survey by the Ministry of Transport of China, more than 63,000 people died in traffic accidents in China in 2018, and more than 80% of the accidents were related to the dangerous driving behavior of drivers. Therefore, driver behavior monitoring technology has important research significance for road safety and intelligent transportation. [0003] The automatic driver behavior recognition...

AI Technical Summary

Problems solved by technology

However, the recognition accuracy of this type of algorithm is not high at present, and there are mainly the following difficulties:

[0004] (1) The movement trend of the driver's behavior is relatively slow, and the similarity of the global information of different behavior categories is high, so the motion information is not significant

Artificially designed motion features are difficult to characterize for this problem

[0005] (2) Dense optical flow field can be used to model driver behavior, but optical flow prediction is time-consuming, so the real-time performance of this method is low

Short-term spatio-temporal characteristics are difficult to determine the current driving behavior representation, which can easily lead to high false alarm rate

Images