Driver behavior recognition method based on lightweight convolutional neural network

Abstract

The invention discloses a driver behavior recognition method based on a lightweight convolutional neural network, and the method comprises the steps: obtaining a driving behavior public data set, andobtaining pictures of different driving behavior classifications; preprocessing the pictures, randomly disturbing a data set, and dividing the data set into a training set and a test set; performing data enhancement on the training set; designing a lightweight convolutional neural network, and inputting the training set into a data input network for feature extraction; performing probability prediction of each driver behavior category on the extracted feature vector by using a classifier, calculating a loss function according to the predicted probability of the training set category label pair, and guiding the next training direction of the network through back propagation; and storing the trained driver behavior classification model after the training is completed. According to the method, the lightweight convolutional neural network is designed by adopting information circulation between the convolution module and the enhanced channel, the driver behavior classification model which is small in size, simple in operation and high in accuracy is trained, and the method is suitable for vehicle-mounted mobile terminals to recognize and classify the driving behaviors.

Description

technical field [0001] The invention relates to computer vision and deep learning technology, in particular to a driver behavior recognition method based on a lightweight convolutional neural network. Background technique [0002] Driver distracted driving has become the main cause of traffic accidents, and the driving behaviors exhibited by driver distracted driving include using mobile phones while driving (including calling, sending messages, browsing websites, playing games, etc.), eating, and Behaviors such as passenger interaction. Nowadays, in-vehicle assisted driving systems have been more and more widely used in various vehicles. Efficient and accurate identification of the driver's driving behavior to prevent traffic accidents caused by driver distraction has become one of the assisted driving systems. extremely important function. For vehicle-mounted assisted driving systems, how to use the existing driver behavior data sets to learn the characteristics of vario...

AI Technical Summary

Problems solved by technology

In the selection of convolutional neural network, ResNet, InceptionV3 and other networks are characterized by high precision but complex structure and large model size; MobileNetV2, ShuffleNet and other networks can reduce the amount of calculation, but bring the loss of precision

[0004] It can be seen that using the existing convolutional neural network to train the driver behavior data set, it is difficult to simplify the network and obtain a small model under the condition of ensuring accuracy, so the application on the vehicle mobile platform is limited

Images