An abnormal detection method for drivers based on computer vision

Abstract

A computer vision-based driver abnormality detection device and detection method, wherein the detection device includes a vehicle-mounted camera and a video recording, storage and return device, and the vehicle-mounted camera is connected to the video record, storage and return device, and is characterized in that it is also included in the detection center The detection server and data storage server are connected through the network, and the detection server is connected with the video record storage and return equipment through the network; there are also central video image processing and target detection equipment in the detection center, and the central video image processing It is connected with the target detection device and the detection server, and the detection server is connected with the video record storage return equipment through the network. The invention solves the problems of large manual analysis workload and slow response in the existing monitoring means, real-time automatic detection and monitoring of the driving conditions of drivers, passengers and vehicles, saves a lot of manpower, is accurate and effective, and achieves the purpose of detection.

Description

technical field [0001] The invention relates to a computer vision-based driver's driving abnormality detection device and a detection method, which relate to the fields of computer vision and machine learning, and is a computer vision-based driver's driving situation detection method. Background technique [0002] The driving situation of drivers is directly related to the occurrence of vehicle traffic accidents. Good driving behavior can reduce the occurrence of traffic accidents and casualties during vehicle driving; while bad driving behaviors such as fatigue driving, drunk driving, answering phone calls and replying text messages while driving Or Internet access, hands off the steering wheel, etc. will cause the driver to be distracted and even the vehicle to drive abnormally, which will be more likely to cause traffic accidents. [0003] For a carrier vehicle, the driver's dangerous driving behavior not only easily leads to road traffic accidents, but also causes accide...

AI Technical Summary

Problems solved by technology

[0004] The current driving behavior of the driver is usually to detect the driving status of the vehicle through external sensors. It is necessary to install corresponding sensors on the outside of the vehicle or on the transmission system, which is costly and requires more additional acquisition and processing equipment; In the case of abnormal driving, the analysis of video surveillance data is mostly manual monitoring and spot checks, with a large analysis workload, small coverage, and poor results, and manual review cannot monitor the scene in real time, making it difficult to monitor all abnormal driving situations and make decisions. Reminders and Quick Responses

Images