Traffic jam degree multi-dimensional analysis method based on deep learning

Abstract

The invention discloses a traffic jam degree multi-dimensional analysis method based on deep learning. The method comprises the following steps: 1) setting a camera preset position and calibrating a camera; 2) initializing a convolutional neural network model; 3) acquiring a real-time video stream; 4) checking the working state of the camera; 5) performing vehicle target detection on the region ofinterest by using a convolutional neural network model; 6) tracking a vehicle target; ;7) collecting traffic jam;8) initially detecting traffic jam and making jam prediction; 9) analyzing traffic jamdegrees; 10) reporting and traffic jam event and setting sleep state. According to the invention, multi-dimensional traffic parameters such as the road accommodation rate and the average speed are adopted to carry out accurate and rapid traffic congestion degree analysis, the robustness is good, and the detection accuracy is high.

Description

technical field [0001] The invention relates to the technical field of deep learning, in particular to a multi-dimensional analysis method of traffic congestion degree based on deep learning. Background technique [0002] In recent years, with the continuous advancement of urbanization development, the number of cars has continued to increase, which brings convenience to traffic, but also makes traffic congestion, traffic accidents and other problems increasingly prominent. Among them, the problem of traffic congestion will not only affect people's travel efficiency, but even lead to serious traffic accidents and endanger people's travel safety. Therefore, it is particularly important to accurately detect and analyze the degree of traffic congestion in real time. [0003] The most direct way to detect traffic congestion is to manually observe the road congestion on the surveillance video, but it is impossible to observe the situation of all road sections for 24 hours at the...

AI Technical Summary

Problems solved by technology

[0005] In order to overcome the shortcomings of low detection accuracy, slow detection speed, and robustness in the prior art, the present invention provides a multi-dimensional analysis method of traffic congestion based on deep learning, and uses deep convolutional neural network (CNN) features for vehicle target detection , and use multi-dimensional traffic parameters such as road capacity and average speed to conduct accurate and rapid congestion analysis

Images