Detection and early warning method and system at traffic intersection and curve

Abstract

The invention provides a detection and early warning method and system at a traffic intersection and a curve. The method comprises the following steps: continuously collecting original images at a first preset distance from a target intersection or a target curve through camera equipment; carrying out defogging processing on the original image through a defogging algorithm based on a conditional generative adversarial network to obtain a fogless image; detecting vehicles and pedestrians in the fog-free image through an image detection network; and when the target object is detected, displaying the target object and the distance between the target object and the intersection or the curve through warning equipment. According to the invention, complex road sections such as curves and intersections can be monitored in real time, accurate and timely road condition information is provided for drivers, the blank of rural curve and intersection early warning schemes is filled up, the method is suitable for complex rural environments, and the occurrence rate of rural traffic accidents is greatly reduced.

Description

technical field [0001] The invention relates to the technical field of image detection, in particular to a detection and early warning method and system located at traffic intersections and curves. Background technique [0002] In my country's rural traffic, a large number of traffic accidents occur every year, causing many people to be injured, disabled or even killed. There are many reasons for traffic accidents in rural traffic. One of the important reasons is that the road conditions in rural areas are complex, and drivers cannot obtain traffic status information in a timely manner, especially on road sections where it is difficult to see the road conditions such as curves or intersections, which leads to Many traffic accidents. Therefore, advanced camera-based detection technology is introduced to monitor complex road sections such as curves and intersections in real time, extract and display road condition information, provide accurate and real-time road condition inf...

AI Technical Summary

Problems solved by technology

However, some of the current detection technologies are not well suited for rural traffic. Since rural roads are mostly close to geographical environments such as mountains and farmland, the wind and sand are relatively large, and the fog is also heavy. Most of the snapshots used for detection are not clear.

[0003] The application scenarios of the detection algorithm in the existing technology are all scenes with good environment and clear captured pictures, but they are not very friendly to the further complex and unfavorable environment.

It has not been used in the rural environment, part of the reason is that it is difficult to achieve better working results because of the complexity of the rural environment

The current popular deep learning-based target detection technology is based on the photos taken by the camera, so the quality of the input image is high, and if the input photo is noisy, it will have a great impact on the detection results.

Images