Automatic traffic off-site zebra crossing area detection method based on AI technology

Abstract

The invention discloses an automatic traffic off-site zebra crossing detection method based on an AI technology, belonging to the field of image recognition. According to the method, a Darknet-53 network is used as a skeleton network to construct a zebra crossing recognition model; the model inputs a picture containing zebra crossings; the Darknet-53 network extracts feature maps of three sizes from the input picture; and multi-scale target detection is carried out through nine anchor frames of different sizes, bounding box data of each zebra crossing in the picture is output, the final bounding box data comprises a bounding box center point coordinate, a bounding box width, a bounding box height, a first slope, a second slope, a target category and confidence, and a bounding box can be converted into the corresponding zebra crossing through the slopes. According to the method, zebra crossings can be effectively and accurately recognized even in a complex and changeable scene, and a recognition speed is greatly higher than the recognition speed of manual recognition.

Description

technical field [0001] The invention belongs to the field of image recognition, and in particular relates to an off-site traffic automatic detection method for zebra crossings based on AI technology. Background technique [0002] Zebra crossing detection technology has been studied for a long time, but most of the existing technologies are based on traditional image processing methods, such as edge detection through Canny algorithm, and then use Hough transform to extract straight lines, and finally identify zebra crossings based on the extracted straight lines. However, this method has high requirements on images, and lighting, pedestrians, vehicles, and bad weather will all affect it. It has strong limitations and cannot cope with complex traffic scenes. In recent years, with the improvement of traffic facilities and illegal capture systems, the occurrence of traffic accidents has been reduced to a certain extent. However, its illegal data is manually screened to classify...

AI Technical Summary

Problems solved by technology

Because the actual traffic scene is complex and changeable, traditional image methods cannot be effectively processed, so there is an urgent need for a detection method that can effectively and accurately identify zebra crossings even in complex and changeable scenes

Images