A Traffic Sign Detection Method Based on Adaptive Threshold

Abstract

The invention discloses a traffic sign detection method based on an adaptive threshold. The present invention first adopts the red and blue method to preprocess the image to obtain the red and blue gray scale image, and then performs multiple threshold value processing on the gray scale image. In each thresholding process, contour detection is performed on the binary image generated by the thresholding process; then, according to the shape characteristics of the traffic sign, geometric constraints and sign contour matching of Hu invariant moments are performed, and the current threshold image is obtained by filtering and screening. The set of suspected traffic sign areas; finally, the multiple thresholding results are combined, and the final traffic sign ROI is determined according to the frequency of detection of the outline area of ​​the sign. The present invention enables the traffic sign candidate area in the image to obtain a better threshold processing effect, has the characteristics of threshold self-adaptation, not only has high detection efficiency and time performance, but also well solves the adaptability problem under different lighting conditions .

Description

technical field [0001] The invention belongs to the technical field of traffic sign recognition, and relates to an efficient traffic sign detection method in traffic sign recognition of vehicle-mounted video images, in particular to a traffic sign detection method based on an adaptive threshold. Background technique [0002] With the rapid development of transportation, traffic safety is becoming more and more important. Subsystems such as pedestrian detection and traffic sign recognition in Advanced Driver Assistance Systems are important methods to reduce vehicle traffic accidents, and traffic sign detection is a key part of the traffic sign recognition system, which affects traffic sign recognition. The system recognizes the important factors of efficiency and real-time performance. [0003] The traffic sign recognition system can be divided into two stages: detection and classification recognition. According to the different characteristics of traffic signs used, the de...

AI Technical Summary

Problems solved by technology

Most of the detection methods based on color features are based on fixed thresholds in some color spaces. Although these methods have good real-time performance, they are easily affected by lighting conditions, and the detection efficiency is very low, which cannot meet the detection rate requirements of the traffic sign recognition system.

The detection method based on shape features has better detection performance for some specific types of traffic signs, but it is not suitable for all types of traffic sign detection and lacks good adaptability; at the same time, because such methods do not consider color features, Will generate more interference, affecting the performance of the detection algorithm

Images