A traffic sign detection method based on center-boundary connection model

Abstract

The invention discloses a traffic sign detection method based on a center-boundary connection model. The steps are: use a dynamic programming method to match the connections in the training samples, and find out the corresponding relationship of each connection of the polygons in all the training samples; assuming that each The length of the connection in all training samples conforms to the Gaussian distribution, the slope angle of the connection and the gradient direction of the boundary point conform to the Von Mises distribution, and the parameter distribution of the connection is estimated by the expected maximum method. The parameter distribution of each connection in the polygon constitutes the center of the polygon ‑Boundary connection model; the potential center and boundary points of the polygon are obtained by using the rotationally symmetric projection method; with the projection center as the potential center of the polygon, all the corresponding boundary points connected in the polygon center‑boundary connection model form the detected polygon. This detection method can realize the rapid detection of traffic signs.

Description

technical field [0001] The invention belongs to the technical field of image processing and computer vision, and in particular relates to a traffic sign detection method. Background technique [0002] Traffic sign detection is an important research direction in computer vision. As an auxiliary facility for road safety, traffic signs play an important role in regulating traffic behavior, indicating road conditions, ensuring road efficacy, guiding pedestrians, and driving safely. As a key technology of intelligent transportation systems, traffic sign detection is an integral part of systems such as autonomous driving, assisted driving, and traffic sign maintenance. [0003] Traffic sign detection is to use the color or shape features of traffic signs to detect and segment the traffic sign area in the image. Whether the detection is accurate or not directly affects the accuracy of subsequent traffic sign classification. [0004] Traffic sign detection methods can be roughly d...

AI Technical Summary

Problems solved by technology

Compared with the projection method, this method has a higher detection rate and a lower false alarm rate. However, this method has higher requirements on the edge quality, and the detection effect decreases rapidly when the edge is incomplete.

[0008] Based on the above analysis, there is currently no method that can effectively detect traffic signs. It is under this research background that the inventor developed this case

Images