A road sign recognition method and system

Abstract

The invention discloses a road sign recognition method and system. The method includes the following steps: S1, performing color threshold binarization on a single frame image in a road image in HSV space, and obtaining sign connected domains according to sign color feature positioning; S2 1. Calculate the aspect ratio of the connected domain of the signage, and exclude false targets; calculate the distance from the edge of the outline in the connected domain of the signage to the center of the target, and obtain the target positioning area in which the outline of the connected domain of the signage is a circle; S3, intercept the corresponding grayscale image and Calculate its average gray scale, and use the average gray scale value as the threshold value to carry out binarization operation on the grayscale image of the target positioning area to obtain the binary image of each target positioning area; S4, extract the connectivity in the binary image of the target positioning area Domains, and sort the connected domains according to the coordinates S5, calculate and determine the numbers or characters inside the target positioning area through the neural network pattern recognition algorithm; S6, calculate the confidence. The invention has high efficiency and high accuracy for identifying the signboard.

Description

technical field [0001] The invention relates to the fields of machine vision and artificial intelligence, in particular to a road sign recognition method and system. Background technique [0002] Video image road sign recognition is to judge the restriction information of signs on driving. The traditional method is: first locate the position of the sign with the image processing method, and then segment each number and character of the sign. Finally, template matching or neural network algorithm is used for identification. [0003] The traditional method can effectively determine the speed limit signs, but cannot distinguish the width limit and height limit signs. The main reason is that traditional methods only focus on the recognition and judgment of numbers and characters in road signs, while ignoring the characteristics of the signs themselves. [0004] The traditional template matching method needs to scale the target size to be equal to the template size, then compa...

AI Technical Summary

Problems solved by technology

[0006] The technical problem to be solved by the present invention is to provide a road sign recognition method and system for the defects that the template matching method in the prior art is inefficient, easily affects the result, and the commonly used neural network algorithm is easily affected by noise.

Images