Traffic sign plate recognition method and device

Abstract

The invention discloses a method and a device for detecting and identifying traffic sign boards by using MCT (Modified Census Transform, modified statistical transformation) features. A method for identifying traffic sign boards according to one aspect of the present invention includes: using the MCT (Modified Census Transform) feature in the input image to detect the stage of the traffic sign board candidate area; using the MCT feature histogram of the candidate area to verify Whether described candidate area corresponds to the stage of traffic sign board; And utilize the MCT feature histogram to verified candidate area (hereinafter referred to as area of ​​interest), classify described area of ​​interest into the traffic sign board of corresponding traffic sign board step by step identification stage.

Description

technical field [0001] The invention relates to a method and a device for identifying a traffic sign board, more specifically, a method for detecting and identifying a traffic sign board and a device thereof by using MCT (Modified Census Transform) features. Background technique [0002] As mobility by vehicles becomes popular, traffic sign boards are installed on roads to provide drivers of vehicles with information necessary for driving smoothly and safely. [0003] Generally, traffic sign boards can provide traffic information sign boards, mileage sign boards, weather information status boards, air pollution status boards, traffic accident status boards, municipal information boards and advertisement boards, etc. according to the type. [0004] Recently, as the number of drivers with poor eyesight or elderly drivers has increased, the need to develop a driving safety support system that can accurately convey information on traffic sign boards has also increased. [0005]...

AI Technical Summary

Problems solved by technology

Here, the preprocessing process needs to increase the calculation, thereby reducing the detection and recognition speed, and the use of color detection technology will cause the detection performance to decrease when the lighting changes and the applied paint changes color.

In addition, the detection technology using morphology requires a large amount of calculation, and when it is difficult to detect in real time or when the calculation amount is small, the detection accuracy will decrease.

Images