Signal Recognition Of Video-based Traffic Control Indicating Lights

Video based traffic control lights indicating signal recognition research is aimed atcontaining a video traffic lights command for processing，To test the video scene of thetraffic lights, description, identification and classification,and to understand and describefurther. Traffic light color discrimination mainly through to the RGB values indetermining its area, and then converted to the corresponding value of HSV space. Itsgoal is to describe the color of the traffic lights, and on this basis,it can help visuallydisabled people to distinguish the traffic lights, so that they can drive safely.This article mainly aimed at the general traffic light for testing, not including speciallights. This paper adopts the color space conversion and many,basic methods such asstatistics. based on these methods we can do some work for the identification andclassification of traffic light colors. In addition, color space conversion get a goodapplication in prediction of traffic light colours.Main work is as follows:1,learning some basic knowledge about digital image and video format, especially theBMP image format, and other basic knowledge, such as color space. BMP image fileincludes three parts: the bitmap file header data structure, it includes the type of BMPimage file,display information and so on; Bitmap information data structure contains aBMP image width, height, the compression method, definition of the information and soon.2, learning VC and OpenCV and grasping the relationship between them;3, learning automatic identification of hand tag and Automatic identification ofautomatic detection,including some other models.4, the use of a variety of color space conversion algorithm and other algorithms torealize the recognition of traffic lights.