Traffic Signs Detection And Recognition In Complex Dynamic Condition And Its Application

The road traffic sign recognition system is very important in ITS. It will promote the level of intelligentialize and automation for traffic management, which is a significant content of ITS research. Road traffic sign recognition system mainly includes two parts, namely detection and recognition of traffic sign. Among them, the traffic sign detection is the key issue that needs to be addressed, and it is the premise to realize the recognition of traffic signs. After studying classic traffic sign detection algorithms, this paper proposes an improved algorithm and combines SIFT algorithm to realize the recognition of traffic signs. The main works of this paper are presented as following:The existing algorithms are relatively unitary for detection object, and mostly only detect several signs in a certain class. With the increasing of traffic signs, the correct rate of detection is decreased obviously. In addition, these methods consider the color and shape information, but they ignore the deterministic relation between shape and geometric with traffic signs. In this paper, a fast and efficient algorithm is proposed. According to the prior information of the shape and color of traffic signs, the algorithm proposes a decision tree of traffic signs, which can filter out the regions of interest and divide traffic signs into 10 sub classes. This paper adopts the unified method to detect all kinds of signs and classify the results of detection. Experiments show that the algorithm can reduce the complexity, and improve the real-time performance and efficiency of the TSR system.The traffic signs always suffer distortion, invisibility or slope because of the complexity of road conditions, the changing illumination conditions and the vibrating of car. These problems will bring great difficulties to traffic signs detection. This paper analyzes the influence of various complex environments on traffic sign detection, and propose the corresponding solutions to complete or adjust sings. Experiments show that this method can gain a good detecting performance of traffic signs in complex scenes, which makes the algorithm more robust.Combining with SIFT algorithm, this paper implements the detection and recognition algorithms in PC, Android, IOS and DSP platform respectively. It can recognize traffic signs as an on-board equipment in reality environment. Experiments show that the algorithm has a good practicability and portability.