Study On Detection & Recognition Of Traffic Speed-Limit Signs Base On Generalized Structure Elements

Intelligent Transportation System(ITS), which is regarded as the research object of transportation, service control and manufacturing, aims to solve the problem of traffic congestion, vehicle safety improving and vehicle intelligence enhance. Meanwhile,Traffic Sign Recognition(TSR) is one of the significant gradients of the research of service control, which contains many technical means, such as electronic information,video processing, digital image processing, artificial intelligence and pattern recognition and so on.Currently, many problems are commonly appear in the aspect of the detection and recognition methods for the traffic speed limit signs, for examples, the non-ideal recognition accuracy rate, the sluggish recognition speed, difficulty to recognize the damaged traffic speed limit speed sign and its like. Thus, a novel template matching recognition algorithm and a generalized structuring element is proposed in this paper aims to research the identification method for the traffic speed limit signs, which is used to solve the problems mentioned above.The main research contents can be expressed as below:(1) Image preprocessing. The advantages of HIS space are expressed after comparing the advantages and shortcomings of RGB and HSI, and then the HSI space is selected to process the traffic speed limit signs. Meanwhile, the processing of histogram enhancement, Wiener filtering and geometric transformation and carried out in this space.(2)Generalized structure elements filtering. The theory of generalized ideological structuring element is introduced to the detection and recognition of traffic speed limit sign. Three kinds of generalized structure elements are built for the three different speed limit sign, so the extended generalized structure elements can realize the aim. According to the generalized structural elements which are included in the generalized structure elements, the detection for the images is realized. The images are to be recognized and the generalized structure elements include red circular pattern, blue circular pattern and black circular pattern.(3) Figure extraction. Based on the interesting region, the key figures of the traffic speed limit signs are extracted and normalized, according to the figure extraction technology.(4) Sign recognition. After the normalized figures are got, according to the three different traffic speed limit signs, the corresponding template has 16 feature component were identified. The template’s matching number is 10 and every single feature has a threshold.(5) Experimental Analysis. In order to realize the verification of the algorithm the tool of C# language is used to develop software. Experiments show that the algorithm for recognizing the traffic speed limit signs which are damaged has a high right rate. For the general speed limit sign, the right rate is 91.67%, and for the damaged speed limit signs, the right rate is 89.83%.The recognition accuracy rate for the above situations are90.80%.The main innovation of this paper as follows:(1)The method for detecting and recognizing the traffic speed limit sign according to the structuring element is proposed.(2)The algorithm for detecting and recognizing the damaged traffic speed limit sign is proposed.