New Traffic Lights Identification Algorithms Based on Image Recognition
The traffic lights identification algorithms is an important part of the platform for automatic driving technology and it is also applied in the remote control system. As the absolute standard of behavior instruction, lights are the most firsthand object to judge. Image recognition technology has been applied in the Gray Spaces and HSI space, but Comparing with RGB color space, the latter is easier to understand and realize. This paper puts forward a kind of identification algorithm for single or multiple lights. According to the target image numerical characteristics in the RGB color space and judging from the target area needed and the related parameters extracted, The feasibility of this algorithm is proved in theory and application and the experimental results are satisfactory in the Matlab environment.
Paul P. Lin and Chunliang Zhang
Y. L. Shao et al., "New Traffic Lights Identification Algorithms Based on Image Recognition", Applied Mechanics and Materials, Vols. 105-107, pp. 1856-1859, 2012