Technology Research, Car License Plate Recognition Based On Intelligent Algorithms

As information technology and intelligent technologies, traffic management system of information, intelligence has become a major trend of development. License plate recognition system is the core of intelligent transportation system components in the modern transportation charge management system has played a pivotal role. In recent years, license plate recognition system, key technology research has become a hot area of intelligent transportation problem. In this paper, license plate recognition system in the image pre-processing, vehicle license location, character segmentation, character recognition and other key parts of several of the more in-depth, comprehensive study, and license plate character segmentation techniques to do a new attempt. The effect of segmentation of the characters directly affect the performance of subsequent image processing.In this paper, the technology license plate character segmentation using improved ant colony algorithm. Image segmentation is the image processing and image analysis of crucial importance. The basic ant colony algorithm search is then, the computation, it is not conducive to convergence of the algorithm. Therefore, this article proposes a two-dimensional histogram and improved ant colony clustering algorithm for image segmentation method. Improve the application of traditional image segmentation method segmentation ineffective problem. Ant colony clustering algorithm is a kind of discrete, parallel, robustness and ability of evolutionary fuzzy clustering method. This algorithm is based on this tradition, by two-dimensional histogram to set the initial cluster centers to reduce the number of cycles of ant colony algorithm to define a new boot function, and by improving the ant colony pheromone update mechanism to improve the speed of clustering. Experiments show that the algorithm is a more accurate and rapid method of image segmentation.As the artificial neural networks and the human brain has a high degree of similarity, based on neural network license plate recognition technology research house, good fault-tolerance and associative memory function, adaptive, self-learning ability and fault-tolerant characteristics, especially improved network on behalf of the neural network has a good self-learning ability, strong classification capabilities, fault tolerance, or robustness can be achieved non-linear mapping from input to output. Therefore, choosing neural networks as classification of characters to identify the main method. Application of...