Real-Time License Plate Positioning Algorithm Optimization and Hardware Implementation

Aiming at the problems of limited application environment and high consumption of hardware resources of the number plate positioning system implemented by the traditional edge detection algorithm, an improved Canny edge detection algorithm is proposed in this paper. The algorithm uses Cordic IP cores to optimise the Sobel operator to compute gradient magnitude and direction, replacing complex operations and reducing hardware costs. The introduction of the Otsu algorithm allows the system to select appropriate thresholds for different environments. The improved and optimised algorithm is implemented on FPGA and simulated to locate 2260 number plates in different environments with a positioning accuracy of 99.58%, an average time consumption of 2.43 ms and a hardware resource consumption of less than 67%. The experimental results show that the system can be applied to a variety of complex environments, providing a feasible technical solution for automatic licence plate recognition system.