Optimization Of Target Detection Algorithms In Complex Detection Environment And Its Implementation In FPGA

Target detection algorithm in complex detection environment is still facing severe challenges.The main problems it faces are the complexity of image processing algorithm,processing speed and detection effect.Target detection algorithm based on FPGA vision processing platform and CORDIC is the key to solve the above problems.In this paper,the maximum approximation angle CORDIC algorithm based on fixed rotation sequence and the Sobel edge detection algorithm based on CORDIC are studied.On this basis,a target detection platform based on FPGA is designed and implemented.To overcome the shortcomings of CORDIC algorithm such as more iterations,high complexity and large storage space,a CORDIC algorithm with fixed rotation sequence is proposed.The algorithm rotates every bit of high binary value twice,which not only reduces the process of judging rotation direction,but also simplifies the calculation of proportional constant K.Therefore,compared with the original algorithm,it reduces the complexity and storage space greatly.The advantages and disadvantages of traditional Sobel target edge detection algorithm are analyzed.Three aspects are optimized based on CORDIC algorithm.The first aspect is to use fast median filtering algorithm to improve the anti-noise ability of the algorithm.The second aspect is to propose an improved adaptive threshold calculation module.Finally,the optimized CORDIC algorithm is used to realize the fast calculation of Sobel edge gradient,and the optimization and implementation of the algorithm are described.Methods The simulation results show that the optimized Sobel algorithm outperforms the traditional algorithm in edge detection under noise and different illumination environments.In view of the above theory,a target detection platform based on FPGA is designed and implemented by taking the driving monitoring unit in the driving visual aided system as an example.This paper introduces the requirement analysis,scheme selection,hardware and software implementation of the platform and test verification of the platform.The results show that the optimized target edge detection algorithm has good detection effect and maintains good real-time performance.