Algorithm And Implementation Of Infrared Image Enhancement For Human Visual System

With the development of infrared imaging technology,real-time infrared image enhancement is drawing more and more attention in the world.Compared with visible light imaging technology,infrared imaging has strong abilities like environmental suitability,penetrability and strong concealment.At the same time the resolution and frame rate of infrared imaging system is also rising,high-end infrared camera can be in the resolution of 640×512 and its frame rate can reach more than 200 Hz.However,the processing performance of existing systems can not meet the requirement of real-time enhancement with high-resolution,high frame rate infrared images.In this paper,a real-time infrared image enhancement system based on FPGA is designed to enhance the infrared image with high resolution and high frame rate.The main research contents include an infrared image enhancement algorithm based on human vision,hardware system design,hardware real-time implementation and optimization.At first,this paper introduces the existing infrared image enhancement algorithms and their hardware implementation.Aiming at the current situation of infrared image hardware implementation at home and abroad,an infrared image enhancement algorithm which is suitable for hardware implementation is proposed.This algorithm can significantly improve the contrast and enhance the detail of infrared images,then avoid the defects such as gray scale overtaking and ring effect in traditional algorithms.Meanwhile,the algorithm has high parallelism and pipeline operation.It’s suitable for real-time implementation of FPGA.To implement the algorithm above,this paper designs a kind of infrared image real-time processing system,which uses FPGA as the processor.Its hardware architecture adopts hierarchical structure to separate the core processing board from the video input and output interface board which improves the generality of the system.Then,the infrared image enhancement algorithm proposed in this paper is realized.In the process of transplanting algorithm,this paper presents a low-delay and large-scale bilateral filter implementation method which has low hardware resources usage.And its running clock frequency can reach more than 200 MHz.Secondly,A simple real-time 64-bit logarithmic operation method without look-up table is put forward to obtain the result in 1 clock cycle.Finally,this paper optimizes the realization of traditional histogram equalization which reducing the logical resource usage by 50%.Then it realizes the logarithmic histogram equalization which can handle the maximum resolution of 65535×65535 image,while its operating clock frequency can reach more than 100 MHz.Based on the improvement above,the real-time hardware implementation of infrared image enhancement algorithm based on human vision is proposed.The system can process multi-format infrared images,including 1280×1024@60Hz and 640×512@234Hz,in real time,to meet the requirements of the current mainstream infrared imaging system.In this case,the allocation of FPGA resources is reasonable,it is convenient to implement this algorithm on other platforms,and improve the versatility of the algorithm hardware.