基于ZYNQ的CLAHE图像增强算法实时加速设计  

Real-time Accelerated Design of CLAHE Image Enhancement Algorithm Based on ZYNQ

作　　者：李晓琪 王云峰 吴倩楠 洪应平[1] Li Xiaoqi;Wang Yunfeng;Wu Qiannan;Hong Yingping(Key Laboratory of Instrumentation Science and Dynamic Measurement,Ministry of Education,North University of China,Taiyuan 030051,China)

机构地区：[1]中北大学仪器科学与动态测试教育部重点实验室,太原030051

出　　处：《单片机与嵌入式系统应用》2023年第11期49-53,共5页Microcontrollers & Embedded Systems

基　　金：国家自然科学基金青年科学基金项目(51705475)。

摘　　要：在低照度环境下,CMOS传感器采集的图像整体效果偏暗,用传统的图像处理方式进行增强速度较慢,满足不了实时性要求,为此,本文设计了基于ZYNQ的CLAHE图像增强算法。CLAHE增强算法主要通过对分块后的直方图进行限幅操作,再利用双线性插值算法消除块状效应来对较暗的图像进行处理,能够提高图像的整体质量并保留局部的细节,防止在像素增强的过程中将噪声放大。在HLS设计的算法中循环使用pipeline流水线操作指令以提高计算的并发性和吞吐率。最后,将综合生成的IP核固化到ZYNQ的PL端,能够在几乎没有延迟的情况下处理多组图像并在HDMI显示器上显示,验证了图像算法增强的实时性。In low illumination environment,the overall effect of the image collected by CMOS sensor is dark,and the processing speed of image enhancement by traditional image processing methods such as image acquisition board or computer is slow,which can not meet the requirements of real-time.So a CLAHE image enhancement algorithm based on ZYNQ is proposed.CLAHE enhancement algorithm mainly carries out amplitude limiting operation on the histogram after segmentation,and uses bilinear interpolation algorithm to eliminate block effect to process the dark image,which can improve the overall quality of the image and retain local details to prevent noise amplifi-cation in the process of pixel enhancement.Then,pipeline operation instructions are used in the algorithm designed by HLS to improve the concurrency and throughput of the calculation.Finally,the synthesized IP core is solidified on the PL side of ZYNQ,and multiple sets of images can be processed and displayed on HDMI display with almost no delay,which verifies the real-time enhancement of image algorithm.

关 键 词：CLAHE 双线性插值 图像增强 HSV颜色模型 

分 类 号：TN919.5[电子电信—通信与信息系统]