一种基于FPGA的图像自动增强算法与实现

摘要

Since there are many too dark or too bright areas in pictures that shot in the space,the corresponding images are uneven distributed,and the image quality is damaged.How to improve the processing speed of the algorithm and how to implement the algorithm by the embedded system are also key problems for space application.In this paper,the automatic image enhancement algorithm which is suitable to be implemented by the FPGA is proposed to solve these problems.The basis of the proposed algorithm is the piecewise linear transformation algorithm.In practical,firstly,the K-means clustering is used to segment the histogram into several sections automatically.Secondly,the quantitative relationship between the histogram distribution and the coefficients of the piecewise linear function is established.As a result,the coefficients can be automatically calculated.Thirdly,the corresponding FPGA system is implemented.And the high-performance parallel pipelined technology is used to ensure the real-time processing ability of the system.The simulations and the experimentations show that the proposed FPGA system is characterized as real-time processing ability and good adaptability.It can achieve good processing effects for different sceneries,and can be used in various practical applications.%图像灰度分布不均匀是影响画质的重要因素,集中表现在对空间场景成像时,图像中存在大量过亮或过暗的区域.同时,算法的实时性和嵌入式平台的可实现性是设计实现空间应用载荷需要着重考虑的问题.针对上述问题,提出一种适合于FPGA实现的图像自动增强算法.算法以分段线性变换方法为基础,采用K均值聚类对视频图像的灰度直方图进行动态的区间划分;构建分段函数系数与直方图之间的定量关系,自动计算线性变换系数;采用高效的并行流水线结构设计,实现了基于FPGA的硬件处理系统.仿真和成像试验结果表明,该FPGA硬件系统实时性好,适应性强,针对不同的场景均取得较好的处理效果,具有广阔的应用前景.