多核DSP图像去雾优化方法  

Optimization Methods for Image Haze Removal Based on Multi-core DSP

作　　者：白林亭 武永卫[1] 谢建春[2,3] 文鹏程[2,3] 刘小剑[2,3] 

机构地区：[1]清华大学高性能所,北京100084 [2]中航工业西安航空计算技术研究所,西安710065 [3]机载弹载计算机航空科技重点实验室,西安710065

出　　处：《电光与控制》2015年第10期14-18,共5页Electronics Optics & Control

基　　金：航空科学基金(2014ZC31004)

摘　　要：数字信号处理器(DSP)专门针对信号处理而设计,以其独特的特点在信号处理、图形图像处理、通信、交通、航空航天等领域得到越来越广泛的应用。在图形图像处理领域中,图像去雾一直是研究热点,开发时多使用Matlab和C语言仿真,将其直接移植到DSP中运行效率通常较低,难以做到实时处理。针对该问题,提出一种快速有效的图像去雾算法,优化算法过程,获得更好的处理结果;利用TI公司的多核DSP和开发工具,对C代码进行实时优化和并行处理,取得很好的优化效果,能够对低分辨率进行图像去雾的实时处理。The Digital Signal Processor（ DSP） is specially designed for signal processing. In recent years, the DSP has found wide application in signal processing, image processing,communication, traffic and aerospace etc. As a practical part of image processing,the image haze removal has been researched for decades. During the research and development of haze removal algorithm, the Matlab and C programming language were usually used, and were supported by DSP. Yet, the efficiency was too low by simply running the source code. To solve this problem, a novel algorithm is proposed for optimizing the algorithm and improving the processing effect. The multi-core DSP and development tool of TI Company are used for realtime optimization of C code and parallel processing. A fine optimization effect is obtained,which shows that the method can be used for processing hazy images with low resolution ratio in realtime.

关 键 词：图像处理 多核数字信号处理器 图像去雾 实时优化 并行处理 

分 类 号：TN956[电子电信—信号与信息处理] O221.6[电子电信—信息与通信工程]