基于DM642夜晚视频增强算法研究与实现-控制理论与控制工程专业论文.docxVIP

系统处理，优化算法程序、分析比较 DSP 处理效果后，得到一种最适 合 DSP 高速处理、增强效果最明显的处理算法，这一算法不仅提高了 夜晚视频的整体对比度，而且保持了夜晚视频的色彩同时降低了夜晚 视频的噪声。本文提出的 DSP 系统设计方案，实现了高速、稳定、灵 活的夜晚视频增强与播放平台，达到了设计的要求。既可以满足一般 的监控需求，在夜晚情况下，提高监控视频质量，同时，本系统可以 升级成恶劣天气的通用视频处理平台。 关键词：夜晚, 视频增强, 图像处理，DM642,直方图均衡，单尺度 Retinex 算法，MEAM 算法 THE RESEARCH AND IMPLEMETION OF THE REAL-TIME NIGHT ENHANCEMENT ALGORITHM BASED ON DM642 PLATFORM ABSTRACT With the development of modern science and technology, real-time image processing has brought widespread concern. However, in the night or low visibility conditions, due to inadequate lighting human can not easily observate the surrounding environment. Similarly, for the ordinary camera monitoring system it is difficult to capture the details in a certain range. Such a situation brings an adverse impact to the night road traffic on video surveillance. So we can improve quality of the original video through image processing algorithms. With the development of DSP technology, more efficient image processing algorithms can be realized through high-speed DSP processing chip. Meanwhile real-time DSP system with the function of video processing has a broad market prospect. In this paper, the high-speed DSP technology makes it possible to implement the real-time night video enhancement algorithms. The system uses an ordinary camera to capture video, through encoding and decoding algorithm by DSP chip then did image processing and optimization, after that output the video signal to screens for display. Not only the quality of night vedio was enhanced, but also the speed of the video processing was improved. This technology has an important value in practice. This paper was focused on the night video enhancement algorithms and its implementation in the DM642 system. Firstly it introduces the basic algorithms for image enhancement in night, including the space-domain and frequency-domain algorithms, such as gray-scale stretch, a variety of histogram equalization algorithm, based on single-scale Retinex transf