基于TMS320DM6467的监控视频去雾算法实现

Implementation of Surveillance Video Defogging Algorithm Based on TMS320DM6467

有雾天气情况下拍摄的视频成像质量退化严重,极大地影响了安防监控、视觉导航等系统的整体性能。面向DSP嵌入式平台,提出一种实时监控视频去雾系统实现方案。首先在暗通道先验单幅图像去雾的基础上,提出了一种监控视频去雾算法;其次,基于TMS320DM6467处理器的硬件资源,对视频格式转换、除法运算、透射率计算、数据搬移等环节进行代码优化,最终在硬件平台上实现了整体系统。实测结果表明,对于D1格式的视频,在兼顾去雾质量的同时,达到了实时处理要求。

The quality of video image is seriously degraded in foggy weather,which greatly affects the overall performance of the system,such as security monitoring,visual navigation and so on.Based on DSP embedded platform,an implementation scheme of real time video defogging system is proposed.Firstly,on the basis of single image fog removal method using dark channel prior,an algorithm is designed to eliminate fog in the video.Second,based on the hardware resources of TMS320DM6467 processor,several strategies for code acceleration are adopted,such as video format conversion,division operation,transmittance calculation,data transmission and so on.Finally,for Dl format video,the proposed scheme achieves real-time processing on the embedded hardware platform.