Computer Engineering and Applications ›› 2015, Vol. 51 ›› Issue (20): 183-187.

Previous Articles     Next Articles

Video surveillance smoke detection system based on OMAPL138

XIE Weibo, HE Wenqiang   

  1. College of Computer Science & Technology, Huaqiao University, Xiamen, Fujian 361021, China
  • Online:2015-10-15 Published:2015-10-30

基于OMAPL138的视频监控烟雾检测系统

谢维波,贺文强   

  1. 华侨大学 计算机科学与技术学院,福建 厦门 361021

Abstract: The design is based on TI OMAPL138, and it implements a video surveillance combined with the smoke detection. This system based on Syslink that it’s using to communication with Davinci heterogeneous multi-core, realizes the communication and processing of video signals between the heterogeneous multi-core processors. Meanwhile it has the advantages of small volume, low power consumption, flexible and efficient algorithm, independent operation etc. This system uses the smoking detection algorithm based on the image frequency characteristics, can effectively detect, and use the unique characteristics of the smoke to avoid error detection caused by moving figures and illumination changes.

Key words: video monitoring, smoke detection, OMAPL138, SysLink

摘要: 基于低功耗浮点DSP+ARM双核架构的高性能芯片OMAPL138,开发一个带烟雾检测功能的视频监控与处理系统。该系统利用Davinci异构多核通信的基础组件SysLink,实现了视频信号在异构多核处理器之间的通信与处理,具有体积小,功耗低,算法实现灵活高效,可以脱离计算机独立运行等特点。在OMAPL138的DSP端运行了基于图像频度特征的烟雾检测算法,综合运用频度图像和灰度图像相对于基准图的相关系数,既能检测浓烟也能检测稀薄烟雾,并利用烟雾的独有特征避免了物体移动与光照变化等场景引起的误检测。

关键词: 视频监控, 烟雾检测, OMAPL138, SysLink