摘要
该系统采用双波段视频采集技术,综合可见光与红外的视场信息,提高了其抗干扰能力和火灾识别率,并且实现了对火灾的分级报警。算法中采用编号标记、编号生存期等一系列标记机制,可实现多个火焰目标的稳定跟踪识别,并且采用基于图像轮廓的算法,增强系统实时性并减少系统内存损耗;硬件采用DSP与ARM双系统组合,TM320DM642用于图像处理,LPC2368用于通信与报警,系统运行更加高效。实验结果表明,该系统对火灾探测的准确性高,抗干扰能力较强,并且具有较快的反应速度。
The system uses dual-band video capturing technology to improve its anti-interference ability and recognition rate of fire combi ning the field information of visible light and infrared ray. The system realizes the fire classification alarm. The algorithm uses a series of mark mechanism such as number sign and number life cycle, so that it can achieve stable tracking and recognition of multiple flame targets. Based on the image contour algorithm, the design enhances the real-time performance and reduces system memory loss. The hard- ware system uses a combination of DSP and ARM. TM320DM642 is used for image processing and LPC2368 is used for communication and alarm, so that the system is more efficient. Experimental results show that the system ensures high accuracy of fire detection, good anti-interference ability and the significant rapidity.
出处
《单片机与嵌入式系统应用》
2014年第9期28-31,共4页
Microcontrollers & Embedded Systems
基金
国家自然科学基金(61104109)
江苏省自然科学基金(BK2011703)
江苏省科技支撑与自主创新资金资助(BE2012178)
博士点基金(博导类)(20113219110027)
