基于FPGA的视频火灾检测系统设计

Video fire detection system design based on FPGA

为了提高火焰检测精度和减少装置的成本,在Altera公司Cyclone II系列的FPGA实验板上实现了基于视频的火焰检测原型系统设计与开发。该系统在FPGA上建立混合高斯模型,对视频进行前期处理,极大地减少了类火焰区域的提取时间。随后将处理过的视频数据,通过DMA模式送入SOPC系统中的SDRAM存储,并进行后期算法处理。对20多个火灾视频进行测试与研究,实验结果表明,该系统能有效地检测火灾发生,具有较好的实时性,并且降低了成本,利于集成。

To improve the fire detection precision and reduce the device cost,a video-based fire detection system is realized on the FPGA experimental board exploited by the Altera company.Firstly,the Gaussian mixture model is built on the FPGA,then the pre-processing is done to the video,which greatly reduce the extraction time of the fire region.Next,the processed video data is stored in the SDRAM of the SOPC system via the DMA mode.Meanwhile,the post-processing algorithm is carried out.Eventually,more than twenty fire videos are tested.Experimental results show that on the one hand,the fire can be detected effectively and in real time,on the other hand,the total cost can be cut down a lot,and is also good for integration.