Video Fire Monitoring System Based On DM642

The traditional fire detection method is to sense the temperature and smoke, light-sensitive detection and infrared detection of radio. However, in the relatively harsh environmental conditions, there are a lot of interference such as the traditional methods of fire detection in the case of dust, electromagnetic, moisture, air conditioning, vibration and interference would be difficult to work in real-time fire detection. The false negative rate and false alarm rate in many fire detection system cannot be well solved. Image fire detection methods explore the digital image processing technology, so it can effectively improve the detection accuracy compared with the traditional fire detection method, greatly reducing the detection time, providing a richer scene of the fire information, fire detection and greatly reducing the rate of false positives and false negative rate. Fire monitoring system is PC-based development of video images collected by a particular structure, and then the acquisition of image transmission to the background processing, greatly increasing the burden of monitoring stations, stability and power consumption are the problem. Therefore, this paper the design and development of a large space of DSP-based video fire detection system, the system has low power consumption, stable performance, easy to upgrade, and fire a high recognition rate.This paper, on the basis of TI’s C6000DSP, designed the DM642for core hardware model system, described the DSP module, video capture module, EMIF expansion module; TI’s DSP/BIOS real-time operating system and RF5program to develop a reference framework based on the development multithreaded tasks including video capture, detection algorithm processing, network transmission applications architecture; research-based fire flame and smoke of the video image segmentation and recognition algorithms proposed a new smoke detection methods, to improve the conventional sparse light method in the multi-target detection must be manually marked; selected the effect of the actual operation is preferably suitable for DSP algorithms; these algorithms using the design of the hardware system realization and program the system and algorithm optimization and improvement. Finally, experimental tests the functionality and performance of the system. Draw the present system can be accurate, real-time detection of fire。...