Design And Implementation Of Video Fire Detection System Based On FPGA

Fire as a frequent natural disasters, it brings a great threat to the human life and social security. Fire in large span space not only destroys the social life but also makes a great demage to the society. The traditional fire detection method can’t satisfy. In recent years, with the rapid development of computer technology and digital image processing technology, the fire detection method based on video is becoming an impotant development. The FPGA can meet the needs of high speed image processing with its hardware features. SOPC technology can make the design flexible, software and hardware in-system programmable and updatable.This paper designs a video fire detection system based on FPGA. The system uses DE2development board of Altera as the hardware platform. CMOS image sensor monitors the environment in real time. The monitor displays the image getting from CMOS image sensor. When there is fire in the environment, then the system will send the alarming signal to notify the monitoring staff by the headphones and the green LED of DE2development board. The green LED shows the fire grade in binary.The system is achieved in a synergistic manner of software and hardware. Hardware: first, on the basis of the requirement analysis, the paper designs and implements all the modules, including image capture module, SDRAM module, image display module and fire detection module; and then the paper builds the system in hardware design language by Quartus II. Using the SOPC builder in Quartus II to construct the sopc_system module for the basis of analysis algorithm behind. Algorithm:after having research of the fire detection based on video, the paper adopts the absolute difference of edge area algorithm, RGB rule to pretreat the image and the average of the intensity in the area from the RGB rule, and FFT algorithm to ananlyze the data to detect fire, so that the detected result can be more accurate. Software:the algorithms are programmed and debugged by C language in the Nios II IDE environment. Moreover, the thesis optimizes Nios II software code with its own optimization and adds a floating-point unit to the Nios II system. The floating-point unit can reduce the excuting time of FFT by hardware instead of software. After a period of debugging and running, the system achieves the expected results.