The Design And Implementation Of Fire Smoke Detection System Based On FPGA

Fire brings a great threat to human life and property security. Fire hazard can be greatly reduced if we can realize early warning of the fire. At present video fire detection technology can be roughly divided into two groups, based on fire smoke and flame. Both of them are analyzed by flame and smoke color and motion feature.Smoke happens before fire and relative to fire flame, smoke image can be more easily captured into the camera. So fire can be detected in the initial stage, which is helpful to realize the early warning and control of fire.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 updata.This paper designs a fire smoke 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 captured by 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 LEDG of DE2development board.This paper uses modular method to design smoke detection system. On the basis of the requirement analysis, the paper designs and implements all the modules, including image capture module, SDRAM data buffer module, image display module and smoke detection module. Among them the smoke detection module is the core of the system, which decides the performance of the system. Using a synergistic manner through software and hardware, the smoke detection module is realized. First of all, using the algorithm of frame difference, erosion and dilation, realizes suspected smoke area extraction with hardware description language. This method ensures to process real-time sequential image. And then we detect the suspected smoke areas with smoke dynamic features including smoke area rising feature and smoke regional stability feature. Finally, the paper builds smokedetect module using SOPC Builder, which reads image data captured by CMOS image sensor monitors into SRAM. Based on smokedetect hardware module this paper realizes smoke background fuzzy characteristics based on the2d discrete wavelet transform in Nios II IDE environment. After a period of debugging and running, the system realizes real-time smoke detection.