基于视觉识别的多点分散式火灾监控方法研究

Research on Method of Distributed and Multipoint Fire Monitoring Based on Visual Identification

为解决传统火灾监控方法火点识别精度不高、实时性差、布线复杂等问题,提出了基于视觉识别和无线通信的火灾监控新方法。在火点识别算法中,采用火焰静态特性和动态特性相结合的方法将摄像头采集的现场图像进行火点真伪识别。在火点信息传输中,采用无线通信技术实现上位机和PLC控制器的通信。实验证明,选取合适的火点脉动识别阈值k_f,系统可极大减少非火点光源的干扰,提高了火灾的识别精度,增强了识别的实时性,同时,采用无线通信解决了系统布线复杂、可靠性不高的问题,为多点分布式环境下的火灾监控提供了一种新的解决方案。

In order to solve the problems of traditional fire monitoring methods such as low accuracy, poor real- time performance and complicated wiring, a new method of fire monitoring based on visual recognition and wireless communication is proposed. In the fire spot recognition algorithm, the authenticity of fire spot in the image captured by camera is identified by the method of combining flame static characteristics and dynamic characteristics. In information of fire transmission, wireless communication technology is used to realize the communication of upper computer and PLC. By selecting the appropriate value of recognition threshold, the ex- periments prove that the system can greatly reduce the interference of non-fore point light somme, improve the i- dentification precision of fire, and enhance the real-time performance of recognition. At the same time, wireless communication is used to solve the problems of complicated wiring and low reliability, which provides a new solution for the fire monitoring in muhipoint distributed environment.