Machine Vision-Based Intelligent Fire Fighting Robot

摘要

Designing a visual monitoring system to detect fire flame is a complex task because a large amount of video data must be transmitted and processed in real time. In this work, an intelligent fire fighting and detection system is proposed which uses a machine vision to locate the fire flame positions and to control a mobile robot to approach the fire source. This real-time fire monitoring system uses the motion history detection algorithm to register the possible fire position in transmitted video data and then analyze the spectral, spatial and temporal characteristics of the fire regions in the image sequences. The fire detecting and fighting system is based on the visual servoing feedback framework with portable components, off-the-shelf commercial hardware, and embedded programming. Experimental results show that the proposed intelligent fire fighting system is successfully detecting the fire flame and extinguish the fire source reliably.