Research On Fire Detection Based On Vision And Path Planning Of Mobile Robot

Forest resources are one of the most important resources in the world. In recent years, forest fires occur frequently, which cause great damage to people’s property.Therefore, it is particularly important to prevent forest fires. With the development of science and technology, image processing technology is gradually being known by people. The common forest fire monitoring method is fire detection based on video.But this method has strict requirements on the number and quality of the camera, high cost, and cannot fully monitor forest fires. But, fire detection robot can solve this problem. This paper applies fire detection algorithm based on video to mobile robot and plans patrol path of mobile robot.Traditional forest fire detection algorithms are mostly simple analysis of smoke or flame to get results and the results are affected by light and other factors which causes low accuracy. In fact, forest fires can produce a lot of smoke, which can be considered as the initial signal of fire. So it is really important to detect forest fire with characteristics of smoke and flame.This paper combines the characteristics of smoke and flame to judge if fire occurs. First, analyzing color of smoke to obtain color rule and extracting color region.Then analyzing flutter characteristics of smoke and detecting smoke region by moving target detection algorithm. At last, combining color and flutter features to get suspected smoke region and estimating motion direction to get real smoke region.After determining the presence of smoke, it is necessary to detect flame. This paper analyses color feature of flame and improves color rules to extract fire region.And analyzing flutter characteristics of flame and detecting fire region by moving target detection algorithm. At last, analyzing flicker frequency of flame to judge iffire occurs.This paper plans patrol path in forest environment. For the characteristics of mobile environment, it builds two kinds of environment models including obstacle environment and barrier free environment. And improved ACO algorithm is used to plan patrol path. Simulation results show that improved ACO algorithm can quickly find the best path.In order to verify the effectiveness of the proposed algorithm, this paper detects fire in different scenes based on mobile robot platform. The experimental results show that the robot can detect smoke and flame accurately and effectively in the no-smokeand flame interference, and has certain anti-interference ability.Finally, the algorithm is verified by the mobile robot that it can achieve a better recognition effect to fire in the process of the motion of robot。...