基于无人机视觉的森林火情预测-《计算机技术与发展》

文章信息/Info

Title:: Prediction of Forest Fire Based on Machine Vision ofUnmanned Aerial Vehicle

作者:: 王君¹ ; 蒲磊¹ ; 黄宁¹ ; 陈楷升¹ ; 梁薇薇² ; 杨振威¹; 1. 中山大学南方学院电气与计算机工程学院,广东广州 510970;
2. 重庆邮电大学自动化学院,重庆 400065

Author(s):: WANG Jun¹ ; PU Lei¹ ; HUANG Ning¹ ; CHEN Kai-sheng¹ ; LIANG Wei-wei² ; YANG Zhen-wei¹; 1. School of Electrical and Computer Engineering,Nanfang College of Sun Yat-sen University,Guangzhou 510970,China;
2. School of Automation,Chongqing University of Posts and telecommunications,Chongqing 400065,China

Keywords:: forest fire prevention early warning; fire rescue; multi-feature fusion; prediction of fire; UAV vision

摘要:: 森林火灾危害严重,容易引起生命和财产损失,救援任务相对困难。然而,在森林火灾发生初期,准确的火情检测和火情识别,可以提前预警,降低生命和财产损失。利用无人机视觉监控火情图像,辅助救援,提高救援效率,保障救援中的人员安全。该文利用无人机实时获取森林状态图像,通过图像处理来自动识别火情。然而,单一特征难以准确分析和判断火情趋势。为了提高森林火灾识别的准确度,对森林火情的识别进行了研究,提出处理提取图像特征,并将图像颜色、尖角数以及烟雾综合特征结合的方法,能实现对火情区域内的有效辨别,并预测火情的蔓延趋势。实验结果表明,该方法对于火情判断的准确程度较高,提升了识别能力,为后续的救援任务提供有利参考。

Abstract:: Forest fire damage is serious,easy to cause loss of life and property,and the rescue mission is relatively difficult. However,in the early stage of forest fire, accurate fire detection and fire identification can lead to early warning and reduce the loss of life and property. The unmanned aerial vehicles are used for visual monitoring of fire images to assist rescue, improve rescue efficiency, and ensure the safety of rescue personnel. We use the unmanned aerial vehicles to acquire real-time forest image and automatically identify fire situation through image processing. However,a single feature is difficult to accurately analyze and judge the fire trend. In order to improve the accuracy of forest fire identification,we study the recognition of forest fire and put forward the method of processing and extracting image features,combining image color,sharp angle number and comprehensive characteristics of smoke,which can realize the effective identification of the fire area and predict the spread trend of fire. Experiment shows that the proposed method has high accuracy for fire judgment,enhances the reliability of forest fire detection system,and provides convenience for real-time rescue.