CAJ | 학술논문

为了能够在多种检测环境中，对森林烟火实现较为精准的定位，达到降低大规模森林火灾发生的可能性，文中研究了视频中基于慢运动物体检测、烟雾颜色区域检测、上升烟区检测、阴影检测与去除等四个子算法；利用最小均方算法对以上四个子算法进行加权；结合OpenCV图像分析处理技术和C++编程，设计了基于图像处理的森林烟火检测系统。实验结果表明，系统具有自动化、智能化程度高，对运行环境要求宽松，使用简单便捷等诸多优点。
위료능구재다충검측배경중，대삼림연화실현교위정준적정위，체도강저대규모삼림화재발생적가능성，문중연구료시빈중기우만운동물체검측、연무안색구역검측、상승연구검측、음영검측여거제등사개자산법；이용최소균방산법대이상사개자산법진행가권；결합OpenCV도상분석처리기술화C++편정，설계료기우도상처리적삼림연화검측계통。실험결과표명，계통구유자동화、지능화정도고，대운행배경요구관송，사용간단편첩등제다우점。
In order to accurately find the location of the fire and make the forest fire losses to a minimum degree in a variety of detection environment,fire detection based on four algorithms like slow moving video object detection,smoke-colored region detection,rising vide-o object detection and shadow detection and elimination is researched. Sub-algorithms are weighted using least mean square. Wildfire de-tection system based on image processing combining OpenCV image analysis processing technology and C++ programming is designed. Experimental results show that the system has high automation and intelligence,loose requirements of runtime environment,convenient operating and many other advantages.