视频的早期烟雾检测技术研究与应用-毕业论文（设计）.doc

PAGE \* MERGEFORMAT PAGE \* MERGEFORMAT 2 论文题目： 视频烟雾识别技术研究与实现 专 业： 学 生： 签名：__________ 指导教师： 签名：__________ 摘 要 火灾是日常生活中一种严重的自然性灾害，具有频发性和不可预测性等特点，给人类社会造成了不可估量的生命财产损失。火灾的前期通常是以烟雾的形势表现出来，因为传统感烟、感温火灾检测技术仍然存在诸多很难解决的问题，不具备实时性预警，所以针对火灾前期烟雾特性的视频图像烟雾检测方法得到许多专家和学者的关注，本文也是基于这个课题，初步研究和设计了一些视频图像烟雾检测的基本算法。 视频图像烟雾检测（VISD）方法通过把一段视频分解成每一帧图像，区分图像烟雾和背景的不同点，比如颜色、形状、纹理等，结合图像在RGB颜色空间以及HSV颜色空间的图像特征，根据颜色的不同可以大致将烟雾区域提取出来，再结合一些形态学操作，膨胀、腐蚀、连接等，得到更加精确的烟雾区域，从而能够更好地做到火灾早期的预警。 【关 键 词】烟雾检测，视频处理，图像提取，图像特征，图像分割 【论文类型】理论研究型 Title： Research and Application of video-based smoke detection technology Major: Electronic information science and technology Name: Signature____________ Supervisor: Signature____________ ABSTRACT Fire is a serious natural disaster in daily life, with frequent and cannot be predicted, caused incalculable loss of life and property to the human society. Fire - usually is expressed with smoke situation, because the traditional smoke, temperature sensing fire detection technology still has many difficult problems to solve, do not have real-time warning, so the detection method for video images for fire smoke the smoke characteristics by many experts and scholars, this paper is based on this topic, a preliminary study on the basic algorithm and some video smoke detection design. Video smoke detection (VISD) method for the decomposition of a video into every frame image, distinguish the smoke and the background image, such as color, shape, texture, combined with the image in RGB color space and the HSV color space, according to the different colors can be roughly the smoke extraction from the region, then combined with some morphological operation, expansion, corrosion, connection, get the smoke area more accurately, so it can be better to do early fire warning. 【Key words】 Smoke detection, Video Processing, Image Extraction 【Type of Thesis】 Th