基于图像处理的大空间建筑火灾报警系统研究

发布时间：2019-03-31 19:59

【摘要】：火灾图像自动监测技术是基于图像处理的优点和计算机语言的高速运算性，提出的一项利用火焰图像的信息来分析判断火灾的监测技术。并实现了基于图像的火灾报警系统，本文针对该系统做了三方面的研究。首先，由于火灾火焰识别算法复杂且对环境要求高，提出了基于图像多特征的火焰识别方法。通过图像差分法对运动火焰疑似区域进行检测，然后将检测到的区域设为感兴趣区域，再通过对感兴趣区域内的疑似火焰目标进行纹理特征和形状特征的图像分析，排除不具有火焰颜色的物体，进一步进行判断决策，最终判断出是否有火焰产生。实验结果表明，，该方法具有较高的探测率，能降低误报漏报率且抗干扰能力强。文章第二部分研究了通过双目标定对火焰位置进行三维定位，其次对摄像机的几种标定方法进行了研究，最后介绍了火灾图像中心坐标的计算方法和几何特征的参数计算，并提出了一种改进的火灾图像中心坐标的计算方法。文章最后设计了基于图像处理的图像型火灾报警系统。系统的硬件设计，包括了系统的总体构成以及硬件组成；系统的软件设计，包括软件的设计要求、设计共嫩、需求分析、开发语言的选择和软件的整体结构；最后针对本文的系统设计，进行了火灾识别仿真实验。
[Abstract]:Fire image automatic monitoring technology is based on the advantages of image processing and the high-speed operation of computer language. Based on the information of flame image, a fire monitoring technology is proposed to analyze and judge the fire. The fire alarm system based on image is realized, and three aspects of the system are studied in this paper. Firstly, because the fire flame recognition algorithm is complex and requires high environment, a fire recognition method based on image multi-features is proposed. The image difference method is used to detect the suspected area of moving flame, then the detected region is set as the region of interest, and then the texture and shape features of the suspected flame target in the region of interest are analyzed. Exclude the object without flame color, make further decision, and finally determine whether there is flame or not. The experimental results show that the proposed method has higher detection rate, lower false alarm rate and strong anti-jamming ability. The second part of the paper studies the three-dimensional positioning of flame position by binocular calibration. Secondly, several calibration methods of camera are studied. Finally, the calculation method of fire image center coordinates and the parameter calculation of geometric characteristics are introduced. An improved method for calculating the central coordinates of fire images is presented. Finally, an image-based fire alarm system based on image processing is designed. The hardware design of the system includes the overall structure and hardware composition of the system, the software design of the system, including the software design requirements, requirements analysis, the choice of development language and the overall structure of the software. Finally, according to the system design of this paper, the fire identification simulation experiment is carried out.
【学位授予单位】：长安大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TU892;TP277

【参考文献】