Research And Application Of Fire Source Matching Technology Based On The Image In Spacious Buildings

With the rapid development of image processing techniques and computer vision technology, image fire detection and location technology has become a hot research topic in recent years. Automatic fire extinguishing system mainly uses image analysis technology and stereo matching technology of the computer vision to complete accurate positioning of the flame. The fast matching of the flame plays a very important role in automatic fire extinguishing system. At the same time, it also is a hot issue in the research of the fire extinguishing system. The main problems existing in flame image matching are:(1) how to extract the characteristics of flame from the image of large amount information;(2) the influence of complex background on the fast extraction of effective feature points;(3) the effect of light on the characteristics of flame, etc. This thesis focuses on the problem how to quickly extract the flame characteristics of flame image.Based on studying the theoretical knowledge of many kinds of image matching algorithms, this thesis makes a simulation experiment analysis that SURF algorithm is more stable and faster. But the false matching points are more and the feature points are less. Therefore, this thesis mainly aims at improving the problem of SURF algorithm. It includes two aspects:(1) Matching algorithm based on image segmentation and SURF algorithm.Firstly, image is transformed to HSV color space. Secondly using color gradient information to extract the edge of flame and segment the image. Thirdly, applying SURF algorithm to extract and match feature points. The color gradient information isused to remove the interference of the background feature points to the flame characteristic.(2) Matching algorithm based on image segmentation and ORB(Oriented FAST and Rotated BRIEF) algorithm. In the HSV color space, edge detection and image segmentation are performed by using color gradient information, and then the ORB feature points are extracted and matched.Simulation result shows that the accuracy and speed of the extraction and matching of flame features are obviously improved. By using the parameters of the camera calibration, the depth information of the three-dimensional space is calculated.Finally, this thesis designs the system interface by using Visual Studio 2012 according to the principle of the automatic fire extinguishing system.