Research On Infrared Stereo Vision Based Tunnel Fire Positioning System

As an important transportation infrastructure, Highway tunnel plays a huge role in thetransportation industry. With a line type and semi-enclosed environment, once a fire occurinside the tunnel, if not timely controlled, will cause incalculable damage to people andproperty. Currently domestic work undertaken mainly focused on fire detection and alarm,There is no depth functional integration of new fire detection technology and thetraditional fire-fighting equipment. In order to reduce personnel, property damage, protectthe safe and reliable operation of the tunnel， the new concept of automatic fireextinguishing robot will promote firefighting equipment to get rid of the entirely dependenton staff. In this paper, an infrared stereo vision based tunnel fire positioning system wasproposed, as the "eyes" of automatic fire-fighting robot, can positioning tunnel fire sourcetimely and accurately, help the fire-fighting robot to extinguish fire. A long-wave infraredthermal imager was utilized to convert infrared radiation of object into an analog infraredimage, the analog signal was digitized and entered into the computer by the image framegrabber, then infrared image processing algorithm was utilized to identify and calculate thespatial location coordinates of fire, finally the fire extinguishing system will be startedautomatically to extinguish fire. The main contents are as follows:1. Color space based multicolor image segmentation algorithm was applied to extractthe red, yellow and white color region (high temperature target)in the infrared fire images,then determine whether the suspicious area is the flame of fire based on the static anddynamic characteristics of fire, and achieved good recognition results.2. The mathematical model of stereo vision positioning system was established. Thescope and accuracy of positioning system were analyzed to determine the parameters ofstereo vision system based on fire protection standards. Internal and external parameters ofthe stereo vision positioning system were calibrated, rotation and translation matrix wereutilized to describe the relative position of the two thermal imagers, feature point in twoinfrared images was matched, and finally, fire coordinates position was calculated by the triangulation principle.3. Tunnel fire detection, localization algorithm was realized by software programmingbased on the study above theoretical. Feasibility of the system was validated and verifiedby analyzing the experimental data and diagram, the system was proved to be effective,with high accuracy and reliability.