Study On Method To Improve Ranging Precision For Anti-collision Warning System Of Tramcar

The vision ranging technology is an effective measure for the mine locomotives'safe transportation. Anti-collision warning system based on monocular vision grab image through the infrared CCD camera, make processing of digital image, identify target and range distance. Based on the distance, the system outputs warming signal, so that prevent collision between tramcars. But, the vibration between locomotive and track or the collision between carriages makes the image sequence unstable. Quality of image stabilization determines the accuracy of target detection and the precision of distance measurement, which determine the performance of system. In addition, the front locomotive's detection is also important. Accurately determine the locomotive, and judge its concrete position, has the very important influence to the range precision.In the paper the algorithms of improving the ranging precision are designed mainly include as follows: First, according to the orbital characteristic, selecting the best edge detection algorithm, proposing the Hough transformation based on the known point to extract track edge. Second, the paper proposed an electronic image stabilization algorithm based on the invariance of orbital character. Third, about the locomotive determination, this paper researched in two conditions. One is opposite direction, the paper proposed a range algorithm based on tramcar's light. And another is the same direction, through the circular and symmetrical of locomotive wheels realize locomotive's examination and range.Finally, by the examination, the paper verified the accuracy of locomotive's recognition and increased the range precision. This paper analyzed the reason of range error, and the further study was put forward.