The Recognition Of Obstacle Between The Metro PSD And Train Based On Video Detection

Video image processing technology is one of the core technology of computer vision, isbased on the video scene analysis, behavior description, and understanding the basis of manysubsequent processing, such as link. Video image processing technology in the field ofintelligent video surveillance, traffic, human-computer interaction, robot visual navigation,virtual reality and so on various aspects has awide range of applications.For guangzhou metro line network’s scale expanding and increasing passenger flow, andthe efficient operation of metro safety needs, this paper puts forward a kind of based on thevideo detection technology using long high definition camera to recognize the obstaclesbetween the subway train and the platform screen door solutions. At the same time, due to theactual operating environment in the narrow space of guangzhou metro, and highlights thelight at the end of the platform features, principle of this scheme is mainly for the use of longdistance high definition camera to highlight light at the end of the subway platform, whichcan identify shade highlight how to judge whether passengers or other object, in order todistinguish whether there are obstacles exist in the narrow space between the subway trainand metro shielding door, so that the metro subway train driver efficient and quick andconvenient to control the closing of the door.This article mainly on camera recognition highlight band image processing technologyhas carried on the exploration and research, mainly related to the research contents are asfollows.First, in view of the actual use of the camera corner problems put forward a kind oftechnology solutions to the camera corner, this scheme can completely solve theoreticallycorner; According to the actual use of highlighting strip light source made some improvement，At the same time this article puts forward the corresponding monitoring schemefor theplatform may be built into a corner in the future.Second, the common edge detection algorithm is analyzed, according to the differentdetection algorithm is the actual effect of selecting one of the mostappropriate algorithms.Third, according to the stated theory, using the Visual Studio2010and OpenCV visionlibrary platform to achieve the obstacle detection platform, combining with the actual experimental results analysis, proves the rationality of the platform technology, can achievethe desired effect.In the end, this paper made a summary and some prospects of the application of theobstacle detection platform.