【作者】 朱为鹏；

【导师】 肖世德；

【作者基本信息】 西南交通大学 ， 机械电子工程， 2003， 硕士

【摘要】 随着铁路安全技术装备的完善，现在铁路运营安全的主要威胁来自因装载、冲撞、制动、震动等因素引起的货车所载货物超出铁路货运装载限界。目前，我国铁路货运管理部门主要采用人工测量方法来检测货车超限情况。但是人工测量只能实现对静止货车装载情况的检测，并且检测效率低，检测精度也难以保障。为了确保铁路货运的运营安全，实现对运行中货车的装载情况的自动监测是必要的。 本文以数字图像处理原理中图像分割理论为基础，结合货车装载状态监测的实际情况和货车装载状态监测系统对图像分割的要求，以保证超限漏检率最小为目标，给出了货车图像自动分割的算法，获得货车及其所装载货物的分布区域，再根据该分布区域，得出货车装载状态是否超限的结论，从而实现对运行中货车的装载情况的自动监测。采集货车运行图像对算法进行测试，测试结果表明，该算法适应性较好，检测精度和速度基本能满足要求。更多还原

【Abstract】 The threats of the security of raiIway freight mostly come from cargoes exceeding over the railway gauge which caused by loading, collision, braking and shaking etc. Today, the loading situation is most Iy checked manua My, wh i ch can on Iy check the Ioad i ng s i tuat i on of the static freight train with the low efficiency and precision. The automat i c mon i tor i ng on the I oad i ng s i tuat i on of mov i ng f re i ght tra i ns should be implemented to ensure the operation security of railway fre i ght.The paper presents a new algorithm for freight train image segmentation with the aim of minimizing missing gauge-exceeding detection rate based on the image segmentation theory. The prof i le of the freight train and cargoes on the image is obtained with the arithmetic and whether cargoes exceeding the gauge or not is concluded by the location of the freight train and cargoes on the image. The algorithm is verified by images of moving freight train. It is proved that the algorithm has good adaptability, acceptable precision and satisfied processing speed in normal climate.更多还原