中文摘要：

针对列车运行故障图像动态检测系统（Trouble of moving Freight car Detection System，TFDS）中挡键丢失故障，提出一种基于形状上下文的列车挡键丢失图像识别算法。取正常挡键区域图像作为模板，到待测TFDS图像中遍历，采用形状上下文描述图像的形状特征，加权形状上下文距离与弯曲能量以定义形状距离作为图像匹配的相似度指标，最后根据模板图像是否遍历出与其相似的区域图像作为挡键丢失的判断依据。采用Matlab编程，通过截取大量测试图像实验发现，所定义的形状距离阈值取0．16，对测试图像中有无挡键能很好地区分。采用形状上下文描述，自定义形状距离作为图像匹配的相似度指标具有很高的可靠性，该算法为TFDS图像故障识别提供了一种新的思路。

英文摘要：

For the missing fault of side frame key in trouble of moving freight car detection system （TFDS）,an image recognition algorithm for side frame key of train was proposed based on shape context.A regional image with side frame key was taken as a template for searching in the TFDS images under testing.The shape charac-teristics of these regional images were described by shape context.A new shape distance was defined as the simi-larity index for image matching by weighting shape context distance and bending energy.Accordingly,the judg-ment of missing side frame key was made if the similar region in the TFDS image was searched by the template. Through a large number of testing image experiments using Matlab programming,the testing images with side frame key can be judged well by choosing 0.16 as the shape distance threshold.Using shape context descriptor, the shape distance was defined as the similarity index for image matching with high reliability.This algorithm provides a new approach to fault recognition in TFDS.