摘要： 根据运行中的货车车厢图像,首先采用最大方差法确定货车车厢图像的二值化阈值,然后使用特殊的结构元素进行数学形态学运算提取二值图像中的目标区域,最后计算矩形拟合因子来判别目标区域是否为矩形,从而判别货车车门是否打开,解决了货车运行中的安全状态检测问题。本算法能有效抵抗椒盐噪声以及乘积性噪声的影响,在进行图像处理和识别时无需进行图像预处理。

关键词: 数字图像处理, 目标检测, 数学形态学

Abstract: According to the running-train image, the paper first uses the method of the maximum square error method to make sure the threshold for image segmentation, then uses mathematical morphological operation to extract aim region in the binary image, finally calculate rectangle factor to distinguish whether the aim region is rectangle, and decides whether train door is open, and solve the problem of safe-status detecting when train is running. The algorithm can overcome the affect of salt&pepper and speckle noise, so needn't pretreat the image.

中图分类号: 

• TP391.41