基于序列图像的阴影检测研究

摘要

针对由于入射光的照射,目标阴影与背景有明显的区别,而被当做目标的问题进行了研究.分析了单幅图像的阴影检测方法,提出了序列图像阴影检测方法,建立了序列图像高度测量模型；采用相邻序列图像对应点匹配的约束条件,降低了匹配的复杂度；研究了序列图像的特征点匹配,获得平面目标的高度信息表示；基于阴影的二维表面属性,设计了利用空间三维信息中的高度信息检测阴影的算法,可以对任意形状的阴影进行检测.通过试验表明,该方法简单,适用面广,可以有效地检测阴影.%For the light rays, shadows and background targets are significantly different. The shadow detection method based on single image is analyzed widely, a shadow detection method is proposed based on image sequence, and the space height measurement model is established. The matching constraints are researched on corresponding point of the sequence im-age, the complexity is reduced. The height information is calculated with the feature points matching of sequence image. The shadow detection algorithm by using the height information of three-dimensional spatial information is designed, which can de-tect the shadow of any shape. The experiment shows that the method is simple and applicable for many fields to detect shadow effectively.