计算机工程与应用 ›› 2017, Vol. 53 ›› Issue (22): 1-7.DOI: 10.3778/j.issn.1002-8331.1707-0487

• 热点与综述 • 上一篇    下一篇

融合色彩比和梯度不变性的运动阴影检测

廖  娟1,朱德泉1,李  勃2,刘  路1,陈启美2   

  1. 1.安徽农业大学 工学院,合肥 230036
    2.南京大学 电子科学与工程学院,南京 210046
  • 出版日期:2017-11-15 发布日期:2017-11-29

Moving shadow detection based on color ratio and texture invariance

LIAO Juan1, ZHU Dequan1, LI Bo2, LIU Lu1, CHEN Qimei2   

  1. 1.School of Engineering, Anhui Agricultural University, Hefei 230036, China
    2.School of Electronic Science and Engineering, Nanjing University, Nanjing 210046, China
  • Online:2017-11-15 Published:2017-11-29

摘要: 为解决运动前景的准确分割受运动阴影影响的问题,提出了一种融合色彩比和梯度不变性的运动阴影检测算法。该算法分析了阴影像素的色彩比和区域纹理梯度的光照不变性,利用亮度变化特性和色彩比不变性初步确定候选运动前景中的阴影像素,然后在候选阴影区域利用纹理梯度不变性进行去错处理,两者的结合弥补了单一特征或单一类型特征的阴影检测性能差的缺陷,提高了阴影检测率和阴影分辨率,能够准确地将阴影和前景区别开来。

关键词: 运动阴影检测, 色彩比不变性, 梯度不变性, 色彩张量

Abstract: To solve the disturbances caused by moving shadows in accurate segmentation of moving object, a moving shadow detection algorithm based on color ratio and texture invariance is proposed. The algorithm analyzes the illumination invariance of the pixel color ratio and region texture gradient, and uses brightness change characteristics and color constancy to detect shadow pixels preliminarily in moving object region in pixel level. Then it utilizes texture gradient invariance to eliminate mistakenly identified shadow region at the regional level. The combination of the two types of features can make up for the defect of poor performance of shadow detection method based on a single feature or single type characteristics, and improve the detection and resolution ratio of moving shadow. The proposed algorithm can distinguish moving shadows and objects under different scenes.

Key words: moving shadow detection, color ratio invariance, texture invariance, color tensor