摘要
自然光状态下拍摄的图片大都存在阴影,影响图像处理的精度。针对这一问题,本文研究了不同时间下目标物与其阴影之间的特定对应关系,结合阴影部分图像的基本特征,建立了阴影与本体在图像中的数学模型。在此基础上,提出了一种基于RGB色彩空间的自然光状态下图像中目标物阴影检测与去除方法。通过合理、必要的假设,阐述了目标物去除阴影的步骤,将阴影部分在原始图像中有效地去除。实验表明,该方法对阴影去除的准确度在91%以上,单幅图片的处理时间小于20ms。
Images obtained under natural light mostly have shadows, which seriously effect the precision of images process. For this problem, the paper researches the special relation between objects and their shadows in different time. Considering the basic characteristics of shadow images, the mathematic model of objects and shadows are founded. Based on this work, a method for shadow removal under natural light in RGB space is proposed. The detailed steps of removing shadows are also introduced by reasonable hypothesis. Experiments show that the ac- curacy of removing shadows is above 91% and processing time for a single image is less than 20ms by proposed method.
出处
《中国农机化学报》
北大核心
2013年第6期120-124,共5页
Journal of Chinese Agricultural Mechanization
基金
山东省自然科学基金(ZR2012CQ026
ZR2011EL038)
山东省高等学校科技发展计划项目(J11LD16
J12LB63)
