摘要
最大类间方差法(Otsu法)因其计算简单、自适应性强而成为被广泛使用的图像阈值自动选取方法.在分析Otsu法原理的基础之上,提出了一种改进的最大类间方差法.为了提高分割效果,该方法同时考虑了背景和目标的类间距离和类内距离.与同类方法相比,提出的方法将目标和背景所占的比例作为权值修正了现有的方法,使得衡量类内距离的目标与背景的平均方差按照目标与背景的面积划分.Lena、Cameraman标准测试图像以及杂草图像的仿真结果验证了本方法的有效性.
Otsu algorithm is widely used in image segmentation thanks to its simplicity and self-adaptation. After studying Otsu thresholding algorithm, an improved method is developed in this paper. By combining the between class distance and within class distance of object and background, a better segmentation is achieved with the new proposed thresholding method. The method proposed in this paper outperforms the present similar methods by using the proportion of object and background as weight values, to make the average variance of the object and the background segmented by their respective areas. The simulation results of standard testing images as well as weed images show the effectiveness of the proposed method.
出处
《应用科技》
CAS
2010年第2期52-54,60,共4页
Applied Science and Technology
关键词
OTSU法
图像分割
阈值化
方差
Otsu algorithm
image segmentation
thresholding
variance
