一种可有效分割小目标图像的阈值选取方法

摘要

目标检测与识别中常遇到目标与背景大小之比很小的小目标图像分割问题,此时现有的阈值分割方法几乎都失效.为此,提出了一种基于背景与目标的面积差和类内方差的小目标图像分割阈值选取方法.指出了目前图像阈值分割方法不能有效分割小目标图像这一缺陷,给出了基于背景与目标面积差和类内方差的一维直方图、二维直方图区域直分及更为有效的二维直方图区域斜分阈值选取公式,导出了相应二维斜分阈值选取的快速递推算法;最后在实验结果中给出了本文方法的图像阈值分割结果和运行时间,并与Otsu、最大熵及Fisher闻值选取快速方法进行了比较.结果表明:本文方法能准确地分割小目标图像,且运行时间短,抗噪性好.%Segmentation difficulties of small object images often occur in target detection and recognition,where the object is much smaller than its background. The existing threshold methods almost fail. Thus,a threshold selection method is proposed on the basis of the area difference between background and object and within-class variance. The threshold selection formulae according to one-dimensional histogram,two-dimensional histogram vertical segmentation and two-dimensional histogram oblique segmentation are given. A fast recursive algorithm of threshold selection in two-dimensional histogram oblique segmentation is derived. The segmented images and processing time of the proposed method are given in experiments.It is compared with some fast algorithms, such as Otsu, maximum entropy and Fisher threshold selection methods. The experiment results show that the proposed method can effectively segment the small object images and has better anti-noise property.