低对比度图像中改进的二维熵阈值分割法  

Improved 2-D Entropy Thresholding Method in Low Contrast Image

作　　者：王栋[1] 朱明[1] 

机构地区：[1]中国科学院长春光学精密机械与物理研究所图像处理室,长春130033

出　　处：《仪器仪表学报》2004年第z3期355-358,362,共5页Chinese Journal of Scientific Instrument

摘　　要：为了实现对低对比度图像中目标的识别和提取,采用二维熵值理论,重新定义了二维直方图坐标的含义,利用二维最大熵作为判定准则,对图像进行阈值切割。利用低对比度图像的特点,采取了压缩循环边界条件方法并结合递推加速算法,对比传统的阈值切割方法,取得了良好的分割效果,同时也提高了原始二维最大熵的计算效率,增强了低对比度下图像目标识别率,有良好的应用前景。Two-dimensional (2-D) entropy method is proposed for target recognition and extraction in low contrast image, through redefining coordinate meanings of 2-D histogram and employing 2-D maximum entropy as criterion to threshold the kind of image. In virtue of characteristics of low contrast image, with compressing periodic boundary condition and combining recursive method, comparing to traditional thresholding theory, a well experimental result has been obtained, and calculational efficiency is improved as well. So there is a good application foreground for the method.

关 键 词：阈值分割 二维最大熵 低对比度 

分 类 号：TH7-55[机械工程—仪器科学与技术]