基于积分图像的快速二维Otsu算法  

Fast two-dimensional Otsu algorithm based on integral image

作　　者：郎咸朋[1,2] 朱枫[3] 郝颖明[3] 欧锦军[3] 

机构地区：[1]中国科学院沈阳自动化研究所,沈阳110016 [2]中国科学院研究生院,北京100080 [3]中国科学院沈阳自动化研究所沈,沈阳110016

出　　处：《仪器仪表学报》2009年第1期39-43,共5页Chinese Journal of Scientific Instrument

基　　金：国家自然科学基金(60705011)资助项目

摘　　要：Otsu自适应阈值算法是图像分割的经典方法之一,在其基础上发展起来的二维Otsu阈值算法却因为计算复杂而制约了其应用。本文针对二维Otsu的耗时瓶颈问题,引入积分图像简化了二维直方图最佳阈值搜索过程的计算复杂度,从而大大缩短了算法处理时间。实验结果表明,本文提出的算法在分割效果上与原始算法完全相同,但计算时间远远小于后者,具有很好的实时性。Otsu adaptive threshold algorithm is one of the classic image segmentation methods. The application of the two-dimensional Otsu threshold algorithm based on original Otsu algorithm is restricted due to its computation complexity. To solve the time-consuming problem of the two-dimensional Otsu algorithm, in this paper, we improve the original algorithm with integral image and put forward a new fast algorithm. The segmentation result of our fast algorithm is the same as that of the original one, but the computation time is significantly reduced, which demonstrates well good real-time performance.

关 键 词：二维Otsu算法 阈值分割 积分图像 

分 类 号：TP2[自动化与计算机技术—检测技术与自动化装置]