基于局部灰度分析的红外图像轮廓跟踪算法  

Contour Tacking for Infrared Image Based on the Characteristic Analysis of the Local Gray Level

作　　者：夏涛[1] 黄士科[2] 陈海清[1] 

机构地区：[1]华中科技大学光电子工程系,湖北武汉430074 [2]中国空空导弹研究院,河南洛阳471009

出　　处：《激光与红外》2006年第2期151-154,共4页Laser & Infrared

基　　金：国家重点预研项目(413010702)。

摘　　要：根据飞行目标红外图像特点,提出了用最大距离法用于局部阈值分割,用链码法来做改进的轮廓跟踪算法。该算法把消除颗粒噪声、采样离散化误差作为边缘平滑的重点;通过对图像局部灰度特性的分析解决了轮廓跟踪过程中出现的断边、分支问题。经验证,采用该算法可得到清晰、光滑的轮廓图像,并且具有较快的运算速度,可实现实时处理,实时显示。In accordance with the characteristics of airborne targets in IR image, a modified boundary tracking algorithm is proposed. It includes two stages：local-threshold segmentation based on the maximum distance method and boundary tracking based on chain-coding. The contour smoothing processing focused on the common zigzag and blurred effect occurred in conventional segmented images. The problems of fragments and furcations were also solved by the characteristic analysis of the local gray level. The experimental results show that the algorithm keeps the contours of segmented image sharp and smooth. With fast processing speed, the algorithm is possible to achieve real - time processing and real-time display.

关 键 词：图像处理 轮廓跟踪 阈值分割 边缘平滑 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]