计算机工程与应用 ›› 2008, Vol. 44 ›› Issue (20): 178-180.DOI: 10.3778/j.issn.1002-8331.2008.20.054

• 图形、图像、模式识别 • 上一篇    下一篇

一种新的纹理分析算法在目标探测中的应用

姚 军,蒋晓瑜,杜登崇,刘燕群   

  1. 装甲兵工程学院 控制工程系,北京 100072
  • 收稿日期:2007-12-26 修回日期:2008-03-06 出版日期:2008-07-11 发布日期:2008-07-11
  • 通讯作者: 姚 军

Application of new texture analysis algorithm in target detection

YAO Jun,JIANG Xiao-yu,DU Deng-chong,LIU Yan-qun   

  1. Department of Control Engineering,the Academic of Armored Force Engineering,Beijing 100072,China
  • Received:2007-12-26 Revised:2008-03-06 Online:2008-07-11 Published:2008-07-11
  • Contact: YAO Jun

摘要: 针对在复杂图像中目标探测难度高、效果差这一问题,对图像的纹理特征进行研究,提出了一种新的基于纹理分析的目标探测方法。首先,利用小波边缘检测算法,获取图像的边缘信息。接着,采用水平模板与垂直模板对图像进行扫描操作,并根据目标区域的纹理特征,滤除大部分非目标区域,获得目标的粗定位结果。最后,结合边缘信息和粗定位结果,采用边缘分析的方式定位目标区域,实现目标的成功探测。通过目标靶图像实验,验证了该算法的准确性,稳定性和抗干扰性,在复杂背景条件下其识别率达到了99.3%。该算法通用性强、识别准确、效率高,是一种有效的目标探测方法。

关键词: 目标探测, 纹理分析, 边缘检测, 小波分析

Abstract: In order to solve the problem of target detection in complicated image,image texture features such as geometric figure,gray-value uniformity,edges and measurements etc are analyzed,and a new target detection approach based on texture analysis is presented.Firstly,image edge feature is obtained using the edge detection algorithm based on wavelet analysis.Then after the scanning operation on image using the horizontal template and the vertical,most of background areas in image are excluded,and the coarse range of target is acquired.Finally,the edge analysis algorithm is taken;the accurate range of target is obtained depending on target edge distribution and coarse range of target.The target is detected successfully.Experimental results indicate that it can satisfy the target detection requirements of precision,stabilization and strong anti-jamming.The preciseness rate of detection is up to 99.3%.It is an effective target detection approach and can be used in engineering widely.

Key words: target detection, texture analysis, edge detection, wavelet analysis