应用图像方向和宽度谱检测机场跑道  

Detection of airport runway based on the orientation and width spectrums of images

作　　者：倪维平[1] 严卫东[1] 吴俊政[1] 张晗[1] 芦颖[1] 郑刚[1] 

机构地区：[1]西北核技术研究所,陕西西安710024

出　　处：《红外与激光工程》2014年第11期3655-3662,共8页Infrared and Laser Engineering

基　　金：国防预研基金;国家863计划

摘　　要：为有效解决成像制导中的机场跑道检测问题,提出了一种基于方向和宽度谱分析的处理方法。首先归纳了机场跑道形态和灰度特性,进而根据局部灰度特征进行跑道区域预分割,然后以Sobel算子提取边缘信息,再利用局部PCA计算边缘方向,并对各方向值出现频次进行累积,形成方向谱。通过分析方向谱的显著峰信息得到跑道的主方向,在此基础上,确定正交宽度谱的扫描计算模式。结合宽度谱显著峰信息的分析和跑道主方向进行跑道区域的指示和几何宽度值的计算。对机场跑道图像的处理结果显示,文中方法具有较高的检测精度,同时对跑道形状畸变显示出良好的鲁棒性。For the detection of airport runway in imaging guidance, a method based on the analysis of orientation and width spectrums was proposed. Firstly, the characteristics of structure and gray of runway in infrared images were analyzed, and then the runway regions were pre﹣segmented based on the local gray properties. Secondly, sobel operators were used to detect the edges, based on which the orientation spectrum was constructed with the edge orientations calculated with local PCA method. According to the significant peak of orientation spectrum, the main direction of runway was obtained, as well as the mode of the projection cross width spectrum construction. Thirdly, the width spectrum was calculated, from which the projection width could be obtained. With the main orientation and projection width, the geometric width of runway was then calculated, which fulfilled the detection of the runway in infrared image. Experimental results with the infrared images indicate that the method proposed has good performance for runway detection, not only with the high detection rate, but also with nice robustness for geometric distortion.

关 键 词：成像制导 局部PCA 机场跑道 方向谱 宽度谱 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]