聚类与几何特征相结合的遥感图像多类人造目标检测算法

针对高分辨率光学遥感图像中人造目标的检测问 题,对传统的相位编组直线段提取算法和k-means 聚类算法了改进,提出了一种k-means聚类和几何特征 相结合的检测方法。根据自然物体和 人造目标在几何外形上表现出的不同特性,首先运用改进的相位编组算法对图像进行快速的 直线段提取； 然后以获取的直线段中心点为处理对象,运用k-means聚类算法 对提取的直线段进行密度聚类；最后,根 据每个类中的直线段数目和构成的几何基元情况,进行人造目标的判定。实验结果表明,本 文算法对遥感 图像中的房屋、汽车、船舰和飞机跑道等多类人造目标可达到90%以 上的检测精度,并具有较高的检测速度,对于一幅512pixel ×512pixel的图像,整个检测过程在100ms 以内。

A fusion algorithm of clustering and geometric features for multi-class man-made objects detection in remote sensing images

After improvement of the traditional phase-grouping baseline algorith m and k-means clustering algorithm,this paper proposes a new algorithm to detect man-made objects in hi gh-resolution remote sensing optical images,which is based on the difference of geometric shape between man -made objects and natural ones. First,an improved phase-grouping method is employed to extract lines in the image efficiently;Then,the center points of extracted-lines are treated as processing objects and the lines are c lassified into k classes by k-m eans clustering algorithm;Finally,man-made objects are picked up based on the numb er of the lines and the geometric elements in every class.The experimental results show that the method proposed in this paper is valid for detection of man-made objects with the detection precision above 90%,and so on,and the detection speed reaches 10frames persecond to the image with 512pixels×512pixels.