摘要
提出了一种融合位置特征和形状特征的机场掩体目标检测方法。以形态学方法提取机场道路的骨架图,然后用道路网的拓扑关系提取掩体目标的位置特征——断头路,用轮廓序列矩检测断头路区域内构筑物形状特征,实现对机场掩体目标的融合检测。并以高分辨率遥感图像为例进行了飞机掩体目标的检测实验,结果表明,该方法能较好地检测机场掩体目标。
Combining the features of position and shape,the method of detecting the airport shelter in remote sensing images is proposed in this paper. Firstly,the Mathematics Morphology to extract the skeleton of airport road networks was introduced in preprocessing step. Then topological relationship of the net is employed to locate the End Road of shelter targets,which can indicate the position well. In succession,contour sequence moments are used as salience features to extract building shapes. On that basis, the fusion detection of shelters can be achieved. The experiment is carried out with high resolution satellite images of airports the results show that this method can recognize the shelters in airport area accurately.
出处
《遥感技术与应用》
CSCD
北大核心
2009年第3期379-384,I0005,共7页
Remote Sensing Technology and Application
关键词
断头路特征
轮廓序列矩
特征融合
目标检测
End road feature
Contour sequence moments
Feature fusion
Target detection
