摘要
针对星载光学遥感图像城市街区复杂背景问题,提出一种车辆目标自动检测方法。首先,利用场景中植被背景的物理属性,通过多光谱波段抑制场景中的植被背景,然后,在分析城市街区地物形态反射率特性的基础上,利用全色波段并结合二值形态学方法抑制场景中的建筑物,最后,引入著名的RX算法对抑制后的图像进行车辆目标检测。将文中提出的方法应用于实际Quickbird影像的车辆目标检测,结果表明所提出的方法具有鲁棒性强,执行效率高,不需要人工辅助等方面的特点,可用于城市街区车辆目标的自动检测。
It is difficult to detect vehicles in city area by using paceborne optical remote sensing images, because the background in city area is too complex. In this paper, an automatic vehicle detection method was proposed to address the issue by using background segmentation method. Firstly, the physical property of the vegetation was analyzed and used to suppress the vegetation background of a scene by using the multi- spectral information of the scene. Next, the reflectance characteristics of city area cover types were analyzed. Based on the reflectance characteristics of building roofs and roads, the building background in the scene was removed by employing the binary morphological method on the panchromatic band image. Finally, the famous RX algorithm was introduced to detect the vehicles on the vegetation and building background suppressed image. The proposed method is applied to the actual Quickbird image for vehicle target detection. The results show that the proposed method has strong robustness, high efficiency, and automatic characteristics, and can be used for vehicle detection in city area.
出处
《红外与激光工程》
EI
CSCD
北大核心
2014年第11期3751-3755,共5页
Infrared and Laser Engineering
基金
国家自然科学基金(61301290)
关键词
车辆检测
光学遥感图像
城市街区
形态学
vehicle detection
optical remote sensing images
city area
morphological
作者简介
李昭慧（1979-），女，博士生，主要从事光信息和光电图像处理方面的研究。E-mail：zh79100l@163．com
导师简介：张建奇（1960-），男，教授，博士，主要从事成像系统性能评估及光电景象生成技术方面的研究。E-mail：jqzhang@mail．xidian．edu．cn
