中文摘要：

针对机载激光雷达（light detection and ranging,LiDAR）数据与航空彩色影像的数据特点,提出了一种面向对象的多源数据融合分类方法。该方法根据影像光谱特性将航空影像分割为若干个同质区域,通过综合考察每个区域内LiDAR数据的滤波结果、空间离散度、高差值和航空影像光谱信息,判断各区域归属为哪一类。实验表明,该方法能够有效地分离房屋、树木和裸露地3种基本地物。

英文摘要：

We propose a new object-based fusing and classification approach of combination of airborne laser scanning data and aerial images. Firstly, a image segmentation, based on spliting and merging with pyramid data, can be processed on aerial image to get a set of disjoint regions. Then, to every region, filtering results, height difference information and sparse character of airborne laser scanning data and line segments, which is extracted from aerial image, can be investigated to classify this region to a kind class. Experiment results show that this method can divide point clouds to three fundamental objects efficiently and reliably.