摘要:
蒙古高原广袤的干旱半干旱地区存在大量未经规划的自然道路,也称临时道路或越野公路。此类道路由车辆任意行驶碾压导致,会对地表生态及其稳定性造成影响,加剧干旱半干旱地区土地退化。由于自然道路数量众多、分布不规则且易随着区域发展而改变,因而高效精确获取这些信息是大范围草原地区的一个紧迫需求和难点。该文以蒙古国自然道路为主要提取目标,基于国产高分一号(GF-1)影像采用面向对象的方法进行研究区道路信息提取。首先对覆盖研究区的GF-1影像进行数据预处理,采用多尺度分割方法进行影像对象分割; 然后对需要提取的自然道路进行特征分析,通过计算光谱、几何特征参数及随机选取道路样本统计样本特征值,选出能表达自然道路特征的参数构建道路提取规则集; 进而结合多种方法的组合应用,最终通过最邻近分类法实现对道路的初步提取,采用阈值分类法等分类算法优化道路信息。结果表明,提取的蒙古国古尔班特斯苏木研究区自然道路长度为3 708.745 km,密度为0.129 km/km2,总体呈现东南密集,西、北部稀疏的分布特征,与本地区实际的煤矿企业生产和居民城镇生活情况相符。研究表明所提方法可以较完整地提取出研究区自然道路,可为蒙古高原等广大干旱半干旱地区自然道路提取提供方法借鉴。
Abstract:
Many unplanned natural roads, which are also known as temporary roads or unpaved roads, exist in the vast arid and semi-arid regions of the Mongolian Plateau. These natural roads, which were formed due to the arbitrary running of vehicles, will influence the surface ecology and its stability and aggravate land degradation in arid and semi-arid regions. They have a large quantity, are distributed irregularly, and tend to change with regional development. Therefore, there is an urgent need for the efficient and accurate information acquisition of natural roads in large-scale grassland regions, which is a challenge. Based on domestic high-resolution satellite (GF-1) images, this study extracted information on the natural roads in Mongolia using the object-oriented method. First, the data of GF-1 images covering the study area were preprocessed, and the image objects were segmented using the multiresolution segmentation method. Then, the characteristics of the natural roads were analyzed for information extraction. By calculating the parameters of spectral and geometric features and randomly selecting road samples to statistically analyze the characteristic values of samples, the parameters that could characterize the natural roads were selected to construct a set of rules for information extraction of roads. Finally, information on roads was extracted and optimized by combining multiple methods for classification, among which the nearest neighbor classification method was used for preliminary extraction while other classification algorithms such as threshold classification were used for optimization. Consequently, natural roads with a length of 3 708.745 km were extracted in the study area, with a density of 0.129 km/km2. This result shows that the natural roads in the study area are densely distributed in the southeast and sparsely distributed in the north and west overall. These distribution characteristics are consistent with the actual production of coal mine enterprises and the living of local residents in the study area. Therefore, the method proposed in this study can extract almost complete information about natural roads in the study area and thus can be used as a reference for the information extraction of natural roads in vast arid and semi-arid regions of the Mongolian Plateau.
下载: