论文部分内容阅读
传统光谱分类法的局限性促使了遥感“图谱耦合”认知理论的发展,使其更加注重了空间信息的应用。然而,已有的分类方法虽也融入了空间形态、空间关系的应用,在精度上有一定的提高,但在空间规律定量描述、地物实际分布边界跟踪等方面仍存在不足。本文发展了一种空间邻接支持下的遥感影像分类方法:通过基准地物的精确提取进而搜索与其邻接的目标地物,对邻接范围内的地类混淆以及非邻接范围内的目标类误分一并进行修正,并以近海地物分类为例进行试验,获得了更为精确、合理的分类结果,也为后续逐步精确地提取各地物提供了一种便捷有效的途径。
The limitations of traditional spectral classification promote the development of remote sensing “coupling theory ” theory and make it pay more attention to the application of spatial information. However, although the existing classification methods are integrated into the application of spatial morphology and spatial relations, they have some improvements in accuracy. However, there are still some shortcomings in the quantitative description of spatial laws and the boundary tracking of the actual distribution of the features. In this paper, we develop a remote sensing image classification method based on spatial adjacency support: by accurately extracting the reference features, we can search the adjacent objects, confuse the adjacent objects and the non-adjacent objects And make some corrections. Taking the classification of offshore objects as an example, we have obtained more accurate and reasonable classification results and provided a convenient and effective way for the subsequent step-by-step accurate extraction of various objects.