结合结构和光谱特征的高分辨率影像分割方法

A High Resolution Image Segmentation Method by Combined Structural and Spectral Characteristics

影像分割是高分辨率遥感影像基于对象分析的先决和关键步骤。传统的遥感影像分割方法往往仅利用影像的光谱或结构信息。本文提出一种结合结构和光谱特征的分割方法。首先使用形态学运算提取结构信息,并与光谱信息结合,采用光谱角距离来衡量结构-光谱特征的相似性,进行区域生长获得初始分割结果,然后通过区域合并改善初始结果获得最终结果。研究中采用城市地区的高分辨率遥感影像,通过目视评价和定量评价并与现有的其他分割方法进行比较,验证了所提出方法的有效性。

Image segmentation is a key and prerequisite step for object-based analysis of very high resolution(VHR)images.Previous image segmentation methods usually use either structural or spectral information of the image alone.A novel image segmentation method for VHR multispectral images using combined structural and spectral information was proposed in this paper.The method can be summarized as follows.First,morphological derivative profile was calculated to quantify structural characteristics.Then the similarity of neighboring pixels was measured with an angle distance of pixels' structural and spectral characteristics.The initial segmentation results were achieved using a region growing procedure based on similarity of pixels in structural and spectral characteristics.The result was further refined by a region merging procedure to generate final segmentation result.The proposed method was evaluated by comparing with existing image segmentation methods through visual inspection and quantitative measures.Experimental results indicate that the proposed method achieves a better performance compared to the existing method.The proposed method applies well in high resolution multispectral image of urban areas.