基于纹理特征与数学形态学的高分辨率影像城市道路提取

从遥感影像上提取道路面临如下问题:根据光谱特征分类的图像存在建筑物与道路混淆,受交通工具、行道树等影响易产生道路中的孔洞与道路边缘的毛刺等。该文利用南京市IKONOS高分辨率影像丰富的空间纹理信息,将纹理信息加入原始波段参与分类,可以解决建筑物与道路的混淆问题。另外,道路具有丰富的形态特征,应用数学形态学的相关算法和形状指数,可以很好地去除道路提取过程中产生的各种噪声。

City Road Extraction from High Resolution Remote Sensing Imagery Based on Texture and Mathematical Morphology

Extracting road feature from satellite data can provide immediate,accurate and effective data for traffic geographic information database.It can greatly improve the timeliness of electronic maps,and reduce the cost of electronic map.However,many problems are still to be solved,such as the confusion of existence buildings and roads on the classification images based on the spectral characteristics,and the holes and spurs due to the existence of transports and standing trees.IKONOS high-resolution images have rich texture information,so the texture information was added into the original bands to help the classification in this paper.The approach improved the effect of the classification.On the other hand,the mathematical morphology and shape parameters were used in the classification.According to the result of experimentation,the proposed approach can extract the main roads in the image effectively.