基于多尺度分割的遥感影像滨海湿地分类

基于多尺度的高分辨率遥感影像分类方法研究,可以为滨海湿地动态监测、规划保护提供更详尽的湿地分类信息和更快速的数据获取方法,对湿地保护具有重要意义。选取连云港青口河入海口处湿地为研究区,以高分辨率遥感影像WV-Ⅱ和航空遥感影像为数据源,利用多尺度分割方法将影像分割成不同层次的实体对象;在不同层次,以实体对象为单元,结合光谱、形状、纹理等不同影像特征,进行滨海湿地分类研究,结果表明:利用该方法分类后,研究区各种湿地类型都达到较高精度。基于多尺度分割的影像分类方法能充分利用各种影像特征完成湿地分类,有效地减少了遥感影像中的"椒盐"现象,提高了分类精度;选择适宜的分割尺度和分割参数是基于多尺度分割的遥感影像分类方法提高精度的前提。

Coastal Wetland Classification based on Multi-scale Image Segmentation Using High Spatial RS Images

Through the classification method study of high spatial resolution RS image based on multi\|scale image segmentation,the wetland information can be obtained in more detailed type and rapid way,and that is important for wetland protection.Taking the coastal wetland,located in Qingkou River estuary in Lianyungan,as study area,and the WV\|Ⅱhigh spatial resolution RS image and Arial image as test data,the images were divided into different level objects by multi\|scale segmentation,and then the wetland classification method was studied combing with spectral,shape,texture characteristics using the object as basic unit in different segmentation levels.The results showed that each wetland type in the study area had better classification precision by the method.Based on multi\|scale segmentation,various image characteristics could be fully used for the wetland classification.So,this method can obviously reduce the disturbance of salt\|and\|pepper noise in the classification results to get more accurate results.It is found that appropriate segmentation scale and parameter play an essential role in the process of classification based on multi\|scale segmentation.