摘要
在对高空间分辨率遥感影像进行分类时,为解决不同地物其空间尺度不同的问题,采用多尺度分割的面向对象分类技术,提出采用"对象完整面积个数最多法"的分割方法,研究得出道路、农田、居民地、裸地、水域的最优分割尺度分别为20、30、50、80和100,采用成员函数法对分割后的影像进行分类,并将分类后的结果与基于像元的监督分类结果进行了对比.研究结果表明:使用"对象完整面积个数最多法"实现最优分割的面向对象分类技术的分类精度高于基于像元的分类精度.
When classifying the high-spatial resolution remote sensing images, there has the problem of different features have different spatial scales. In order to solve the problem, this study utilized the technology based on object-oriented methodology that uses the multi-scale segmentation to classify the high-spatial resolution remote sensing images, and came up with a method to calculate the maximum number of the complete area of objects. Based on this method, the optimal segmentation scale of road, farmland, house, vacant and water are 20, 30, 50, 80 and 100 respectively. Using this method of member function to classify the objects that have been segmented, and compared the results of the classification with the results of supervised classification based on pixel. The result of study indicates that the accuracy of classification based on the technology of object-oriented that uses the method of maximum number of the complete area of objects to realize the best segment is higher than the accuracy of the classification iust based on oixel.
出处
《辽宁工程技术大学学报（自然科学版）》
CAS
北大核心
2014年第1期56-61,共6页
Journal of Liaoning Technical University (Natural Science)
基金
江苏高校优势学科建设工程资助项目(PAPD)
海岛(礁)测绘技术国家地理信息局重点实验室基金资助项目(2011B07)
关键词
高空间分辨率
多尺度分割
面向对象
最优分割尺度
对象完整面积
成员函数法
监督分类
精度评定
high-spatial resolution
multi-scale segmentation
object-oriented
optimal segmentation scale
complete area of objects
member function
supervised classification
accuracy assessment
