论文部分内容阅读
针对目前面向对象方法在高分辨率遥感影像中提取绿地专题信息的特点,以及受到城市区域范围尺度限制因素的影响,本文提出一种利用高分辨率遥感影像提取城市区域范围绿地专题信息的方法,结合影像分割、影像面向对象模板统计和基于多阀值的模糊分类技术的优势,林地、密草地和疏草地专题信息被提取出来。研究中提出利用基于面向对象多尺度分割算法和最大似然监督分类法的提取结果进行精度评价,实验结果证明,在城市区域范围尺度上,该方法计算简便且实现了93.72%的高精度和总Kappa系数为0.8236的评价结果。
In view of the current object-oriented method to extract thematic information of green space in high-resolution remote sensing images and the influence of scale constraints in urban areas, this paper presents a method for extracting thematic information of urban green space using high-resolution remote sensing images, Combining with the advantages of image segmentation, image object-oriented template statistics and fuzzy classification based on multiple thresholds, thematic information of woodland, grassland and grassland were extracted. In the research, the accuracy evaluation based on object-oriented multi-scale segmentation and maximum likelihood classification is proposed. The experimental results show that this method is simple and easy to be implemented and achieves high accuracy of 93.72% Kappa coefficient of 0.8236 evaluation results.