下载中心
优秀审稿专家
优秀论文
相关链接
首页 > , Vol. , Issue () : -
摘要

获取精细湿地资源信息对国际湿地城市湿地修复保护、管理利用以及区域可持续发展至关重要,目前面向国际湿地城市的湿地精细分类研究较为缺乏,尤其是针对湿地中水体的详细类别划分较少。本研究以全球典型国际湿地城市常德市为案例研究区,基于Google Earth Engine(GEE)云计算平台和2020年Sentinel 1/2 时序遥感数据和地形数据,首先采用最小冗余最大相关算法和递归梯度提升树算法进行湿地分类特征集优选,进而构建集成基于像元的随机森林和面向对象的知识规则决策模型的城市湿地精细分类智能模型,实现常德市湿地资源精细分类。结果表明:(1)湿地分类特征优选前后特征数由63减少为16,总体精度下降0.9%,其中主要是旱期的水体指数、植被频率和建成区指数的特征重要性较大,特征优选可以减少特征数据信息冗余、提高分类效率;(2)常德市湿地精细分类结果包括河流、湖泊、水库、养殖池/坑塘、运河/水渠、泥滩地、草滩地和芦苇湿地8种湿地类型,总体精度达91.53%,Kappa系数为0.89,可以满足国际湿地城市精细湿地分类的需求;(3)常德市湿地主要分布在东部西洞庭湖平原区域,呈现出东多西少的空间格局。本研究利用多源遥感数据、GEE云计算平台、机器学习算法以及知识规则模型能够准确高效地提取国际湿地城市精细湿地信息,有望迁移至其他城市湿地制图应用,在服务国际湿地城市创建及优选、湿地资源修复保护与可持续开发利用等方面具有较大应用潜力。
Obtaining refined wetland resource information is very important for the restoration, protection, management and utilization of wetlands in international wetland cities and regional sustainable development. At present, there is a lack of refined wetlands classification research for international wetland cities, especially for the detailed classification of water bodies in wetlands. This study takes Changde City, a typical international wetland city in the world, as the case study area. Based on Google Earth engine (GEE) cloud computing platform and sentinel 1/2 time series remote sensing data and terrain data in 2020, the minimum redundancy maximum correlation algorithm and recursive gradient boosting tree algorithm were first used to optimize the wetland classification feature set, Then, an intelligent model of urban wetland refined classification integrating pixel-based random forest and object-oriented knowledge rule decision model is constructed to realize the refined classification of wetland resources in Changde City. The results show that: (1) the number of features before and after the optimization of wetland classification features is reduced from 63 to 16, and the overall accuracy is reduced by 0.9%. Among them, the characteristics of water index, vegetation frequency and built-up area index in the dry period are of great importance, and feature optimization can reduce the redundancy of feature data information and improve the classification efficiency; (2) The results of the fine classification of wetlands in Changde City include 8 wetland types: rivers, lakes, reservoirs, aquaculture ponds / pits, canals / canals, mudflats, grassland and reed wetlands, with an overall accuracy of 91.53% and a kappa coefficient of 0.89, which can meet the requirements of the fine wetland classification of International wetland cities, indicating that the precision of the urban wetland fine classification method framework is high, which can meet the needs of the international wetland city fine wetland classification; (3) Wetland in Changde city is mainly distributed in the eastern West Dongting Lake Plain, showing a spatial pattern of more in the East and less in the West. Using multi-source remote sensing data, gee cloud computing platform, machine learning algorithm and knowledge rule model, this study can accurately and efficiently extract the fine wetland information of international wetland cities, which is expected to be transferred to other urban wetland mapping applications, and has great application potential in serving the creation and optimization of international wetland cities, wetland resource restoration, protection and sustainable development and utilization.