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

海湾处于水陆交互的区域,生态系统较为脆弱,资源环境极易受损。大尺度、长时序和高精度的土地利用/覆被变化(land use/cover change,LUCC)制图是海湾区域国土空间规划和环境保护的基础。现有的制图方法多是针对原始遥感图像,难以充分挖掘和联合利用特征空间和变换空间蕴含的信息潜力,导致传统方法在地表异质性较高的海湾区域应用效果较差。本文面向杭州湾区域,基于Landsat长时间序列卫星影像和谷歌地球引擎(Google Earth Engine,GEE),提出了融合遥感指数和主成分分量的随机森林遥感图像分类方法,实现了1985-2020年(5年时间间隔)的LUCC制图及时空格局分析。结果表明:(1)融合遥感指数和主成分分量的随机森林算法能够准确提取杭州湾LUCC信息,8个时相的平均总体精度(overall accuracy,OA)和Kappa系数分别为92.83%和0.9108。(2)研究期间内,建设用地(278.26 km2至2984.76 km2,年均增加77.33 km2)、水体(509.32 km2至680.21 km2,年均增加4.88 km2)、裸地(768.99 km2增长至1078.13 km2,年均增加8.83 km2)的面积呈现增加趋势,而林地(2159.49 km2至1881.52 km2,年均减少7.94 km2)、耕地(6998.45 km2至4800.59 km2,年均减少62.80 km2)、滩涂(181.65 km2至161.50 km2,年均减少0.58 km2)的面积呈现减少趋势。(3)研究期间内,耕地是最主要的转出源,总面积占比由64.23%减少至41.43%,耕地面积转出以建设用地(2268.05 km2)与裸地(630.20 km2)为主;耕地面积转入以水体(376.22 km2)与林地(352.22 km2)为主。本研究为杭州湾区域土地资源的科学管理提供数据支持,所得LUCC数据集对区域可持续发展具有重要意义。
The bay is in the area of land and water interaction, the ecosystem is relatively fragile, and the resources and environment are easily damaged. Large-scale, long-time series and high-precision land use/cover change (LUCC) mapping is the basis for territorial spatial planning and environmental protection in the bay region. Most of the existing mapping methods are aimed at the original remote sensing images, and it is difficult to fully tap and jointly utilize the information potential contained in the feature space and the transformation space, resulting in poor application effect of traditional methods in the bay area with high surface heterogeneity. Based on Landsat long-term satellite images and Google Earth Engine (GEE) for the Hangzhou Bay area, this paper proposes a random forest remote sensing image classification method that integrates index and principal component components, and implements the analysis of the spatio-temporal pattern of LUCC mapping from 1985 to 2020 (5-year interval). the result shows: (1) The random forest algorithm integrating index and principal components can accurately extract Hangzhou Bay LUCC information, and the average overall accuracy (OA) and Kappa coefficient of the eight-time phases are 92.83% and 0.9108, respectively. (2) During the study period, the area of construction land (278.26 km2 to 2984.76 km2, an average annual increase of 77.33 km2), water area (509.32 km2 to 680.21 km2, an average annual increase of 4.88 km2), and bare land (768.99 km2 to 1078.13 km2, an average annual increase of 8.83 km2) showed an increasing trend, while wood land (2159.49 km2 to 1881.52 km2, an annual average decrease of 7.94 km2), cultivated field (6998.45 km2 to 4800.59 km2, an annual average decrease of 62.80 km2) and tidal-flat area (181.65 km2 to 161.50 km2, an annual average decrease of 0.58 km2) showed a decreasing trend. (3) During the study period, cultivated field was the main transfer source, and the proportion of the total area decreased from 64.23% to 41.43%. The transfer-out of cultivated field area is mainly construction land (2268.05 km2) and bare land (630.20 km2); the transfer-in of cultivated field area is mainly water body (376.22 km2) and forest land (352.22 km2). This study provides data support for the scientific management of land resources in the Hangzhou Bay region, and the obtained LUCC dataset is of great significance to the region’s sustainable development.