基于Planet遥感数据的沿海滩涂提取和细分类型识别研究
Research on Extraction and Subdivision Type Identification of Coastal Mudflat Based on Planet Remote Sensing Data
刘连胜 1温进杰 2李慧珊 1岳文1
作者信息
- 1. 广东省土地调查规划院,广东 广州 510075;自然资源部陆表系统与人地关系重点实验室,广东 广州 510075
- 2. 华南农业大学 资源环境学院,广东 广州 510642
- 折叠
摘要
根据高时间分辨率Planet遥感卫星每天过境时间筛选出 5 个研究区近 3 年最高潮位和最低潮位对应日期的遥感数据,采用U-Net深度学习算法识别高潮位和低潮位水陆分界线,提取每个研究区的滩涂范围,并通过目视解译识别沿海滩涂细分类型.5 个研究区的结果表明,高时间分辨率的Planet遥感数据可以快速通过最高潮位和最低潮位的水陆分界线提取沿海滩涂范围,并识别沿海滩涂自然资源细分类型,解决了目前沿海滩涂自然资源登记细分类型单元划分的难题,为自然资源资产清查中沿海滩涂实物量量测和沿海滩涂自然资源登记提供新的技术支持.
Abstract
Based on the daily overpass times of high temporal resolution Planet remote sensing satellite,this paper selects imagery cor-responding to the highest and lowest tides over the past three years for five study areas.The U-Net deep learning algorithm is em-ployed to identify the land-water boundary at high and low tide and extract the mudflat extent in each study area.Coastal mudflat sub-division types are identified through visual interpretation.The results from the five study areas demonstrate that high temporal resolu-tion Planet remote sensing data can be used to rapidly extract coastal mudflat extents by the land-water boundaries at the highest and lowest tides and identify subdivision types of coastal mudflat natural resources.This approach addresses current challenges in subdivi-ding coastal mudflat natural resource units for registration and provides new technical support for physical measurement of coastal mud-flats and natural resource registration in the inventory of natural resource assets.
关键词
沿海滩涂/细分类型/U-Net模型/Planet影像/高时间分辨率Key words
coastal mudflat/subdivision type/U-Net model/planet imagery/high temporal resolution引用本文复制引用
出版年
2024
NETL
NSTL
万方数据