自然资源高频次卫星遥感监测----光伏用地遥感监测产品

基于高频次卫星遥感影像，利用自动提取技术，深度学习地物特征，针对光伏用地，构建样本知识库，实现光伏用地的自动提取，形成了光伏用地专题监测产品。
光伏用地提取主要包括影像筛选、光伏用地图斑自动提取、图斑后处理等步骤。
根据光伏电站特点以及为了避免提取过多无效图斑，基于2米级分辨率卫星影像，针对全国/省/市/县1季度开展1次的光伏变化提取，提取面积超过400平方米（≥100个像素）的光伏图斑，主要提取光伏用地位置、分布、面积和变化等信息。

Leveraging high-frequency satellite remote sensing imagery, automatic extraction technologies, and deep learning-based ground feature extraction, we constructed a sample knowledge base targeting photovoltaic (PV) land use, realized automatic extraction of PV land parcels, and developed thematic monitoring products for PV land use.
The PV land extraction workflow mainly includes three steps: image screening, automatic extraction of PV land parcels, and post-processing of land parcels.
To align with the characteristics of photovoltaic power stations and avoid extracting excessive invalid parcels, we conduct PV change extraction once per quarter at the national, provincial, municipal, and county levels using 2-meter-resolution satellite imagery. We extract PV land parcels with an area exceeding 400 square meters (≥100 pixels), and mainly acquire information such as the location, distribution, area, and changes of PV land use.

搜集汇总

背景与挑战

背景概述

该数据集基于高频次卫星影像和深度学习技术，自动提取全国范围内面积超过400平方米的光伏用地图斑，提供季度更新的位置、分布及面积信息。提取过程涵盖影像筛选、图斑自动提取和后处理等步骤。