中国西北五省（宁夏、内蒙古、新疆、甘肃、青海）的亚米级光伏面板遥感数据集（NWC-PV）

本研究聚焦中国西北五省（宁夏、内蒙古、新疆、甘肃和青海），该区域生态环境脆弱但具备发展太阳能的天然优势。其丰富的土地资源、高海拔、长日照与强太阳辐射条件，加之多样的干旱半干旱地貌（涵盖沙漠、草原、盐碱地等）、平坦地势及较低的土地利用压力，显著降低了光伏项目的建设成本与难度。同时，降水少、光照充足的气候特征，使其成为大型地面光伏电站的理想选址。为全面评估 DPVM 模型的泛化能力，研究选取新疆伊犁、甘肃金昌、青海海南州、内蒙古乌拉特前旗，以及宁夏青铜峡和中卫市等典型地区作为研究区域。这些区域在地形、地貌与气候条件上差异显著，可有效验证模型在不同环境下的性能表现。This study focuses on five provinces in Northwest China: Ningxia, Inner Mongolia, Xinjiang, Gansu and Qinghai. This region has a fragile ecological environment but boasts inherent advantages for solar energy development. Its abundant land resources, high altitude, prolonged sunshine duration, intense solar radiation, coupled with diverse arid and semi-arid landforms including deserts, grasslands, saline-alkali soils and the like, flat terrain and low land use pressure, have significantly reduced the construction costs and technical barriers of photovoltaic (PV) projects. Meanwhile, the climatic characteristics of low precipitation and ample sunlight make this region an ideal site for large-scale ground-mounted photovoltaic power stations. To comprehensively evaluate the generalization ability of the DPVM model, this study selects typical research areas including Yili in Xinjiang, Jinchang in Gansu, Hainan Tibetan Autonomous Prefecture in Qinghai, Urad Front Banner in Inner Mongolia, as well as Qingtongxia and Zhongwei cities in Ningxia. These regions have significant differences in terrain, landforms and climatic conditions, which can effectively validate the model's performance across diverse environments.

搜集汇总

背景与挑战

背景概述

该数据集是一个专注于中国西北五省（宁夏、内蒙古、新疆、甘肃、青海）的亚米级光伏面板遥感数据集（NWC-PV），基于高分辨率卫星影像（分辨率0.27-0.6米）构建，包含4791个图像块和47623个光伏面板实例，手动标注并覆盖多样化的干旱和半干旱地貌，适用于光伏面板检测、分割和遥感智能解译任务，旨在评估模型在复杂环境下的泛化能力。