杨凌农业示范区经济作物种植结构多源多时相遥感数据集

卫星遥感能够及时获取大尺度下的地物分布，为经济作物种植结构的信息获取提供良好的数据与技术支撑。本数据集以杨凌农业示范区为研究区域，由遥感数据、地面真值数据、杨凌边界以及分类结果四部分组成。遥感数据由2021年4月到9月的哨兵2号、高分1号（包括高分1号C星）、高分2号以及高分6号等卫星数据构成，经过辐射校正、大气校正、正射校正、图像融合以及影像配准等遥感影像处理。通过实地调查、Google Earth目视解译、小区域的无人机近地遥感等多种方式，建立地面真实分布验证区。在质量控制方面，遥感数据整体含云量很少、颜色均匀、空间分辨率为2m；地面真值图通过实地调查进行绘制，真实可靠。数据集采用随机森林算法验证，总体分类精度为86.17%。本数据集可为相关算法在经济作物种植结构获取方面的研究及应用提供训练样本，也可为杨凌示范区的土地利用分类及变化、农作物长势监测等方面提供数据支撑。

Satellite remote sensing can timely obtain ground feature distributions at large scales, providing excellent data and technical support for acquiring information on economic crop planting structures. This dataset takes the Yangling Agricultural Demonstration Zone as the study area, and consists of four parts: remote sensing data, ground truth data, Yangling boundary, and classification results. The remote sensing data is composed of satellite data from Sentinel-2, GF-1 (including GF-1 C satellite), GF-2, GF-6 and other satellites from April to September 2021, and has undergone remote sensing image processing procedures including radiometric correction, atmospheric correction, orthorectification, image fusion and image registration. Ground truth distribution validation areas were established through multiple methods including field surveys, visual interpretation via Google Earth, and unmanned aerial vehicle (UAV) close-range remote sensing in small-scale regions. In terms of quality control, the remote sensing data generally has low cloud cover, uniform color, and a spatial resolution of 2 m; the ground truth map was drawn based on field surveys, which is authentic and reliable. The dataset was validated using the random forest algorithm, with an overall classification accuracy of 86.17%. This dataset can provide training samples for the research and application of relevant algorithms in acquiring economic crop planting structures, and also offer data support for land use classification and change, crop growth monitoring and other aspects in the Yangling Agricultural Demonstration Zone.

搜集汇总

背景与挑战

背景概述

该数据集聚焦杨凌农业示范区的经济作物种植结构，整合了多源多时相遥感数据，包括Sentinel-2和高分系列卫星数据，时间覆盖2021年4月至9月。数据集经过严格处理和质量控制，空间分辨率为2米，分类精度达86.17%，适用于经济作物种植结构研究、土地利用分类及作物生长监测等应用。