雄安马蹄湾村航空高光谱遥感影像数据集

2017年10月，利用中国科学院上海技术物理研究所设计的可见光和近红外成像光谱仪，获得了雄安新区的航空高光谱遥感图像。光谱仪的总视场角为40.6 °，瞬时视场为0.25 mrad，有效推扫像素为2834，最大速高比为0.04。飞行高度2000 m，飞行区域覆盖雄县、安县、荣县、白洋淀。东西长48千米，南北宽27.5千米，总面积1320平方公里。在东西向方向发现了21条飞行线，而Matiwan村位于第10和第11条飞行线。飞行天气晴朗无云，能见度状况良好。数据分类前进行了辐射校正、几何校正以及图像镶嵌和裁剪。雄安新区 (马蹄湾村) 航空高光谱遥感图像的光谱范围为400-1000 nm，具有250带和0.5 m的空间分辨率。图像尺寸是3750 × 1580像素。此处标注的土地覆被类型为19种，主要为经济作物。采用随机森林分类法对雄安新区马蹄湾村的航空高光谱遥感影像进行了分类。利用了光谱的前三个主成分及其相应的八个空间纹理特征和植被指数，例如归一化差水指数和归一化差植被指数。总分类精度为97%，kappa系数为0.98。根据混淆矩阵，刺槐，梨树和Acer复合体的混淆是严重的。这种情况导致刺槐的分类准确性低。

In October 2017, airborne hyperspectral remote sensing images of the Xiong'an New Area were acquired using a visible and near-infrared imaging spectrometer designed by the Shanghai Institute of Technical Physics, Chinese Academy of Sciences. The spectrometer has a total field of view of 40.6°, an instantaneous field of view of 0.25 mrad, 2834 effective pushbroom pixels, and a maximum velocity-to-height ratio of 0.04. The flight altitude was 2000 m, and the flight area covered Xiong County, An County, Rong County and Baiyangdian Lake, with an east-west length of 48 km, north-south width of 27.5 km, and a total area of 1320 km². A total of 21 flight lines were identified along the east-west direction, and Matiwan Village was located on the 10th and 11th flight lines. The flight was conducted under clear, cloudless conditions with good visibility. Radiometric correction, geometric correction, image mosaicking and cropping were performed prior to data classification. The airborne hyperspectral remote sensing images of Xiong'an New Area (Matiwan Village) have a spectral range of 400–1000 nm, 250 spectral bands and a spatial resolution of 0.5 m, with an image size of 3750 × 1580 pixels. A total of 19 land cover types were annotated, mainly cash crops. Random forest classification was adopted to classify the airborne hyperspectral remote sensing images of Matiwan Village in Xiong'an New Area. The first three principal components of the spectrum, along with their corresponding eight spatial texture features and vegetation indices (e.g., Normalized Difference Water Index and Normalized Difference Vegetation Index) were used. The overall classification accuracy reached 97%, with a Kappa coefficient of 0.98. According to the confusion matrix, severe confusion occurred among Robinia pseudoacacia, pear trees and the Acer complex, which resulted in low classification accuracy for Robinia pseudoacacia.

搜集汇总

背景与挑战

背景概述

该数据集是2017年10月获取的雄安新区马蹄湾村航空高光谱遥感影像，光谱范围400-1000 nm，具有250个波段和0.5米空间分辨率，图像尺寸为3750×1580像素。数据经过辐射校正和几何校正处理，标注了19种土地覆被类型（主要为经济作物），并采用随机森林分类法达到97%的总分类精度，适用于高精度地物分类和遥感分析研究。