基于GF-2影像的森林优势树种分类数据集

2017年针对湖北省十堰市九华山林场进行样地实地调查，共计511个样点，所有样点分为水体、硬阔、马尾松、针阔混交、茶园、柳杉、杉木和非林地这8种地物类型。本研究采用高分二号多光谱影像作为数据源，借助样点提取影像光谱特征、纹理特征、主成分变换的前2个主成分PCA1、PCA2和归一化植被指数NDVI、高程、坡度以及坡向总共18个特征，利用支持向量法、最近邻法和随机森林的分类方法对影像进行分类。所有数据包含文章图片、分类结果矢量集和分类结果的精度检验结果。

A field survey of sample plots was conducted at Jiuhua Mountain Forest Farm, Shiyan City, Hubei Province in 2017, with a total of 511 sample plots. All sample plots were categorized into eight land cover types: water body, hardwood forest, Pinus massoniana forest, coniferous-broadleaved mixed forest, tea plantation, Cryptomeria fortunei forest, Cunninghamia lanceolata forest, and non-forest land. Gaofen-2 (GF-2) multispectral imagery was used as the data source for this study. A total of 18 features were extracted from the imagery based on the sample plots, including spectral features, textural features, the first two principal components (PCA1 and PCA2) derived from principal component transformation, Normalized Difference Vegetation Index (NDVI), elevation, slope, and aspect. Image classification was implemented with three classification algorithms: Support Vector Machine (SVM), k-Nearest Neighbor (k-NN), and Random Forest (RF). All datasets include article figures, vector datasets of classification results, and accuracy assessment results of the classification.

搜集汇总

背景与挑战

背景概述

该数据集基于GF-2（高分二号）卫星影像，用于森林优势树种分类，覆盖湖北省十堰市九华山林场区域，包含2017年采集的511个样点，分为8种地物类型。数据集提取了光谱、纹理、植被指数和高程等18个特征，并应用支持向量机、最近邻法和随机森林等多种分类方法，提供分类结果矢量集和精度检验，适用于森林资源监测和遥感分类研究。