NeonTreeEvaluation: RGB Dataset

The authors of the RGB part of the NeonTreeEvaluation Benchmark dataset aimed to address the challenges associated with broad-scale remote sensing for building forest inventories combining RGB, LiDAR, and hyperspectral sensor data from the USA National Ecological Observatory Network's Airborne Observation Platform. The dataset includes over 6,000 image-annotated crowns, 400 field-annotated crowns, and 3,000 canopy stem points, covering a diverse range of forest types. An R package is provided to standardize evaluation metrics, facilitating method comparisons. Additionally, the dataset offers more than 10,000 training crowns for optional use.

NeonTreeEvaluation 基准数据集的 RGB 部分作者旨在应对结合美国国家生态观测网络（USA National Ecological Observatory Network）航空观测平台获取的 RGB、激光雷达和超光谱传感器数据，在广域森林资源清查中遇到的挑战。该数据集包含超过 6,000 个图像标注的树冠、400 个现场标注的树冠以及 3,000 个树冠茎点，涵盖了多种森林类型。同时提供了一套 R 包，用于标准化评估指标，便于方法之间的比较。此外，该数据集还提供了超过 10,000 个可选的培训树冠用于训练。

搜集汇总

背景与挑战

背景概述

NeonTreeEvaluation RGB数据集旨在通过结合RGB、LiDAR和高光谱传感器数据解决大规模森林资源遥感调查的挑战，包含6000多个图像标注树冠、400个实地标注树冠和3000个树冠茎点数据，覆盖多种森林类型，并提供标准化评估工具和超过10000个训练样本。