胡杨林重点监测矢量信息产品（2020）

本数据集基于条件约束下的R-CNN 深度学习的胡杨林提取方法，包括场景分割和目标识别两大步骤，根据胡杨的生长环境和空间分布格局，利用“整体到局部”的思想，首先通过简单的语义分割，将水域和陆域场景分开，然后在陆域和水域分别采取不同的算法实现胡杨的自动识别

This dataset is constructed based on the conditionally constrained R-CNN deep learning method for Populus euphratica forest extraction, which includes two core steps: scene segmentation and target recognition. Adopting the "whole-to-part" principle based on the growth environment and spatial distribution pattern of Populus euphratica, simple semantic segmentation is first employed to separate water and land scenes. Then, distinct algorithms are respectively deployed on land and water regions to achieve automatic recognition of Populus euphratica.

搜集汇总

背景与挑战

背景概述

该数据集为2020年胡杨林重点监测矢量信息产品，覆盖塔河干流下游英苏至阿拉干之间的林区及沙雅胡杨林塔河部分，采用R-CNN深度学习方法进行胡杨林信息提取，数据量179.62 MB，适用于地球科学领域的遥感监测研究。