一种机器学习和遥感方法,用于准确地监测森林分区层面的植被覆盖变化
Wenjie Zhang1,2, Yingze Tian3, Xiaohui Su1,2
1School of Information Science & Technology, Beijing Forestry University, Beijing, China.
准确的森林分区 (FSC) 变化检测对于管理至关重要. 使用Sentinel-2数据的新粒子群优化-反向传播神经网络 (PSO-BPNN) 方法显示了精细森林监测的卓越性能.
科学领域:
- 林业林业 林业 林业 林业
- 遥感 遥感 遥感 遥感
- 机器学习 机器学习
背景情况:
- 准确检测森林分区 (FSC) 中的植被覆盖类型变化对于知情的森林管理至关重要.
- 尽管有各种森林变化检测算法,但FSC层面的微量检测得到了有限的关注.
研究的目的:
- 开发一种FSC级植被覆盖类型变化检测方法.
- 将来自Sentinel-2图像的光谱和纹理信息与森林管理规划和设计调查 (FMPI) 数据结合起来.
主要方法:
- 提取的光谱带,植被指数和纹理特征.
- 使用粒子群优化-反向传播神经网络 (PSO-BPNN) 构建了一个分类模型.
- 将PSO-BPNN与随机森林 (RF),支向量机 (SVM) 和传统反向传播神经网络 (BPNN) 模型进行比较.
主要成果:
- 在FSC尺度变化检测方面,PSO-BPNN始终优于其他算法.
- 实现了91%的整体准确性和0.86卡帕系数用于变更识别.
- 在验证数据集中成功检测到大约80%的改变FSC.
结论:
- 拟议的PSO-BPNN方法为微细森林变化监测提供了强大而准确的解决方案.
- 增强可持续森林资源管理的科学基础.
更多相关视频
09:16Methods of Soil Resampling to Monitor Changes in the Chemical Concentrations of Forest Soils
Published on: November 25, 2016
12:26Integrating Remote Sensing with Species Distribution Models; Mapping Tamarisk Invasions Using the Software for Assisted Habitat Modeling SAHM
Published on: October 11, 2016
相关概念视频
Levels of Organization
Molecules Are Composed of Atoms, and Biomolecules Are Assembled from Molecules:
The most basic levels include atoms, molecules, and biomolecules. Atoms, the smallest unit of ordinary matter, are composed of a nucleus and electrons. Molecules...
The Sense of Self: Reflected Self-Appraisal and Social Comparison
Model Approaches for Pharmacokinetic Data: Compartment Models
Two primary types of compartment models are recognized: mammillary and catenary. The more...
Compartment Models: Two-Compartment Model
Machines
A free-body diagram of the...
Global Climate Change
