| 基于人工智能深度学习的卫星影像分类研究 |
| |
| 引用本文： | 冷天熙,钱发斌,胡文萍. 基于人工智能深度学习的卫星影像分类研究[J]. 林业调查规划, 2021, 46(1): 1-4 |
| |
| 作者姓名： | 冷天熙 钱发斌 胡文萍 |
| |
| 作者单位： | 云南省林业调查规划院,云南 昆明 650051;云南省林业调查规划院,云南 昆明 650051;云南省林业调查规划院,云南 昆明 650051 |
| |
| 摘 要： | 以芒市2019年卫星影像及2019年林地一张图成果为研究对象,基于深度学习的卫星影像分类研究,构建森林资源分类识别模型,以提高森林资源监测能力.将裁剪后的芒市2019年卫星影像分有林地、灌木林地、未成林地及耕地、建设用地5个类别导入自定义的ResNet18模型进行深度学习,并对学习结果进行验证.实验结果显示,在模型训练...
|
| 关 键 词： | 卫星影像分类 人工智能 深度学习 模型训练 森林资源监测 |
Satellite Image Classification Based on Artificial Intelligence In-depth Learning |
| |
| LENG Tianxi,QIAN Fabin,HU Wenping. Satellite Image Classification Based on Artificial Intelligence In-depth Learning[J]. Forest Inventory and Planning, 2021, 46(1): 1-4 |
| |
| Authors: | LENG Tianxi QIAN Fabin HU Wenping |
| |
| Affiliation: | (Yunnan Institute of Forest Inventory and Planning,Kunming 650051,China) |
| |
| Abstract: | Taking the results of satellite images and one map data of forestland of Mangshi in 2019 as the research object,the forest resources classification and recognition model was constructed to improve the forest resources monitoring ability based on the in-depth learning satellite image classification research.The clipped images were divided into five categories:forest land,shrub land,uncultivated land and cultivated land,and construction land which were introduced into the self-defined ResNet18 model for in-depth learning,and the learning results were verified.The experimental results showed that in the process of model training,the loss value of the model decreased gradually with the increase of the number of iterations,and the more training samples,the higher the accuracy. |
| |
| Keywords: | satellite image classification artificial intelligence in-depth learning model training forest resources monitoring |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
