基于CNN算法的遥感影像分类技术在耕地保护中的应用
Application of remote sensing image classification technology in farmland protection based on CNN algorithm
赵晓燕1
作者信息
- 1. 山西省测绘地理信息院,山西 太原 030001
- 折叠
摘要
针对遥感影像在耕地保护中的重要应用,对基于卷积神经网络(CNN)的深度学习技术进行影像分类的研究展开了讨论.分析了CNN算法在遥感影像分类中的优势,如自动特征提取、高精度分类和端到端学习等.结果表明:该算法在耕地等地物分类上能取得较高的分类精度,为评估耕地分布、监测耕地变化等提供了有力的数据支持.最后,本文指出了CNN算法在遥感影像分类领域的潜在发展方向,如与其他算法融合、引入多元数据源等,以期进一步提高分类性能,推动相关技术在实际问题中的应用.
Abstract
Considering the important application of remote sensing images in arable land protection,the research on image classification based on convolutional neural networks(CNNs)deep learning technology is discussed.The advantages of CNN algorithms in remote sensing image classification are analyzed,such as automatic feature extraction,high classification accuracy,and end-to-end learning.Through experimental verification,the algorithm can achieve high classification accuracy for land cover types like arable land and provide data support for evaluating arable land distribution and monitoring arable land changes.Finally,the potential development directions of CNN algorithms in the field of remote sensing image classification,such as fusion with other algorithms,introduction of multi-source data,etc.are pointed out in this paper so as to further improve classification performance and promote the application of related technologies in practical problems.
关键词
卷积神经网络/深度学习/遥感影像/耕地保护/地物分类Key words
Convolutional Neural Network/deep learning/remote sensing image/farmland protection/classification引用本文复制引用
出版年
2024
NETL
NSTL
万方数据