安徽农学通报 >
2025 , Vol. 31 >Issue 10: 109 - 113
DOI: https://doi.org/10.16377/j.cnki.issn1007-7731.2025.10.025
遥感影像自动配准技术及其在农业领域的应用
|
张亮霞(1980—),女,甘肃会宁人,高级工程师,从事水土保持工程设计工作。 |
Copy editor: 何艳
收稿日期: 2024-09-19
网络出版日期: 2025-06-03
Automatic registration technology of remote sensing images and its applications in agriculture
Received date: 2024-09-19
Online published: 2025-06-03
遥感影像自动配准技术是实现多源影像协同分析的关键基础,其通过建立空间映射关系解决影像间几何不一致的问题,为地表信息定量化处理提供支撑。本文总结了基于灰度和基于特征两种配准技术及其相关改进研究,列举介绍了自动配准技术在农业领域的应用,并对遥感影像配准技术进行了展望。当前主流的自动配准方法分为基于灰度和基于特征两类:前者通过优化相似性度量(如互相关系数、互信息)及改进参数求解算法(如蚁群算法、粒子群-Powell混合策略)提升配准精度与效率,但对大尺度变形及数据冗余问题仍存在局限性;后者通过提取并匹配几何特征(如SIFT、SURF及改进算子)实现高效配准,但存在误差累积、局部形变适应不足等缺陷。近年来,深度学习技术通过端到端特征学习与匹配优化,显著提升了配准鲁棒性,例如结合卷积神经网络的光流场校正和自监督学习方法。在农业领域应用方面,基于特征的改进算法(如SNS算法、自适应角点检测及双特征混合模型)成功应用于柑橘种植、水稻监测及丘陵耕地影像配准,提升了配准效率与精度。未来研究需进一步解决复杂地形导致的局部畸变、训练数据不足及非欧结构特征提取等难题,同时深化深度学习与多模态优化算法的融合应用,以推动遥感影像配准技术向更高精度与智能化发展。
张亮霞 , 谢丽萍 , 刘瑞龙 , 夏妍 . 遥感影像自动配准技术及其在农业领域的应用[J]. 安徽农学通报, 2025 , 31(10) : 109 -113 . DOI: 10.16377/j.cnki.issn1007-7731.2025.10.025
Automatic registration technology for remote sensing image is a critical foundation for multi-source image collaborative analysis. By establishing spatial mapping relationships, it addresses geometric inconsistencies between images and supports quantitative processing of surface information. This paper summarizes two mainstream registration technologies (gray-based and feature-based methods) and their related improvements, introduces applications of automatic registration in agriculture, and provides prospects for future development. Currently, gray-based methods enhance registration accuracy and efficiency by optimizing similarity metrics (e.g., cross-correlation, mutual information) and improving parameter-solving algorithms (e.g., ant colony optimization, particle swarm-Powell hybrid strategies). However, limitations persist in handling large-scale deformations and data redundancy. Feature-based methods achieve efficient registration through geometric feature extraction and matching (e.g., SIFT, SURF, and their variants), yet face challenges such as error accumulation and insufficient adaptation to local distortions. Recent advancements in deep learning, including end-to-end feature learning, convolutional neural network-based optical flow correction, and self-supervised methods, have significantly improved registration robustness. In agricultural applications, feature-based enhanced algorithms (e.g., SNS algorithm, adaptive corner detection, and dual-feature hybrid models) have been successfully applied to citrus plantation monitoring, rice growth assessment, and hilly farmland image registration, achieving efficiency gains and high precision. Future research should focus on addressing challenges such as local distortions caused by complex terrains, insufficient training data, and non-Euclidean structural feature extraction, while advancing the integration of deep learning with multimodal optimization algorithms to drive remote sensing image registration technology toward higher precision and intelligent development.
| 1 |
石正一. 基于模型一致分区和组合多特征的高分辨率影像配准研究[D]. 北京:中国科学院大学(中国科学院空天信息创新研究院),2022.
|
| 2 |
|
| 3 |
|
| 4 |
梅益文,回丙伟,郭鹏程. 合成孔径雷达图像与可见光图像配准方法综述[J]. 计算机应用,2024,44():242-249.
|
| 5 |
李星华,艾文浩,冯蕊涛,等.遥感影像深度学习配准方法综述[J].遥感学报,2023,27(2):267-284.
|
| 6 |
|
| 7 |
|
| 8 |
|
| 9 |
|
| 10 |
冯蕊涛,杜清运,罗恒,等. 基于光流校正的复杂地形区多时相遥感影像配准[J]. 遥感学报,2021,25(2):630-640.
|
| 11 |
|
| 12 |
|
| 13 |
|
| 14 |
尚明姝,王克朝.基于多尺度Harris角点检测的图像配准算法[J].电光与控制,2024,31(1):28-32.
|
| 15 |
|
| 16 |
BAY H,
|
| 17 |
|
| 18 |
|
| 19 |
|
| 20 |
|
| 21 |
|
| 22 |
陆健强,李旺枝,兰玉彬,等. 基于点特征检测的农业航空遥感图像配准算法[J]. 农业工程学报,2020,36(3):71-77.
|
| 23 |
周志艳,闫梦璐,陈盛德,等. Harris角点自适应检测的水稻低空遥感图像配准与拼接算法[J]. 农业工程学报,2015,31(14):186-193.
|
| 24 |
宋飞,杨扬,杨昆,等. 基于双特征的丘陵山区耕地低空遥感图像配准算法[J]. 北京航空航天大学学报,2018,44(9):1952-1963.
|
/
| 〈 |
|
〉 |