基于单样例的遥感图像时空融合算法
Spatiotemporal Fusion Algorithm of Remote Sensing Image Based on Single Example
姚振稷 1欧阳恒1
作者信息
- 1. 贵州轻工职业技术学院,贵州 贵阳 550000
- 折叠
摘要
同时具有高空间分辨率和高时间分辨率的卫星图像是研究地表动态的重要数据,在监测土地变化与灾害以及估算农作物生长态势等方面得到广泛应用.然而,由于技术限制等原因,目前没有传感器可以同时拥有高精度的空间分辨率和频繁的重访周期.时空融合是一种将2类具有相似波段数量和带宽的遥感数据整合在一起的技术.基于此,提出一种基于单样例遥感图像的时空融合方法,使用神经网络来表示不同空间分辨率图像之间的映射关系,以提升模型预测目标图像的准确度.从实验结果来看,所提方法取得了良好的应用效果.
Abstract
Satellite images with high spatial resolution and high temporal resolution are important data for the study of land surface dynamics,and are widely used in monitoring land changes and disaster and estimating crop growth situation.However,due to technical limitations and other reasons,there are currently no sensors that can have both high-precision spatial resolution and frequent revisit cycles.Based on this,a spatio-temporal fusion method based on a single sample remote sensing image is proposed.Neural network is used to represent the mapping relationship between images with different spatial resolutions,so as to improve the accuracy of the model to predict the target image.From the experimental results,the proposed method has achieved good application effect.
关键词
遥感图像/时空融合/卷积神经网络(CNN)/精细-粗略图像对Key words
remote sensing image/space time fusion/Convolutional Neural Networks(CNN)/fine-rough image pair引用本文复制引用
基金项目
贵州轻工职业技术学院院级科研项目(23QY15)
出版年
2024
NETL
NSTL
万方数据