北京测绘2025，Vol.39Issue(4)：412-417,6.DOI:10.19580/j.cnki.1007-3000.2025.04.002
基于动态特征融合的遥感影像目标识别
Remote sensing image target recognition based on dynamic feature fusion
刘永禄1
作者信息
- 1. 三和数码测绘地理信息技术有限公司,甘肃 天水 741000
- 折叠
摘要
Abstract
Remote sensing images often have complex and changing backgrounds,filled with significant amounts of noise and interference.Image targets typically exhibit dynamic characteristics.To accurately identify specific targets within the image,this paper proposed a method for precise target recognition in remote sensing images based on dynamic feature fusion.The method combined layer clustering algorithms with graph segmentation to segment remote sensing images.The segmented images were then input into a dynamic feature fusion network.The dynamic feature extraction module of the network was used to obtain the dynamic features from the image.Dynamic feature fusion was completed through feature fusion weights and dynamic label matching strategies.The dynamic label matching method was used to achieve precise target recognition in remote sensing images.Test results show that this method provides good image segmentation results,with distortion and under-segmentation rates both below 2%.The integrity of the target images obtained through segmentation is high,and the method can achieve precise recognition of specific targets in remote sensing images.关键词
动态特征融合/遥感影像/目标精准识别/动态特征/图像分割/动态标签匹配Key words
dynamic feature fusion/remote sensing image/precise target recognition/dynamic feature/image segmentation/dynamic label matching分类
信息技术与安全科学引用本文复制引用
刘永禄..基于动态特征融合的遥感影像目标识别[J].北京测绘,2025,39(4):412-417,6.
