基于单类分类的航空遥感影像变化检测  

Change Detection of Aerial Remote Sensing Images Based on One-class Classification

作　　者：薄树奎[1] 荆永菊[2] BO Shukui;JING Yongju(School of Intelligent Engineering,Zhengzhou University of Aeronautics,Zhengzhou 450046,China;Library,Zhengzhou University of Aeronautics,Zhengzhou 450046,China)

机构地区：[1]郑州航空工业管理学院智能工程学院,河南郑州450046 [2]郑州航空工业管理学院图书馆,河南郑州450046

出　　处：《郑州航空工业管理学院学报》2023年第2期74-78,共5页Journal of Zhengzhou University of Aeronautics

基　　金：国家自然科学基金项目(U1904119)。

摘　　要：航空遥感影像的变化检测在城市规划等领域发挥着重要作用。文章提出一种基于单类分类的变化检测方法,通过对不同时相的航空遥感影像进行单类分类,提取针对该兴趣类别的变化信息。该方法仅需要兴趣类别的训练样本,由训练样本确定一个距离阈值,实现单类分类和变化检测。虽然简化了参数设置和样本训练,但实验结果表明,该方法与最近邻方法具有相近的变化检测精度。Change detection of aerial remote sensing images is playing a crucial role in urban planning and other fields.This paper proposes a one-class classification-based change detection for aerial remote sensing images.The change information for the interest class is extracted by implementing one-class classification for aerial remote sensing images.The method only needs the training samples of the interest category,and a distance threshold is determined by the training samples to achieve one-class classification and change detection.Although the parameter settings are simple,the experimental results show that the method has similar change detection accuracy to the nearest neighbor method.

关 键 词：变化检测 单类分类 航空影像 遥感影像 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]