一种基于双层协同结构改进的高光谱异常检测方法
By using a two-layer collaborative representation structure to pre-detect and purify the background of hyperspectral images, the problem of background dictionary being contaminated by anomalies is solved, thus improving the accuracy and performance of anomaly detection.
CN118154541BActive Publication Date: 2026-07-17XIDIAN UNIV
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- XIDIAN UNIV
- Filing Date
- 2024-03-11
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
Existing hyperspectral anomaly detection algorithms based on collaborative representation suffer from the problem of background dictionary being contaminated by anomalies, which affects detection performance.
Method used
A two-layer collaborative structure is adopted. First, the original hyperspectral image is pre-detected and the background is purified by the first layer collaborative representation. Then, the second layer collaborative representation is used for reconstruction to reduce the pollution of the background dictionary by outliers.
Benefits of technology
It effectively reduces the pollution of the background dictionary by outliers, improves the accuracy and performance of anomaly detection, and significantly improves the background suppression effect.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN118154541B_ABST
Abstract
一种基于双层协同结构改进的高光谱异常检测方法，包括如下步骤：使用第一层协同表示对原始高光谱图像H进行预检测；对预检测出的可能异常点进行均值向量填充，完成背景纯化；使用第二层协同表示对背景纯化后的高光谱图像进行重构，得到重构高光谱图像Y；计算原始高光谱图像H和重构高光谱图像Y的重构误差，得到异常检测结果。与现有技术相比，本发明设计了双层协同表示结构，通过第一层协同表示算法将大部分异常点预检出，并用邻域进行背景纯化，有效降低异常点对背景的污染，从而减弱异常点对算法性能的影响，并具有相对优异的检测效果。
Need to check novelty before this filing date? Find Prior Art