一种基于样本质量自适应的SAR目标检测高效训练方法
By adopting a training method based on sample quality adaptation, the problem of low training efficiency in SAR target detection is solved, and fast and accurate target detection is achieved, meeting the needs of efficient training in emergency situations.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HARBIN INST OF TECH
- Filing Date
- 2025-08-20
- Publication Date
- 2026-07-17
AI Technical Summary
Existing SAR target detection methods are inefficient during training, making it difficult to meet the needs of rapid response in emergency situations. Furthermore, traditional training methods cannot effectively address the problems of unstable SAR image quality and significant differences in sample difficulty, and lack a systematic course learning framework.
We employ a training method based on adaptive sample quality. This method evaluates static priority through image quality and target complexity, dynamically estimates difficulty and adjusts quality, introduces a closed-loop feedback mechanism and anomaly loss rescheduling, and constructs a training sample sequence to achieve efficient training.
It improves the speed and accuracy of SAR target detection, meets the needs of time-sensitive applications, and enhances the performance gain of the model within a limited time.
Smart Images
Figure CN121121044B_ABST