双模态融合建模的红外可见光目标跟踪关联方法与系统
The infrared-visible light target tracking method based on dual-modal fusion modeling utilizes the fusion of multiple features from infrared and visible light images to solve the problems of mismatch and identity switching in complex scenes, thereby improving the reliability and stability of target association.
CN122156258BActive Publication Date: 2026-07-17HUNAN UNIV
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HUNAN UNIV
- Filing Date
- 2026-05-09
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
Existing dual-modal target tracking methods using infrared and visible light are prone to mismatches and identity switching in complex scenarios, leading to unstable associations.
Method used
A dual-modal fusion modeling method is adopted, which uses target detection from infrared and visible light images, Kalman filtering state extrapolation, gradient direction histogram descriptors, and Hungarian algorithm to calculate a comprehensive association score to achieve one-to-one matching.
Benefits of technology
It improves the reliability of target association and the stability of state updates in complex scenarios, and enhances the ability to identify identities and adapt to complex environments.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN122156258B_ABST
Abstract
本发明公开了一种双模态融合建模的红外可见光目标跟踪关联方法与系统，本发明方法包括将红外图像及其对应的可见光图像分别利用目标检测网络进行目标检测得到检测框集合；再利用卡尔曼滤波算法得到两种预测框集合；分别计算可见光图像的检测框集合中的检测框、预测框集合中的预测框之间的可见光运动相似度、红外梯度结构相似度以及双模态外观相似度三项指标；将三项指标进行加权融合得到综合关联分数，并基于匈牙利算法求得红外可见光目标跟踪关联的一对一匹配结果。本发明旨在解决现有红外与可见光双模态目标跟踪中关联依据单一、复杂场景下易发生误匹配和身份切换的问题，提高复杂场景下目标关联的可靠性和后续状态更新的稳定性。
Need to check novelty before this filing date? Find Prior Art