一种面向巷战中多目标精确识别的视觉感知方法、系统及装置
By constructing a multi-target detection network model and combining it with a meta-learning strategy and a multi-scale semantic segmentation network, the problem of rapid identification of unknown targets by unmanned equipment in urban warfare was solved, thereby improving the battlefield adaptability and combat effectiveness of unmanned equipment.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHANGSHA XIANGJI HAIDUN TECH CO LTD
- Filing Date
- 2023-12-19
- Publication Date
- 2026-07-17
AI Technical Summary
Existing multi-target recognition algorithms cannot quickly expand to include unknown targets, leading to misidentification and missed detection of enemy targets by unmanned equipment in urban warfare, thus limiting the application scope and battlefield role of unmanned equipment.
By employing a meta-learning strategy and object detection network model reconstruction method, a multi-object detection network model is constructed. This model utilizes feature extraction, feature fusion, and object prediction networks, combined with a multi-scale semantic segmentation network, to achieve rapid identification and accurate localization of unknown objects.
It enables unmanned equipment to quickly and accurately identify unknown targets in urban warfare, improving battlefield adaptability and combat effectiveness, reducing network training time, and enhancing the accuracy of target tracking and strike.
Smart Images
Figure CN117994558B_ABST