视觉驱动的航母舰载机自动挂弹关键技术研究

Weapon loading for the carrier-based aircrafts on the aircraft carrier deck is one of the most important factors that influence the performance of the aircraft carrier. Conventional weapon loading manners mainly rely on manual operation, which cannot meet the demand for sustained, rapid and safe weapon loading during long and tough confrontation. Automation and intelligence of weapon loading for the carrier-based aircrafts is the trend of the aircraft carrier development. This research proposes to adopt scene perception and analysis method based on machine vision，and explore new methods and new technologies that are suitable for automatic weapon loading. The research mainly includes: (1) novel methods that involve human vision constancy and compound features to enhance the images captured by the vision system of weapon loading vehicle; (2) novel methods that utilize “Adversarial- Transfer” learning and vision saliency to recognize the hanging shelves and hanging point of carrier-based aircraft and extract feature points; (3) a novel distributed camera system that adaptively support vision guidance for automatic weapon loading. Through the above research, we aim at realizing the active vision sensing and guiding for weapon loading vehicle and applying the key technologies to integrated verification prototype system, so as to provide key technical support for improving the intelligent level of carrier-based aircraft ammunition support of our country.

航母飞行甲板舰载机挂弹作业是影响航母作战性能的关键因素之一。常规挂弹方式主要依靠人工操作，难以满足长时间、强对抗情况下持续、快速、安全的挂弹作业需求。舰载机挂弹作业自动化和智能化是未来航母发展的必然趋势。本课题拟采用基于机器视觉的场景感知和分析技术，探索适用于视觉引导下挂弹车执行挂弹作业的新方法和新技术。研究内容主要包括：（1）基于人眼视觉恒常性及组合特征的挂弹车视觉感知能力增强；（2）基于“对抗-迁移”学习及视觉显著性的舰载机挂架挂点识别及特征提取；（3）基于分布式视觉系统的挂弹作业自适应动态视觉引导。通过上述研究，实现视觉引导下的挂弹车自动挂弹作业，并将相关核心技术在原型系统上进行集成验证，为我国航母舰载机弹药保障的智能化水平提升提供关键技术支撑。

航母飞行甲板舰载机挂弹作业是影响航母作战性能的关键因素之一。常规挂弹方式主要依靠人工操作，难以满足长时间、强对抗情况下持续、快速、安全的挂弹作业需求。舰载机挂弹作业自动化和智能化是未来航母发展的必然趋势。本课题拟采用基于机器视觉的场景感知和分析技术，探索适用于视觉引导下挂弹车执行挂弹作业的新方法和新技术。研究内容主要包括：（1）基于人眼视觉恒常性及组合特征的挂弹车视觉感知能力增强；（2）基于“对抗-迁移”学习及视觉显著性的舰载机挂架挂点识别及特征提取；（3）基于分布式视觉系统的挂弹作业自适应动态视觉引导。通过上述研究，实现视觉引导下的挂弹车自动挂弹作业，并将相关核心技术在原型系统上进行集成验证，为我国航母舰载机弹药保障的智能化水平提升提供关键技术支撑。本项目已经按照原定计划如期并超额完成，项目执行期间一共发表三类高质量学术论文10余篇，其中包括IEEE- Cybernetics、IEEE-TNNLS、IEEE-TMM、IEEE-TIM、CVPR、Pattern Recognition、Frontiers of Computer Science、中国科学-信息科学、自动化学报等，并申请相关发明专利2项，目前处于公开审查状态，搜集并标注飞机挂架数据集一个，有利于本方向继续开展研究。本项目的研究成果为基于视觉引导的飞机自动挂载作业提供了有力的理论与方法支撑，有利于提高挂载作业的智能化水平和效率，能更好地服务航母舰面保障作业。