基于NSCT和半张量积在防跑道侵入检测技术的研究  

Research on Anti Runway Intrusion Detection Technology Based on NSCT and Semi Tensor Product

作　　者：杨昌其[1] 赵毅 YANG Chang-qi;ZHAO Yi(Civil Aviation Flight University of China,Guanghan 618000,China)

机构地区：[1]中国民用航空飞行学院,四川广汉618000

出　　处：《航空计算技术》2024年第1期41-45,共5页Aeronautical Computing Technique

基　　金：四川省科技厅重点研发项目资助(2023YFSY0025)。

摘　　要：为进一步提高跑道侵入告警系统的自动化水平,降低设施设备安装维护成本等问题,通过摄像设备获取机场场面的视频图像数据,采用稀疏表示对视频图像数据集进行检测处理;通过利用NSCT算法优化,结合半张量积,对样本数据集进行跟踪训练;根据单目摄像机测距建立坐标转换和测距模型,对机场场面航空器与跑道中线的距离进行准确测量,根据地面保护区设置合适的阈值实现跑道侵入告警。实验结果表明,优化后的模型图像重构质量提高了7.07%,平均处理时间降低了21.95%,模拟环境下跑道侵入告警准确率为95.2%。模型重构质量好,处理时间快、实时性好、准确率高,可以为预防跑道侵入事件的发生提供方法支持。In order to further improve the automation level of the runway intrusion warning system and reduce the installation and maintenance costs of facilities and equipment,this article uses camera equipment to obtain video image data of the airport scene,and uses sparse representation to detect and process the video image dataset;By utilizing NSCT algorithm optimization and combining semi tensor product,the sample dataset is trained for tracking;Establish coordinate conversion and ranging models based on monocular camera ranging,accurately measure the distance between airport surface aircraft and runway centerline,and set appropriate thresholds based on ground protection zones to achieve runway intrusion warning.The experimental results show that the optimized model image reconstruction quality has improved by 7.07%,the average processing time has decreased by 21.95%,and the accuracy of runway intrusion warning in simulated environments is 95.2%.This model has good reconstruction quality,fast processing time,good real time performance,and high accuracy,and can provide method support for preventing runway intrusion events.

关 键 词：跑道侵入 图像识别 NSCT 半张量积 单目测距 

分 类 号：V328[航空宇航科学与技术—人机与环境工程] V355