Research on Autonomous Recognition Technology of Foreign Objects on Airport Runway Based on Multimodal Data Fusion

Aiming at the problems of difficult recognition of weak targets and insufficient utilization of multimodal data in the detection of foreign objects (FOD) on airport runways, a multimodal autonomous recognition system integrating visible light, infrared, and hyperspectral is proposed. By using knowledge graph guided deep neural network architecture and combining non imaging multispectral material recognition technology, precise detection of FOD in complex scenes can be achieved. The experiment shows that in the testing of Zhengzhou Xinzheng International Airport, the false alarm rate of the system is reduced by 62% compared to traditional methods, and the detection efficiency is improved by three times, providing an innovative solution for the construction of smart airports.