Real-time autonomous indoor navigation and vision-based damage assessment of reinforced concrete structures using low-cost nano aerial vehicles

Abstract

This paper presents a novel autonomous inspection framework for low-cost nano aerial vehicles (NAVs) in indoor assessment scenarios. First, a novel autonomous navigation and obstacle avoidance scheme has been developed for effective navigation and image data collection. Second, the collected images have been live-streamed and processed by the real-time vision-based geotagging method and damage detection algorithms to localize damaged structural components. The proposed pipeline has been implemented on reinforced concrete (RC) structures. Parameter studies have been conducted on the influence of navigation formulation, take-off locations, and variation of floor plans. The results show that the proposed navigation algorithm is robust against the take-off location and it can achieve high coverage in all investigated floor plans. The results also demonstrate that the geotagging method and damage detection algorithms can achieve high localization accuracy for structural components, concrete spalling, and steel reinforcement exposure.