Remote Visualization and Detection of Foreign Object Debris in Aerospace Manufacturing using a Low-Cost Depth Camera

Abstract

Performing work within limited access environments such as aircraft wings is ergonomically hazardous. Often, mechanics are required to crawl through a waist-sized access hole to perform work in a wing. Within these spaces, it is difficult to see and easy to leave behind Foreign Object Debris (FOD). FOD that is left in the aircraft after work is completed can cause expensive damage. Mechanics working in limited access environments are at a high risk of developing Muskuloskeletal disorders [1], and recent FOD issues at Boeing have led to rejection of product deliveries by customers [2]. This research explores remote visualization and automated FOD detection methods as a solution to these problems using a custom-built robot. It was found that remote visualization works well only for debris that is large or distinct in color from the background and that automated FOD detection performs very well when statistics are used to inform selection of the detection threshold. For minimal false positive detections, it was found that at least 40 depth images should be used for each of the initial and final sets of depth images. Detection capabilities were tested in an aircraft wing section using a variety of common debris and it was found that collars as small as 0.27 inches in diameter can be detected, even without visible light.