Aircraft maintenance, repair and overhaul (MRO) faces many challenges, for example the line maintenance performed between flights is always time-critical. Furthermore, due to the size and shape of the aircraft, some areas are hard to reach and to be inspected, such as the crown or the vertical stabilizer. To improve the inspection process, a camera can be used to capture the aircraft surface and identify the damage. However, an RGB camera can only help to pick up the damage based on the colour difference and the results might be affected by the aircraft livery. Hyperspectral imaging (HSI) looks at reflectance to distinguish between different materials or chemicals despite being apparently the same colour. It has been widely applied in industries such as food safety, agriculture, pharmaceutics, etc. However, it has rarely been considered in the aerospace industry. This paper introduces a case study of inspecting damage on aircraft using an HSI camera. An HSI camera covering the range 400 - 1000 nm is used for this case study on metal and composite parts of aircraft. The aircraft parts, which came from decommissioned aircraft, are hit by simulated lightning strikes to recreate damage on aircraft. The damage can be identified by using HSI camera with range of 400 - 1000 nm. In the future, other ranges of wavelength may be used to study the damage since there might be more significant features of the reflectance spectrum to be used as markers. Thus, the damage ought to be easier identified since there will be more spectral information provided.
Airplanes are regularly inspected for any external damage between flights and during maintenance, especially when aircrew report possible lightning strike. Even today, the inspection is mainly done visually by authorized ground staff to look for evidence of possible damage, such as cracks and burn marks, etc. The process is not only inefficient and with poor traceability, but also troublesome when there is a need to inspect upper parts of the airframe. Approaches are available to automate the image acquisition, such as mounting a camera onto a moving robot and take multiple shots to cover the whole airplane. However, the acquired images still need to be screened thoroughly by technicians, which becomes an obstacle to automate the visual inspection process. The main reason for needing human intervention is the large number of distractions in the form of other features on the aircraft, not to mention the clutter produced by aircraft company livery. In this paper, novel methods to analyze the two-dimensional (2D) images and identify evidence of possible damage are presented. The methods are based on autocorrelation function (ACF) which is mostly used for fabric analysis. A pre-processing is firstly applied on the airframe image to remove background and enhance its quality. ACF is then implemented to look for abrupt changes which might be indications of damages. Lastly, a post-processing step is taken to filter out possible distractions. The proposed methods can work efficiently in various scenarios, which enables the possibility of automating the aircraft visual inspection process.