Luke Connolly

ORCID ID: 0000-0003-2994-4710

Research Project Title: Development of an Unmanned Aircraft System Prototype for General Visual Inspection purposes utilising Machine Learning Models for flight path control and defect acquisition and analysis

Supervisors/s: Edmond Tobin, Diarmuid O’Gorman, James Garland

Project Funding: Irish Research Council, Project ID: GOIPG/2022/1015



  • Biography
  • Research Project Description
  • Publications and Outputs


I completed my BEng (Honours) in Aerospace Engineering at Institute of Technology Carlow (now SETU, Carlow Campus). I then pursued a master’s by research project focused on visual inspections of aircraft utilising an Unmanned Aircraft System (UAS). From this I decided to continue my research to PhD level, by applying machine learning and spatial mapping techniques to the system to enhance defect detection and flight path assistance.

I have a keen interest in maths and work as a maths tutor here at the University. I focus primarily on helping engineering students grasp the fundamentals of maths and apply these techniques to areas such as calculus and algebra.

Research Project Description

General Visual Inspections (GVIs) are critical for ensuring the safety and airworthiness of aircraft. They are performed every day, before, and after flights. However, these inspections can be time-consuming and pose safety risks to ground personnel who must climb ladders or scaffolding to inspect upper surfaces. To address these challenges, this project proposes the use of an UAS equipped with a stereo-vision camera to assist with flight path control and guidance. Furthermore, it can also improve upon the reliability as with ground personnel, they are prone to human error and complacency.

The stereo-vision camera onboard the UAS allows the system to perceive its environment in 3D, as humans do. It does this as it has two image sensors separated by a baseline. As the UAS navigates around the aircraft, it will record a video that has metadata of XYZ coordinates through translation and rotation values recorded from an Inertial Measurement Unit (IMU). This video will be disseminated for defects utilising a machine-learning model that is trained on images of defects and non-defects. Once this model determines which parts of the video have defects, it will use the metadata from the IMU to determine where that defect is on the aircraft, notifying the operator of where maintenance must occur.

Publications and Outputs

  • Connolly, L., O’Gorman, D., & Tobin, E. (2022). Design and Development of a low-cost Inspection UAS prototype for Visual Inspection of Aircraft. Transportation Research Procedia, 59, 85–94.

This conference paper discusses the design and development of an Unmanned Aircraft System (UAS) specifically designed for visual inspection of aircraft. The paper delves into various aspects of designing such a UAS, including the selection of appropriate motors, flight controller, and camera system. To ensure optimal performance, the paper outlines the weight and current draw analysis that was performed to achieve an efficient thrust to weight ratio, as well as sufficient flight time to perform the visual inspection.

  • Spatial Mapping of light aircraft with stereo-vision camera for use on Unmanned Aircraft System for defect localisation ( in IEEE Xplore)

This conference paper presents a novel approach to spatially mapping aircraft in a hangar environment using stereo-vision camera technology. The paper outlines how this 3D mapping capability can be incorporated into a UAS to enhance visual inspection purposes. The step-by-step procedure for implementing this system is described, ensuring a robust and reliable solution.