Evaluation of Propeller Inspection Using Different Deployment Strategies

Authors

  • Ghita Ikmel University Sidi Mohamed Ben Abdellah https://orcid.org/0000-0002-8228-7273
  • Mohamed Salim Harras Chemnitz University of Technology
  • Wolfram Hardt Chemnitz University of Technology
  • Najiba El Amrani El Idrissi University Sidi Mohamed Ben Abdellah

DOI:

https://doi.org/10.14464/ess.v10i8.661

Abstract

In recent years, the use of Unmanned Aerial Vehicles (UAVs) for various applications has increased significantly. Among these applications, the inspection of infrastructures using UAVs has become a prominent area of research. This paper evaluates the efficiency of the YOLOv5 algorithm for propeller inspection. The algorithm's deployment across various platforms such as PC, Google Colab, and Jetson Nano is examined, with a focus on different deployment formats like PyTorch, ONNX, TensorFlow Lite, and others. The study
highlights the often-overlooked importance of the deployment phase in the development of AI models and underscores its significance for the practical application of AI in real-world scenarios.
Keywords— Computer vision, algorithm deployment, propeller inspection, Deployment strategies, efficiency improvement

Author Biographies

Ghita Ikmel, University Sidi Mohamed Ben Abdellah

Signals Systems & Components Laboratory (LSSC)
Faculty of Science and Technology University Sidi Mohamed Ben Abdellah

Fez, Morocco

Najiba El Amrani El Idrissi, University Sidi Mohamed Ben Abdellah

Prof. Dr. El Amrani El Idrissi

ESS

Downloads

Published

2023-12-23