Combining Visible and Infrared Spectrum Imagery using Machine Learning for Small Unmanned Aerial System Detection is a research paper by Vinicius G. Goecks, Grayson Woods, and John Valasek of Texas A&M.

The demand for technology and solutions to counter commercially available small unmanned aerial systems (sUAS) is on the rise. Advances in machine learning and deep neural networks for object detection, combined with the reduced cost and power requirements of cameras, have paved the way for promising vision-based solutions in sUAS detection. However, relying solely on the visible spectrum has posed reliability challenges in low-contrast scenarios, such as when sUAS fly beneath the treeline or against bright light sources. As an alternative, utilizing long-wave infrared (LWIR) sensors, which capture the relatively high heat signatures emitted by sUAS during flight, can generate images that effectively distinguish the sUAS from its background.

Nonetheless, compared to the readily accessible visible spectrum sensors, LWIR sensors exhibit lower resolution and may generate increased false positives when subjected to heat sources like birds. This research suggests a solution by combining the strengths of LWIR and visible spectrum sensors through machine learning for the vision-based detection of small unmanned aerial systems (sUAS). By leveraging the heightened background contrast provided by the LWIR sensor and synchronizing it with the relatively enhanced resolution of the visible spectrum sensor, a deep learning model was trained to detect sUAS even in challenging environments.

Publication Date– April 2020

Combining Visible and Infrared Spectrum Imagery using Machine Learning for Small Unmanned Aerial System Detection contains the following major sections:

  • Introduction
  • Related Work
  • Methods
  • Results and Discussion
  • Conclusion
  • Limitations and Future Work

There are no noted distribution limitations of this document. C-UAS Hub does not own this content and provides a link for users at the bottom of the page to access it in its original location. This allows the author(s) to track important article metrics related to their work. All credit goes to its rightful owner.

Post Image Credit- envatoelements by halfpoint