Visibility estimation through image analytics
June 20, 2023
Project Report
Author:
Published in:
MIT Lincoln Laboratory Report ATC-453
R&D Area:
R&D Group:
Summary
MIT Lincoln Laboratory (MIT LL) has developed an algorithm, known as the Visibility Estimation through Image Analytics Algorithm (VEIA), that ingests camera imagery collected by the FAA Weather Cameras Program Office (WeatherCams) and estimates the meteorological visibility in statute miles. The algorithm uses the presence of edges in the imagery and the strength of those edges to provide an estimation of the meteorological visibility within the scene. The algorithm also combines the estimates from multiple camera images into one estimate for a site or location using information about the agreement between camera estimates and the position of the Sun relative to each camera's view. The final output for a site is a prevailing visibility estimate in statute miles that can be easily compared to existing automated surface observation systems (ASOS) and/or human-observed visibility. This report includes thorough discussion of the VEIA background, development methodology, and transition process to the WeatherCams office operational platform (Sections 2–4). A detailed software description with flow diagrams is also provided in Section 5. Section 6 provides a brief overview of future research and development related to the VEIA algorithm.