A new algorithm, AEA-RDCP, has been developed to process images for estimating fog density and visibility, significantly enhancing safety in maritime navigation.

This research introduces a novel methodology that improves atmospheric light estimation and refines fog detection algorithms, leading to greater accuracy.

The study specifically aims to enhance ship detection in coastal waters during foggy conditions by utilizing a modified object detection model known as YOLOv8s-Fog.

Improvements to the YOLOv8 model for detecting cone buckets include CA attention, color space transformation, and a new loss function, which collectively boost accuracy and recall rates.

The study was conducted by authors Mingrui Dai, Guohua Li, and Weifeng Shi from the Institute of Computing Technology, China Academy of Railway Sciences in Beijing.

The article detailing this research was submitted on August 13, 2024, revised on September 8, accepted on September 11, and published on September 12, 2024.

For the research, two datasets of marine images were created, one with minimal atmospheric light interference and another heavily influenced by it, allowing for comparative analysis.

The study underscores the necessity for onboard sensors to collect data that ensures safe navigation in poor visibility conditions.

The findings of this research are documented in the journal 'Sensors', volume 24, article number 5930.

The enhanced YOLOv8s-Fog model achieved an average detection accuracy of 74.4%, surpassing the standard YOLOv8s by 1.2%.

In South Korea, there are 23 observer measurement centers and 291 meteorological systems dedicated to measuring visibility, highlighting the importance of accurate visibility data.

Visibility measurement equipment, which relies on optical sensors, is crucial yet can be expensive and prone to significant errors, particularly at sea.