Floating Object Detection and Monitoring in Aquatic Environments

Floating object detection encompasses the identification, localisation and tracking of natural and anthropogenic matter—ranging from driftwood and algal blooms to plastic debris and oil slicks—on the surfaces of rivers, lakes and coastal waters. Advances in optical imaging, acoustic sensing and remote platforms (satellites, unmanned aerial vehicles and autonomous surface vessels) have been integrated with machine-learning techniques to cope with dynamic backgrounds, variable illumination and complex fluid motion. Deep-learning architectures now underpin real-time vision systems that combine background modelling, feature fusion and attention modules to detect small or low-contrast objects. Concurrently, networked deployments using edge computing and next-generation communications support continuous monitoring, rapid hazard warnings and data-driven environmental management on local to global scales.

Research from Nature Portfolio

No recent Nature Portfolio content available.

Floating Object Detection and Monitoring in Aquatic Environments publication trend

The graph below shows the total number of articles in floating object detection and monitoring in aquatic environments across all publications each year (not limited to Nature Index journals).

Technical terms

Convolutional Neural Network (CNN): A deep-learning model composed of convolutional layers for automated feature extraction in images.

You Only Look Once (YOLO): A family of real-time object-detection architectures that predict bounding boxes and class probabilities in a single evaluation.

Attention Mechanism: A module that weights features spatially or across channels to focus the model on salient regions.

Feature Pyramid Network (FPN): A multi-scale architecture that combines semantic information from different convolutional layers to improve detection of varied object sizes.

Mean Average Precision (mAP): A standard metric for object-detection performance, averaging precision across recall levels and object categories.

Background Subtraction: A computer-vision technique that separates moving foreground objects from a static or slowly varying background.

Edge Computing: Distributed processing of data near the source of acquisition to reduce latency and bandwidth usage in monitoring systems.

References

1. Detection of River Floating Garbage Based on Improved YOLOv5. Mathematics (2022).
2. A Floating-Waste-Detection Method for Unmanned Surface Vehicle Based on Feature Fusion and Enhancement. Journal of Marine Science and Engineering (2023).
3. Development of a Lightweight Floating Object Detection Algorithm. Water (2024).
4. A Framework and Method for Surface Floating Object Detection Based on 6G Networks. Electronics (2022).