Model-based Metric 3D Shape and Motion Reconstruction of Wild Bottlenose Dolphins in Drone-Shot Videos

2025-04-22

AI Summary

Accurate markerless 3D reconstruction of wild aquatic animals’ morphology and motion remains challenging under natural underwater conditions due to water-induced occlusions, optical attenuation, and lack of geometric references. Method: This paper introduces the first monocular video-based metric 3D reconstruction method explicitly integrating a water transmission model. Leveraging a parametric dolphin deformation model, it jointly models water-caused occlusion and light-path attenuation, while enforcing monocular geometric constraints and spatiotemporal motion regularization on drone-captured underwater videos. Contribution/Results: The method achieves millimeter-scale 3D reconstruction of wild bottlenose dolphins in situ. It enables, for the first time, non-contact estimation of dolphin mass and volume—validated against manual 2D measurements with mean error <4.2%. This demonstrates its viability as a novel paradigm for wildlife health assessment. The work bridges a critical gap in field-based 3D biometrics for aquatic species and establishes a scalable technical framework for non-invasive monitoring of endangered marine fauna.

Technology Category

• Computer Vision: 3D Computer Vision
• Machine Learning: Calibration & Uncertainty Quantification
• Intelligent Robots: State Estimation

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

We address the problem of estimating the metric 3D shape and motion of wild dolphins from monocular video, with the aim of assessing their body condition. While considerable progress has been made in reconstructing 3D models of terrestrial quadrupeds, aquatic animals remain unexplored due to the difficulty of observing them in their natural underwater environment. To address this, we propose a model-based approach that incorporates a transmission model to account for water-induced occlusion. We apply our method to video captured under different sea conditions. We estimate mass and volume, and compare our results to a manual 2D measurements-based method.

Problem

• Estimating 3D shape and motion of wild dolphins from monocular video
• Assessing dolphin body condition using model-based reconstruction
• Overcoming water occlusion challenges in aquatic animal reconstruction

Innovation

• Model-based 3D reconstruction of dolphins
• Transmission model for water occlusion
• Mass and volume estimation from video