Enhanced Underwater 3-D Reconstruction: Case Study at Los Organos Reef Using 3dFeatureUp and 3dVisualAppUp

Abstract

In recent years, photogrammetry for tridimensional (3-D) reconstruction of underwater environments has gained significant interest, offering a nonintrusive approach to study and monitor aquatic ecosystems. However, underwater photogrammetry requires specialized equipment capable of capturing high-resolution images with adequate and uniform lighting. This article introduces two innovative methodologies, 3-D reconstruction based on Feature Enhancement (3dFeatureUp) and 3-D reconstruction based on Visual Appearance Enhancement (3dVisualAppUp), which are designed to enhance the accuracy and visual quality of 3-D underwater models. These methodologies incorporate novel algorithms for water image enhancement (WaterImgEnh) and water image restoration (WaterImgRest), aiming to address the challenges posed by underwater image acquisition such as light attenuation, color distortion, and visibility reduction. This approach is validated through the 3-D reconstruction of Los Organos Reef, located in Piura, Peru, employing a combination of standard and specialized underwater cameras along with a customized lighting system. The results demonstrate significant improvements in both the structural characteristics and visual appearance of the 3-D models, as compared to those generated by traditional methods or using specialized underwater cameras alone. The 3dFeatureUp method significantly enhances the structural features of the model by merging images from multiple cameras, while the 3dVisualAppUp methodology improves the visual quality of the models by correcting color imbalances and removing water effects.