The software first compares overlapping images or video frames to identify visual features that appear in multiple views. It uses those matches to estimate each camera’s position, then triangulates feature locations in three dimensions. This sequence converts image evidence into spatially organized measurements, allowing the final model to represent submerged geometry rather than serving only as a visual record.
Image overlap gives software repeated views of the same visual features from different camera positions. Those shared features provide the relationships needed to estimate camera locations and triangulate their positions in three-dimensional space. Insufficiently related images would limit the reconstruction process, while well-connected imagery supports a more coherent point cloud or textured mesh.
Triangulation determines the three-dimensional location of visual features by using their positions across multiple views and the estimated camera geometry. The resulting spatial points form the basis of a scaled point cloud or help support a textured mesh. This step is important because it turns matched image features into measurable locations within the submerged environment.
A typical workflow begins by capturing overlapping photographs or video of the submerged target. Modeling software then matches recurring visual features, estimates the positions from which images were captured, and triangulates feature locations. The processed data produces a scaled point cloud or textured mesh that can be used for measurement, visualization, and environmental assessment.
Imagery may be collected by divers, remotely operated vehicles, or autonomous systems. These platforms provide different ways to document submerged environments, structures, and objects that are difficult to access directly. Regardless of the platform, the captured photographs or video must supply overlapping views so software can relate visual features and reconstruct the subject spatially.
Environmental researchers can model coral reefs and seafloor habitats to create measurable records for assessment and conservation planning. Repeated modeling also supports documentation of change over time. Because the resulting records can be visualized and measured, they help communicate site conditions and support comparisons of submerged environments without relying only on direct observation.