UM Logo ART GLEASON  |  Underwater Landscape Mosaics

Large-area mosaics


For several years we have been slowly building capability to make mosaics over larger areas. Examples include:

Historical progress by our team in producing large-area mosaics is summarized in the table below. Processing has become faster and spatial resolution has increased over time as computers have improved. Depth reconstruction (2.5-D processing) has also become feasible recently. The total number of images that can be processed remains a bottleneck, however, even with reasonably powerful computing clusters (64 CPU, 320 GB RAM used for the Flower Gardens dataset). Processing larger areas than these, at higher spatial resolution and in less time, will require new approaches to managing, subdividing, and assembling overlapping datasets.

table of lage-area mosaic statistics

Platforms used for large-areas

DPV / DSLR rig DPV 1

Towed cameras at fixed depth
A) 7.5 m pole with 6 GoPros DPV 1 B) towing near Carysfort lighthouse, FL DPV 1

Towed diver Towed array of gopro cameras with diver (Gareth Williams photo)