Use of automated seabed photomosaicing in forensic analysis of the RMS TITANIC disaster

During the 1998 expedition to the wreckage of the RMS TITANIC, a new technique was used for creating a photomosaic that significantly improved the process and the results of forensic analysis. In excess of 20,000 digital images were collected with two high-resolution digital cameras mounted on the submersible NAUTILE and with an ROV based high-resolution video camera. The individual images varied in brightness, contrast and color as well as orientation and distance. An automated technique was developed to apply corrections for radiometric and geometric distortions, thereby yielding normalized images with proper scale. Once normalized and properly scaled, the individual images were merged using image correlation methods that included boundary feathering and smoothing of lighting and contrast inhomogeneities. The final digital imagery is nearly 1 Gigabyte in size and allows for resolution of objects on site as small as a coin. Submersible and ROV depth and altitude measurements were collected and summed over the wreckage to provide total depth. These measurements were then gridded to produce a three dimensional surface model of the wreck site and surrounding region of debris. By combining the 3D surface model with the photomosaic imagery, a completely interactive three dimensional model of the TITANIC wreckage was created for the use of researchers and the forensic engineers. With this imagery and model, the engineers could, for the first time, detect large scale patterns and features that were obscured in detailed examination. With such patterns, the engineers were able to answer many of the questions that have persisted about the ship and its breakup since the morning of April 13th, 1912