Textural analysis and structure-tracking for geological mapping: applications to sonar images from endeavour segment, Juan de Fuca Ridge

The volume of data collected by side-scan sonar during seafloor surveys has become larger and larger as the resolution of these systems has improved. As a result, new image processing techniques need to be developed to partly automate the interpretation of this increasing wealth of data. The first two steps in the geological analysis of a new image usually are the mapping of linear structures, and of morphologic units. These maps are then used in tectonic and geological interpretations. Linear features are detected by adaptive filtering. Their properties (length, direction, sinuosity, ...) are computed and quantified by the new structure tracking algorithm. Mapping of morphologic units is addressed by a textural analysis method, based on grey-level co-occurrence matrices. The authors have applied these techniques to high-resolution sonar images of the Endeavour Segment, Juan de Fuca Ridge. Separation of geologically meaningful regions can be easily attained using the authors' textural analysis method. Some regions fan be distinguished by numerical methods, whereas visual interpretation had to rely on supplementary data (photos, dives)