We conducted autonomous underwater vehicle (AUV) surveys of a large-scale hydrothermal deposit called the Hakurei site located in the Bayonnaise knoll caldera of the Izu-Ogasawara island arc during the YK11-11 research cruise of the R/V Yokosuka in December 2011. Two dives of the AUV Urashima, a vehicle developed by the Japan Agency for Marine-Earth Science and Technology (JAMSTEC), were devoted to collecting side-scan and multi-beam acoustic data from the southern part of the caldera, including the location of the hydrothermal site. We made detailed corrections for the vehicle's position data and obtained a high-resolution multi-beam bathymetric map. After radiometric corrections, side-scan and multi-beam backscatter mosaics were created by using the corrected position data. These mosaics commonly show lower backscatter intensities on the caldera floor, suggesting that ponded sediments cover the seafloor. On the other hand, they show stronger backscattering on the central cone, probably because it is exposed volcanic rocks with little sediment cover. An elongated zone of a distinctive, scaly pattern of strong backscattering was identified in the side-scan image of the southwestern caldera floor. However, this remarkable pattern appears rather modestly in the bathymetric map and in the multi-beam image. It implies that this pattern is not the result of topographic relief but rather reflects different materials on the seafloor, possibly representing erupted volcanic rocks with a little sediment cover. The side-scan image of the Hakurei hydrothermal site is characterized by short-wavelength alternation of strong and weak backscattering with high contrast, which probably represents rugged topography associated with sulfide chimneys and mounds. These different characteristics of the seafloor were quantitatively classified by means of the side-scan image. The Hakurei site was marked by a distinctive spotty pattern in the multi-beam image, a different expression by the side-scan image. This spotty pattern was utilized to characterize the hydrothermal site and to delineate other areas that have similar characteristics. The resulting distribution is generally in good agreement with the classification result from the side-scan image. We suggest that hydrothermal activity at various scales would occur in several places in the caldera.