Object identification with acoustic lenses

With increasing frequency, underwater work is situated in rivers and coastal areas where visibility is a fraction of a meter. Two acoustic lens sonars, mine reacquisition and identification sonar (MIRIS) and dual-frequency identification sonar (DIDSON) have sufficiently high resolutions and rapid refresh rates that they can substitute for optical systems in turbid water where optical systems fall. MIRIS operates at 2 MHz and uses 64 0.3/spl deg/ beams to form images of objects at ranges up to 10 m. DIDSON operates at two frequencies, 1.8 MHz or 1.0 MHz, and forms 96 beams spaced 0.3/spl deg/ apart or 48 beams spaced 0.6/spl deg/ apart respectively. It images out to 12 m at 1.8 MHz and 40 m at 1.0 MHz. Both sonars have update rates between 5-27 frames/s. Identification of objects with these sonars eliminates the need to send a diver to identify them by tactile means. These sonars provide biologists with a new technology for observing marine life including counting crab and monitoring fish behavior in turbid water. In both sonars, the transmit and receive beams are formed with acoustic lenses with rectangular apertures and made of polymethylpentene plastic and FC-70 liquid. The lenses eliminate the need for beamforming electronics. DIDSON can be commanded to focus on objects from 1 m to its maximum range. The sonars are approximately 30-cm long, 20-cm high, and 17-cm wide. They consume between 20 and 30 watts, which is important to submersibles with a power budget. The paper shows images from these sonars, and discusses the basic principles of beam formation with transducers and lens elements. These sonars form images with "line focused" beams that provide good images in many but not all conditions. If objects were at the same range in the same beam but at different elevations, this type of imaging could not sort them out. An example would be to try to image an object embedded in a pile of debris on the ocean floor. True video, using "point-focused" optics could meaningfully image the object imbedded in the pile as long as it were not totally covered. Fortunately, for the great majority of imaging tasks, MIRIS and DIDSON provide unambiguous, near-photographic quality images.