A fisheries application of a dual-frequency identification sonar acoustic camera

Size: px

Start display at page:

Download "A fisheries application of a dual-frequency identification sonar acoustic camera"

Thomas Houston
6 years ago
Views:

1 ICES Journal of Marine Science, 60: doi: /s (03) A fisheries application of a dual-frequency identification sonar acoustic camera Russell A. Moursund, Thomas J. Carlson, and Rock D. Peters Moursund, R. A., Carlson, T. J., and Peters, R. D A fisheries application of a dualfrequency identification sonar acoustic camera. ICES Journal of Marine Science, 60: The uses of an acoustic camera in fish-passage research at hydropower facilities are being explored by the U.S. Army Corps of Engineers. The Dual-Frequency Identification Sonar (DIDSON) is a high-definition imaging sonar that obtains near-video quality images for the identification of objects underwater. Developed originally for the U.S. Navy by the University of Washington s Applied Physics Laboratory, it bridges the gap between existing fisheries-assessment sonar and optical systems. The images within 12 m of this acoustic camera are sufficiently clear such that fish can be observed undulating as they swim and their orientation ascertained in otherwise zero-visibility water. In the 1.8 MHz high-frequency mode, this system comprises 96 beams over a 29 field-of-view. The high resolution and fast frame rate provide target visualization in real time. The DIDSON can be used where conventional underwater cameras would be limited by low light levels and high turbidity. Ó 2003 International Council for the Exploration of the Sea. Published by Elsevier Science Ltd. All rights reserved. Keywords: acoustic camera, DIDSON, hydropower limits, multibeam sonar. R. A. Moursund: Pacific Northwest National Laboratory, PO Box 999, MSIN: K6-85, Richland, WA, USA. T. J. Carlson: Battelle Portland Office, 620 SW Fifth Avenue, Suite 810, Portland, OR 97204, USA; thomas.carlson@pnl.gov. R. D. Peters: U.S. Army Corps of Engineers, Portland District, PO Box 2946, Portland, OR 97208, USA; rock.d.peters@nwp01.usace.army.mil. Correspondence to R. A. Moursund: tel: þ ; fax: þ ; russell.moursund@pnl.gov. Introduction The U.S. Army Corps of Engineers (USACE) operates eight hydroelectric dams on the lower Columbia and Snake Rivers through which salmon and other anadromous fish migrate. Adult fishways and bypass systems for juveniles are in place at the dams to aid salmon passage. As a steward of the region s natural resources, USACE continues to improve existing systems and is evaluating long-term alternatives for fish passage such as surface bypasses. The Dalles Dam, located at Columbia River mile 192, comprises a navigation lock, a spillway perpendicular to the main river channel, and a powerhouse parallel to it with non-overflow dams on each side. Situated between Bonneville and John Day Dams, it is the second dam upstream of the mouth of the Columbia River (Figure 1). The powerhouse spans 637 m and has 22 main units (MU). Two fish units (FU) are located just west of MU1. The total hydraulic capacity of the powerhouse is m 3 s ÿ1. The spillway spans 421 m and has 23 bays. An ice and trash sluiceway extends the entire length of the powerhouse with skimmer gates that may be opened above each turbine intake and discharged into the sluiceway. During the fish-passage season (April through November), the three gates above MU1 are normally opened to allow juvenile salmonids to pass downstream. The maximum discharge rate of the ice and trash sluiceway through all three gates of one turbine unit is approximately 140 m 3 s ÿ1. At the Dalles Dam in 2001 and 2002, prototype turbine intake occlusion plates with J-extensions were evaluated as a new means of protecting downstream-migrating juvenile salmon by diverting them from the turbine intakes into the spillway or sluiceway. The intake occlusions were formed by a sheet of steel that rested against the upstream side of the trashracks. The occlusion plates covered the upper half of the intakes at FU 1 2 and MU 1 5 and prevented flow from entering the upper half of the turbine intake. At the lower edge of the occlusion, another steel sheet extended 6.1 m into the forebay, perpendicular to the trashracks. Then, at the upstream edge of this horizontal section, there was a 1.5-m-high vertical piece. Thus, the entire occlusion assembly had the shape of the capital letter J when /03/000678þ06 $30.00 Ó 2003 International Council for the Exploration of the Sea. Published by Elsevier Science Ltd. All rights reserved.

2 An acoustic camera in fish-passage research 679 Figure 1. Location of the Dalles Dam in relation to other mainstem hydroelectric facilities in the Columbia River Basin. viewed from the side. A Dual-Frequency Identification Sonar (DIDSON) was deployed at the Dalles Dam to evaluate smolt behavior near gaps in the J-occlusions in 2001 and the presence of large, predatory fish near the structure in Materials and methods Initial tests of a Limpet Mine Imaging Sonar (LIMIS) had been conducted at the Pacific Northwest National Laboratory (PNNL) in Richland, Washington. The LIMIS was developed by the University of Washington s Applied Physics Laboratory as one of several high-resolution, imaging sonars that utilize acoustic lenses for beam forming. LIMIS was a hand-held, single-frequency (2 MHz), 64-beam precursor to DIDSON (Belcher et al., 1999). Tests of how well the sonar imaged free-swimming fish were conducted in various holding tanks. The fish tested were native to the Columbia River basin: chinook smolts (Oncorhynchus tshawytscha), adult rainbow trout (Oncorhynchus mykiss), and juvenile Pacific lamprey (Lampetra tridentata). Based on these images, the USACE obtained a LIMIS unit to conduct tests in the field during the spring migration. By this time, however, the DIDSON had been developed and was being used for field tests conducted at the Dalles Dam in 2001 and The DIDSON was also developed by the University of Washington s Applied Physics Laboratory and it also uses an acoustic lens (Belcher et al., 2001). It has two available frequency modes: a low-frequency mode of 1.0 MHz and a high-frequency mode of 1.8 MHz. These two frequencies bring trade-offs in both maximum range and resolution (Table 1). In high-frequency mode, the 0.3 beamwidths ensure that within the 12-m range all but the smallest targets (<63 mm at the furthest range) intercept multiple beams (Figure 2). Software need only account for beam geometry and display the relative amplitude. The system does so in real time, and for a 12-m range achieves seven frames per second. The result is not just an acoustic still image of the underwater environment, but something akin to an acoustic video camera. For riverine fisheries applications, the high-frequency mode was by far the most useful because it provided the highresolution images that defined the outline, shape, and even fins of target fish. Table 1. Technical specifications of the DIDSON. High-frequency mode Operating frequency 1.8 MHz Beamwidth (two-way) 0.3 horizontal by 12 vertical Number of beams 96 Low-frequency mode Operating frequency 1.0 MHz Beamwidth (two-way) 0.4 horizontal by 12 vertical No. of beams 48 Both modes Field-of-view 29 Power consumption 30 W typical Weight in air 7.0 kg Weight in water ÿ0.61 kg Dimensions 171 mm 307 mm 206 mm

680 R. A. Moursund et al. Figure 2. Diagram of the 96 beams along the horizontal plane of the sonar in high-frequency mode including fish targets and a mechanical structure.

3 680 R. A. Moursund et al. Figure 2. Diagram of the 96 beams along the horizontal plane of the sonar in high-frequency mode including fish targets and a mechanical structure. Physically, the sonar is small and nearly neutrally buoyant in water. The transducer array, acoustic lens, and the associated electronics are all contained in a single underwater housing (Figure 3). The earlier LIMIS system was designed to be diver hand-held and battery-operated. The DIDSON system maintained the low weight, small profile, and low power consumption characteristics. In this research, the DIDSON transmitted data via a telemetry cable to the surface, where a laptop computer was sufficient to control the sonar, display images in real-time, and record the raw data. Removable data storage was used to maintain a continuous archive. Results The system was capable and robust for the real-time imaging of fish. Fish targets were confirmed by the swimming silhouette and the rheotactic behavior of fish. One feature of this sonar design was that both nearby structures and fish could be observed at the same time on the same transmitted pulse. We found it desirable to have some structure in view at the same time that fish were present simply for spatial reference. Structure in this case could be walls, gratings, support cables, or pilings. Figure 3. Photograph of the DIDSON underwater unit as deployed. The single housing contains the transducer array and acoustic lens in the upper compartment and the electronics below. This unit is attached to a three-axis underwater rotator assembly. Discrimination of individual fish within schools was also shown (Figure 4). Figures 4 and 5 show unaltered single frames of the raw, real-time output of the sonar. These data were collected at seven frames per second. The multibeam nature of the DIDSON made it robust in the acoustically noisy environments commonly encountered at hydropower facilities. The close spacing of 0.3 between the beams also provided detailed information about each target. Along the other axis, the 0.3 height of each beam created the illusion of three-dimensionality, even though the data were two-dimensional. This was due to the reflection of acoustic energy across the entire curved surface of the fish combined with the high resolution of the system. The resulting images showed the body depth of fish (Figure 5). This was an important feature of the sonar because it gave additional visual data with which to differentiate targets. Further, because the data were collected at several frames per second, the display over time showed motion characteristics such as the undulation of swimming fish. The system was weak at the identification of fish moving orthogonally to the ensonified plane created by the linear transducer array. In our field trials, the sonar did not work

4 An acoustic camera in ﬁsh-passage research 681 Figure 4. A school of 18-cm ﬁsh shown in front of the operating ice and trash sluiceway. The piernose, trashrack guide, and J-occlusion guide slots are clearly visible. The image is formed with details both along the transducer array and across it. Range marks (white numbers) are in meters. well imaging ﬁsh in a bathymetry-sonar type of deployment. The best geometry to obtain high-resolution images of ﬁsh and the possibilities for distinguishing ﬁsh from other objects was when targets were aligned along the 96-beam plane. Discussion Real-time, acoustic-imaging systems date back to the 1960s (Smyth et al., 1963; Jacobs, 1965). These early examples of acoustic cameras were part of the development pathway to the medical ultrasound devices in use today. For bathymetry and ﬁsheries applications, a variety of imaging sonars are currently available from multi-beam to mechanically scanned systems. The diversity of those sonars forms a frequency, range, and resolution continuum with the DIDSON acoustic camera at the high-frequency, shortrange, and detailed-resolution end. Thus, the DIDSON is useful where the most detailed imaging is needed over short ranges. While three-dimensional sonars are available, see Hansen and Andersen (1993) for example, the twodimensional data and acoustic-lens feature of the DIDSON kept the computational and data-storage costs relatively low. The DIDSON acoustic camera was used at the Dalles Dam, where optical underwater cameras would be limited in sampling range to less than 1 m due to low light levels and high turbidity, and where single-beam sonars were limited by the conﬁned space and complex structures. The sonars used to date for ﬁsh-passage studies at hydropower facilities have been specially constructed with highfrequency, single- and split-beam transducers having very low sidelobes. Acoustic images were taken with the DIDSON of ﬁsh located in conﬁned spaces and near bubbles, which would have confounded single-beam methods. For ﬁsh-passage studies, this high-deﬁnition, imaging-sonar technology occupies a niche between short-range, highresolution optics and long-range, low-resolution sonar systems (Table 2). The DIDSON oﬀered biologists the ability to see underwater objects under conditions of low light and high turbidity, albeit at lower resolution compared to optical vision. This technology can be applied to other areas of ﬁsheries research and aquatic ecology. The interaction of underwater boundaries and ﬁsh, the ecology of nekton and benthos, ﬁshing-gear eﬃciency and selectivity, and

5 682 R. A. Moursund et al. Figure 5. Several ﬁsh in front of an intake trashrack at the Dalles Dam are shown. The bars of the trashrack are visible as well as acoustic shadows created by two of the ﬁsh. Dorsal, ventral, and caudal ﬁns can be seen on the largest ﬁsh. All these ﬁsh were positively rheotactic and actively avoided passage through the trashrack. Range marks (white numbers) are in meters. behavior studies should all prove to be rich areas of new research. The instrument works equally well in either freshwater or seawater, since there is no diﬀerence in the absorption of sound at frequencies greater than 1 MHz. Also, studies can be conducted without the behavioral implications of white light and at a much longer range than infrared light. The DIDSON high-deﬁnition acoustic camera is a signiﬁcant new tool in ﬁsheries and aquatic research. In particular, the DIDSON has demonstrated its value in ﬁshpassage research. The technology has allowed the U.S. Army Corps of Engineers unprecedented insight into ﬁsh movements very near structures and within conﬁned spaces. Detailed images of ﬁsh were positively identiﬁed within a biologically meaningful ﬁeld of view. These data, based on the in situ observations of ﬁsh near hydropower facilities, provide the basis for improving ﬁsh passage in the region. Acknowledgements This research was funded by the U.S. Army Corps of Engineers, Portland District. Table 2. Fish passage observation technology comparison. Display resolution Maximum range Field-of-view Optical camera Acoustic camera Traditional sonar pixels (video) <1 m (in turbid water) pixels (HF mode) 12 m (HF mode) range data bins >50 m 6

6 An acoustic camera in fish-passage research 683 References Belcher, E., Dinh, H., Lynn, D., and Laughlin, T Beamforming and imaging with acoustic lenses in small, high-frequency sonars, pp Proceedings of Oceans 1999 Conference, September, Seattle, Washington. Belcher, E., Matsuyama, B., and Trimble, G Object identification with acoustic lenses, pp Proceedings of Oceans 2001 Conference, 5 8 November, Honolulu, Hawaii. Hansen, R. K., and Andersen, P. A D acoustic camera for underwater imaging. In Acoustic Imaging, pp Ed. by Y. Wei, and B. Gu. Plenum Press, New York. Jacobs, J. E The ultrasound camera. Science Journal (London), 1(4): Smyth, C. N., Poynton, F. Y., and Sayers, J. F The ultrasound image camera. Proceedings of IEE, 110:

Brad Eppard AFEP Coordinator Portland, Oregon December 14, US Army Corps of Engineers Portland District

Brad Eppard AFEP Coordinator Portland, Oregon December 14, 2011 US Army Corps of Engineers Portland District JSATS: Program Overview and Objectives 1. Develop a system for study of subyearling Chinook