A method of generating 3D views of aquatic plants with DIDSON

DIDSON is a highly definition sonar which uses acoustic lenses to make images in dark, turbid waters where optical systems are ineffective. The acoustic lenses of it focus and form an acoustic image on the transducer array at the rear of the sonar. Then electronics convert that acoustic image into a digital image which can be displayed on a computer screen. The combination of high frequency sound waves, acoustic lens configuration, and high resolution transducer array make it suitable for different kinds of applications such as: positive identification and inspection, navigation in close quarters, monitoring in turbines or being used as a security video camera underwater in dark and turbid conditions etc. However, the images formed by DIDSON are a little bit different from the images generated by Optical video cameras. Sometimes, it is high demanding for users to analyze the DIDSON data. In order to make the images being understand much more easily, one method is to improve the algorithm which was being used in the conversion of acoustic images. We used the standard DIDSON to take the video of the water column at several spots in the Yamanaka Lake, Japan. Considering with the depth of lake, we chose the high frequency sound waves (1.8 MHz) which consists of 96 beams, each beam spanning 0.3 degrees in the horizontal plane. Aimed to work out the whole 3D image of the aquatic plants, we set acoustic beam projections of DIDSON towards the bottom of the lake and turned it 45 degrees up from vertical. By doing that, echoes came back from aquatic plants were began with the root and ended with the top of the aquatic plants. Also, we added a concentrator lens which reduces the vertical aperture from 14 to 1 degree in front of the DIDSON. Therefore, each beam pattern is consisting of 96 beams covering a 30 deg by 1 deg spread. It allowed horizontal DIDSON beams to go farther in the water with reduced surface and bottom reverberation and illuminates thin slices which were very- - useful in 3D reconstruction of the growth condition of aquatic plants. In this study, we designed a new system for the detection of aquatic plants with suitable equipments which can be widely used in surveys of underwater ecological reserves. We also developed a software which can generate 3D views of the aquatic plants in accuracy of 1 centimeter. For future work, we hope to apply it into the identification of aquatic plants species.