Stereophotography in marine Biology: Development, Application and Evaluation.


Contact
Philipp.Fischer [ at ] awi.de

Abstract

Stereophotography has been well known since the 1840s and has been continuously applied as a tool in the aquatic sciences since the early sixties. Stereophotography enables reference-free and non-invasive measurements of object coordinates in the three-dimensional domain of the underwater world. Thus, the need for a parallel estimation of quantitative parameters such as distance or volume is not necessary. Furthermore, assessments of photographic time series of communities allow the estimation of quantitative derivatives, including the growth or productivity over time, using an empirically determined relationship between biomass and linear measurements. However, the application of underwater stereophotography as a routine tool for the assessment of species metric traits is only seldom used by the scientific community. The apparently complex set-up of an adequate stereoscopic camera system, as well as the complicate post process on the stereo images for measurements with a high accuracy and precision, deters scientists to use this convincing measurement technique. This study had two main goals: The first main goal was to provide a practical, reliable and sufficiently precise procedure to assess length dimensions by stereo photographic evaluation for ecological studies in the marine environment. The second main goal of this study was to evaluate the possibilities, as well as the technical limitations, of the non-invasive method stereophotography for in situ assessments. In the first part I introduce a stereo camera system that can be easily assembled and rebuild with the use of standard products including a self-developed installation guide and a complete parts list. I provide a complete workflow for the calibration of the stereo camera and a post-process for the correction of distorted digital stereo images of the produced stereo images. Furthermore, I provide the software needed to execute these processes and to measure imaged objects with a high accuracy and precision. I evaluate the combination of the stereo camera, the software and the calibration / correction process concerning the accuracy and the precision of the system. An initial evaluation of the stereo camera revealed an accuracy error of -0.6 % and a precision of 93 %. The correction process slightly affected the accuracy error but enhanced the precision to a value of 10.7 %. The use of a wide angle converter had only little influence on the measurement accuracy after the correction of the stereo images. 95 SUMMARY ____________________________________ In the second part I tested the reliability of the stereo camera and the calibration and correction process in an in situ experiment. Therefore, I tested the hypothesis, that significant differences in length-height relationship of Ctenolabrus rupestris between three different sampling sites (South harbour, Sathurnbrunn, MarGate) around the North Sea island Helgoland. The results show that length height-length relationship measured at the South harbour was significantly different from Sathurnbrunn and Margate. Possible responsible reasons are discussed in relation to the obtained findings in size differences. In the last part of this study I assessed the possibilities as well as the limitations of the non-invasive method stereophotography in combination with DISTA for the use in close-up benthos observation. DISTA is a professional software for photogrammetric calibration and digital stereo analysis and was adapted to under water applications in close cooperation with the Raumbezogene Informationstechnik in Mainz (i3-Mainz) as part of this study. In length measurements of 1 mm the systems accuracy is 6.2 % with a precision of 24 %. The system is able to distinguish significantly between objects of 1 mm in size with a probability of 90 % with a sample size of N = 36. Using this method, it was furthermore possible to assess length measurements of objects across the borders of neighbouring stereo image pairs in a test field of 36 stereo images. A for this task new implemented feature of DISTA failed because of unexpected perspectival problems due to the close distance between the stereo camera and the object used in this set-up. The results show that the camera system in combination with the calibration software Camera Calibration Toolbox and StereoMarker are displaying a high accuracy and precision. The system presented its ability to produce a reliable set of data in the length- and height assessment of benthic associated fish, but has to prove this ability in the observation of pelagic fish. In combination with DISTA the camera works with a high accuracy and precision in a simulation of benthos observation. In conclusion, this study and the obtained results provide a solid basis in the building of a stereo-system and the use of stereophotography in marine biology. Furthermore, it indicates the limitations but gives also suggestions for improvements. With the here gained knowledge further studies in fish or benthos observation could be performed in shorter time periods and in high quality.



Item Type
Thesis (PhD)
Authors
Divisions
Primary Division
Programs
Primary Topic
Peer revision
Not peer-reviewed
Publication Status
Published
Eprint ID
40337
Cite as
Wehkamp, M. (2014): Stereophotography in marine Biology: Development, Application and Evaluation. , PhD thesis,


Share

Research Platforms

Campaigns


Actions
Edit Item Edit Item