A Small-Scale Comparison of Iceland Scallop Size Distributions Obtained from a Camera Based Autonomous Underwater Vehicle and Dredge Survey
An approach is developed to estimate size of Iceland scallop shells from AUV photos. A small-scale camera based AUV survey of Iceland scallops was conducted at a defined site off West Iceland. Prior to height estimation of the identified shells, the distortions introduced by the vehicle orientation...
| Published in: | PLoS ONE |
|---|---|
| Main Authors: | , , |
| Format: | Text |
| Language: | English |
| Published: |
Public Library of Science
2014
|
| Subjects: | |
| Online Access: | http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4193778 http://www.ncbi.nlm.nih.gov/pubmed/25303243 https://doi.org/10.1371/journal.pone.0109369 |
| Summary: | An approach is developed to estimate size of Iceland scallop shells from AUV photos. A small-scale camera based AUV survey of Iceland scallops was conducted at a defined site off West Iceland. Prior to height estimation of the identified shells, the distortions introduced by the vehicle orientation and the camera lens were corrected. The average AUV pitch and roll was and deg that resulted in error in ground distance rendering these effects negligible. A quadratic polynomial model was identified for lens distortion correction. This model successfully predicted a theoretical grid from a frame photographed underwater, representing the inherent lens distortion. The predicted shell heights were scaled for the distance from the bottom at which the photos were taken. This approach was validated by height estimation of scallops of known sizes. An underestimation of approximately cm was seen, which could be attributed to pixel error, where each pixel represented cm. After correcting for this difference the estimated heights ranged from cm. A comparison of the height-distribution from a small-scale dredge survey carried out in the vicinity showed non-overlapping peaks in size distribution, with scallops of a broader size range visible in the AUV survey. Further investigations are necessary to evaluate any underlying bias and to validate how representative these surveys are of the true population. The low resolution images made identification of smaller scallops difficult. Overall, the observations of very few small scallops in both surveys could be attributed to low recruitment levels in the recent years due to the known scallop parasite outbreak in the region. |
|---|