Open Collections

Fish size estimation in sea cages using a fish image capturing and sizing system

University of British Columbia

Field trials were conducted at three British Columbian Atlantic salmon farms to test the performance of a fish image capturing and sizing system (FICASS) against the presently used seine/dip netting technique. FICASS consisted of underwater video camera equipment combined with an S-VHS videorecorder and image processing equipment. Salmon farmers require a new, noninvasive, sampling method to obtain population average fish weight (kg) (AFWIKE) and standard deviation (kg) (SDWIK), because the current level of accuracy is +/- 20% of actual mean fish weight and is very invasive to the fish. Environmental, husbandry (i.e. grading, sorting, feeding histories), swimming and stocking characteristics were recorded and, whenever possible, related to the sizing results. FICASS was developed by a research group at Dr. Petrell's Fish Imaging Laboratory, Chemical and Bio-Resource Engineering Department, University of British Columbia. In the three cages of one kg-sized fish of low stocking density that formed uncompacted types of aggregations and had not been previously size sorted, the AFWIKEs obtained by FICASS changed with sampling position. In 13 out of 15 cages of fish (> 3 kg) which had been size sorted, the AFWIKEs obtained by FICASS did not significantly change with position. In cages of harvest-sized fish and size-sorted three kg-sized fish, AFWIKEs by FICASS varied with position if the fish were unfed. When 3 kg-sized fish had been previously size sorted, FICASS results agreed with seine/dip netting results in 11 out of 13 cages. When 1 kg-sized fish were tightly schooling together, FICASS results was agreed with seine/dip netting results in 4 out of 6 cages sampled. Generally FICASS results agreed with seine/dip netting results when size un-sorted and unfed fish were videotaped at the top of aggregation (4 of 5 positions). FICASS results agreed with harvest data either when fish had been previously size sorted or fed as scheduled, and when sized-unsorted and unfed fish were videotaped in the middle or bottom of a loose, swimming aggregation. In general, different feeding and sorting schedules, and stocking conditions appeared to promote the incidence of sub-populations characterized by size within a sea cage. Ideally fish sizing should occur when the population size is randomized. FICASS can be used as an effective sizing tool because it can be used to non-invasively detect subpopulations, and if detected, the fish sizing could be repeated at a latter date when fish conditions become more randomized.

Thesis/Dissertation

application/pdf

2009-02-11

For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use.

http://hdl.handle.net/2429/4434

Chemical and Biological Engineering

University of British Columbia

UBCV

DSpace