A novel in situ method for estimation of the carapace length of sheltering spiny lobsters, Panulirus japonicus, via stereo photography

Abstract

We developed a practical and reliable non-invasive method for estimating the carapace length of the spiny lobster, Panulirus japonicus, hidden in shelters. The method estimated the carapace length (CL) from multiple reference body parts measured from images taken using compact stereo cameras, based on predetermined allometric relationships. Two versions of the method to estimate CL were tested, one from only non-sexual reference lengths (Method I) and the other from all including sexually differential reference lengths through determining sex (Method II). Fifteen reference lengths, consisting of 4 distances between characteristic points on the cephalothorax, and 11 segment lengths of appendages, were measured using stereo photogrammetry for 53 spiny lobsters in outdoor tanks, and all except the distance between the frontal horns provided reasonably accurate regression models to estimate CL (relative errors at 95% prediction intervals ranging within ±15%) including sexually differential models for 10 appendage reference lengths. A laboratory test with 39 different lobsters in artificial dens showed that Method II enabled correct sexing for >70 mm CL and provided CL-estimates for 95% of the lobsters with absolute relative errors <10%, while Method I provided a slightly lower acquisition percentage (87%) and precision (maximum absolute relative error = 12%). Field tests demonstrated that, flash photography produced sharp images suitable for image processing without evoking escape behavior, and Method II could achieve a similarly high acquisition percentage (ca 90%) of CL-estimates, validating the practicality of this non-invasive method in the field.