RESUMO
Acoustic cameras are increasingly used in monitoring studies of diadromous fish populations, even though analyzing them is time-consuming. In complex in situ contexts, anguilliform fish may be especially difficult to identify automatically using acoustic camera data because the undulation of their body frequently results in fragmented targets. Our study aimed to develop a method based on a succession of computer vision techniques, in order to automatically detect, identify and count anguilliform fish using data from multiple models of acoustic cameras. Indeed, several models of cameras, owning specific technical characteristics, are used to monitor fish populations, causing major differences in the recorded data shapes and resolutions. The method was applied to two large datasets recorded at two distinct monitoring sites with populations of European eels with different length distributions. The method yielded promising results for large eels, with more than 75% of eels automatically identified successfully using datasets from ARIS and BlueView cameras. However, only 42% of eels shorter than 60 cm were detected, with the best model performances observed for detection ranges of 4-9 m. Although improvements are required to compensate for fish-length limitations, our cross-camera method is promising for automatically detecting and counting large eels in long-term monitoring studies in complex environments.
Assuntos
Acústica , Anguilla , Animais , Projetos de PesquisaRESUMO
Total lengths (LT ) of 50 free-swimming fish in a tank, silver carp Hypophthalmichthys molitrix and rainbow trout Oncorhynchus mykiss, were measured using a DIDSON (Dual-frequency IDentification SONar) camera. Using Sound Metrics software, multiple measurements of each fish (LT , side aspect angle and distance from the camera) at different times were analysed by two experienced operators while a subset of data was analysed by two inexperienced operators. The main result showed high variability in intra-fish LT measurements. The number of measurements required to minimise errors and to obtain robust fish measurements (true LT ± 3 cm) was estimated by a bootstrap method. Three to five measurements per fish were recommended for fish surveys in rivers. In this experimental study, aiming to reproduce river conditions, no evidence of fish position (side aspect angle and distance from the camera) effect was detected, but an operator effect (partially explained by training) was observed. General linear mixed models also showed that lengths of the smallest fish (LT < 57 cm) were overestimated and lengths of the largest fish (LT > 57 cm) were underestimated in comparison with their true lengths. In conclusion, we highlight that this technology, like any monitoring methods, returns imperfect observations. We advise DIDSON users to ensure that measurements are carried out correctly in order to draw accurate conclusion from this new technology.