Sex and age determination in European lobsters using AI-Enhanced bioacoustics

Monitoring aquatic species presents considerable challenges due to their elusive nature and complex habitats. Consequently, the development and application of innovative, non-invasive approaches, such as Passive Acoustic Monitoring (PAM), are paramount for effective ecological assessment and management. This study addresses these challenges by focusing on the European lobster (Homarus Gammarus), a key representative species of rocky benthic environments that underpins valuable local fisheries and aquaculture ventures. Comprehensive understanding of lobster habitats, welfare, reproduction, sex, and age is critical for robust aquaculture management, ecological research, conservation strategies and sustainable fisheries. While bioacoustic emissions have been successfully employed to classify various aquatic species using Artificial Intelligence (AI) models, such as fish, this research specifically leverages lobster bioacoustics to classify Homarus Gammarus by age group (juvenile and adult) and sex (male and female). Despite lacking vocal cords, different types of lobsters produce a variety of characteristic sounds, such as stridulation characterised by the European spiny lobster (Panulirus elephas) and Caribbean spiny lobster (Panulirus argus), buzzing or carapace vibrations expressed by the Homarus Gammarus and American lobster (Homarus Americanus), rattling sounds marked by the Tropical spiny lobster (Panulirus ornatus) and clicking or snapping sounds. These varied lobster sounds are amenable to classification using advanced computational AI models. The dataset collection was carried at Johnshaven in Scotland at the local Lobster Shop of Murray McBay and Company. Hydrophones were used and installed underwater in concrete tanks to record the lobster sounds. This investigation explores the efficacy of Deep Learning (DL) models (specifically One-Dimensional Convolutional Neural Networks (1D-CNN) and One-Dimensional Deep Convolutional Neural Networks (1D-DCNN) with varying hidden layers) and six commonly employed Machine Learning (ML) models (Support Vector Machine (SVM), k-Nearest Neighbors (KNN), Naive Bayes (NB), Random Forest (RF), Extreme Gradient Boosting (XGBOOST), and Multi Layer Perceptron (MLP) in classifying buzzing sounds or carapace vibrations of Homarus Gammarus for age and sex classification. The Mel Frequency Cepstral Coefficients (MFCC) served as features for all models across four distinct datasets. MFCC has been reported to extract reliable features from audio signals. The majority of models achieved classification accuracies exceeding 97% for adult versus juvenile differentiation, with the exception of NB, which attained 91.31%. For sex classification, all models except NB surpassed 93.23% accuracy. These compelling results underscore the substantial potential of supervised ML and DL models to extract age-and sex-related features from lobster sounds. Ultimately, this research offers a promising non-invasive approach for lobster conservation, detection, and management within aquaculture and fisheries, thereby enabling real-world edge computing applications for PAM of underwater species.