Mussel Classifier System Based on Morphological Characteristics

Abstract

The recognition, counting, and sorting of mussels in marine cultures for seed production are currently performed by visual examination experts (i.e., entirely dependent on human resources). In this paper, we present the development of an automatic mussel classifier system based on the morphological characteristics for the simultaneous recognition and sorting of five mussel species. The proposed system provides rich statistical information needed for tracking the long-term evolution of culture parameters. In our experimental demonstration, we have achieved a recognition rate of 95% in most of the test probes for the five studied mussel species. A single sample of dozens of specimens can be classified within seconds with real-time capability when the vision interface is not used. Finally, the system has the potential to be extended for the automatic classification of mussels worldwide.

Keywords

• Digital imaging processing
• machine learning
• machine vision
• mussel classification
• real-time classification