Clément Souvignet

Clément Souvignet

Auto-generated profile, based on publicly available information

Clément Souvignet


Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
12 January 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
LNLuiz Claudio Navarro
Luiz Claudio Navarro
LRLucas Nogueira Roberto
Lucas Nogueira Roberto
CSClément Souvignet
Clément Souvignet
TCThaís Franco Cavalheri
Thaís Franco Cavalheri
CECarlos Espirito-Santo
Carlos Espirito-Santo
Rodrigo Otávio de Almeida OzorioRodrigo Otávio de Almeida Ozorio
Rodrigo Otávio de Almeida Ozorio
Accurate, non-invasive biomass estimation remains a critical challenge in aquaculture, hindered by complex optical conditions (e.g., turbidity, foam) and stress from physical sampling. This work validates a novel above-water stereo imaging methodology for precise fish size estimation in shallow or turbid land-based environments where immersed cameras are impractical. The system demonstrated high accuracy: achieving a Mean Absolute Percentage Error (MAPE) of 1.2% for median total length (TL) and 4.9% for median weight in rainbow trout (Oncorhynchus mykiss), and 2.3\% for median TL and 6.9\% for median weight in European Seabass (Dicentrarchus labrax). This accuracy is comparable to submerged systems but achieved without submersion, avoiding biofouling and waterproofing maintenance. It outperforms typical monocular or lab-based methods relying on invasive handling. The workflow features five innovations: (1) coordinate system realignment orthogonal to the water surface; (2) a computationally efficient two-dimensional refraction correction model; (3) a segmentation strategy derived from iteratively generated training datasets for robust detection under occlusion; (4) Voronoi-based centerline extraction for curved fish; and (5) morphological filtering for geometric consistency. This non-invasive approach enhances animal welfare by eliminating physical handling, injury risk, and pathogen transmission. Ultimately, this methodology serves as a robust kernel for system tools driving optimized growth modeling, grading, and harvest planning, thereby advancing Precision Fish Farming (Aquaculture 4.0) towards improving profitability, sustainability, and animal welfare.