Non-intrusive optical measurement of egg geometry and volume for macro- and microscopic ovoids

Egg size is a determining trait in the survival and development of bird and arthropod offspring. Ecologists studying these animal groups need reliable and practical methods to measure egg geometries. Despite the existence of numerous egg measurement techniques, the current scientific literature lacks a non-intrusive method allowing for field measurements without disrupting the measured individuals. Moreover, there has been so far no systematic comparison between the different available methodologies, making it difficult to assess their strengths and weaknesses. This study proposes a new and non-intrusive method for retrieving ovoid-shaped egg geometries, consisting in optimizing a parametric virtual model from N camera views (N-Views method). We tested the N-Views method on chicken (Gallus gallus domesticus) and quail (Coturnix coturnix) eggs, comparing its accuracy with 2 other existing optical measurement methods (Photogrammetry and Silhouette). Across all tested methods, N-Views provided the best estimate of egg volume when compared with the reference manual measurements (Archimedean buoyancy force), resulting in a MAE of 125 mm3 and 1.03% and 0.28% relative error for Coturnix and Gallus eggs, respectively. We further quantified the sensitivity of the N-Views and Silhouette methods to error sources arising from perspective and egg-camera alignment, highlighting the characteristics needed in optical measurement protocols for the estimation of egg geometries. Finally, we discuss the transferability of the N-Views method to the size estimation of microscopic arthropod eggs. This novel approach opens a new door for the non-intrusive exploration of relationships between egg size and life history traits in avian and arthropod species.