Line-scan imaging analysis for rapid viability evaluation of white-fertilized-egg embryos

The process of selecting the fertilized eggs and discarding the nonfertilized or poorly fertilized ones is critical for the efficient egg production. Egg-fertility evaluation is generally done by the candling method, in which the egg is illuminated by a light source and the fertility is confirmed, based on the blood vessels; this method requires expertise and cannot be automated. In this study, a hyperspectral visible/near infrared line-scan imaging system, which utilizes a customized illumination system to determine the viability of fertilized eggs is used in the semi-transmittance mode. The line-scan hyperspectral imaging provided a novel unique patterned image of viable eggs due to the movement of viable embryo in the light illumination environment. The acquired image of each egg (88 eggs) was analyzed using principal component analysis (PCA) and image processing including a texture co-occurrence method. The generated PCA-core images exhibit significant differences between the blood vessels of the fertilized viable and nonviable eggs. The analysis result indicates that by utilizing a single-band (560 nm) image and simple image processing methods, the viability of fertilized eggs can be determined with an accuracy of approximately 99%. A single-band-image-based low-cost system is envisaged in the near-future for the automation of the viability-status evaluation of fertilized eggs.