Assessment of Egg Quality and Hatchability using Hyperspectral Imaging

快速有效地评估鸡蛋质量的能力是鸡蛋包装厂和鸡蛋加工厂的重要目标。目前广泛使用的烛光检查技术可能是费力的，并且不适用于在线实现。不能受精的蛋如果被及时识别和移除，会给孵化场带来严重的安全和经济问题。因此，开发一种有效的、非破坏性的和准确的方法来评估质量并检测蛋的生育力和孵化率，特别是在转移阶段，将对家禽业非常有利。在拟议的研究中，高光谱成像技术的潜力将被开发用于鸡蛋质量评估和孵化率预测的技术。该研究的目标是获得高光谱图像的鸡蛋在维斯/近红外波长，并将这些图像与相应的变化，在各种内部鸡蛋的质量属性相关的蛋白和蛋黄的质量，存在的血液或肉斑点，新鲜度参数（霍夫单位和pH值），生育力和孵化率的完整的鸡蛋。将使用统计方法和先进的人工智能技术开发预测模型和算法，对鸡蛋进行分类，并预测鸡蛋的生育力和孵化率。新的高光谱成像系统将基于使用更宽的光谱范围（400至1700 nm）收集的信息，而不是目前文献中报道的鸡蛋质量评估。将开发的技术将大大减少检测胚胎发育和评估鸡蛋质量，生育力和孵化率的主观性，它也将取代目前用于确定各种鸡蛋质量属性的高度耗时和繁琐的烛光检查方法。该研究团队的专业知识涵盖食品工程、高光谱成像和家禽科学。将培训高素质的人员。工业合作伙伴处于独特的地位，可以促进将技术引入工业。

The ability to assess egg quality rapidly and effectively is an important goal of egg packing houses and egg processing plants. The candling technique that is currently widely used can be laborious and not adaptable to on-line implementation. In-fertile eggs that are identified and removed on time can pose significant safety and economic problems for hatcheries. Thus, the development of an efficient, non-destructive and accurate method for assessing quality and detect fertility and the hatchability of eggs especially at the transfer stage would be very advantageous to the poultry industry. In the proposed research, the potentials of hyperspectral imaging techniques will be exploited to develop techniques for egg quality assessment and prediction of hatchability. The goals of the study are to obtain hyperspectral images of egg at the VIS/NIR wavelengths and to correlate these images with corresponding changes in various internal egg quality attributes related to albumen and yolk quality, and the presence of blood or meat spots, freshness parameters (Haugh Units and pH), fertility and hatchability of intact eggs. Predictive models and algorithms will be developed using statistical methods and advanced artificial intelligence based techniques to classify eggs and predict egg fertility and hatchability. The novel hyperspectral imaging system will be based on the information collected using wider range of spectrum (400 to 1700 nm) than has currently been reported in the literature for egg quality evaluation. The technology that will be developed will greatly reduce subjectivity in detecting embryo development and in assessing quality, fertility and hatchability of eggs and it will also replace the highly time consuming and tedious candling method currently used to determine various egg quality attributes. The research team has expertise covering food engineering, hyperspectral imaging and poultry science. Highly qualified personnel will be trained. The industrial partner is in a unique position to facilitate introduction of the technology to the industry.