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一个重要目标。目前广泛使用的烛化技术既费力又不适合在线实现。未受精的卵被及时识别和移除会给孵化场带来重大的安全和经济问题。因此，开发一种高效、无损和准确的方法来评估鸡蛋的质量和检测蛋的生育能力和孵化能力，特别是在转移阶段，将对家禽业非常有利。在本研究中，将利用高光谱成像技术的潜力，开发鸡蛋质量评估和孵化率预测技术。本研究的目的是获得鸡蛋在可见光/近红外波段的高光谱图像，并将这些图像与鸡蛋内部各种品质属性的相应变化联系起来，这些属性与蛋白和蛋黄质量、血或肉斑的存在、新鲜度参数（哈夫单位和pH值）、完整鸡蛋的受精率和孵化率有关。预测模型和算法将使用统计方法和先进的基于人工智能的技术来分类卵子并预测卵子的生育能力和孵化能力。新的高光谱成像系统将基于使用更宽的光谱范围（400至1700纳米）收集的信息，而不是目前文献中报道的用于鸡蛋质量评估的信息。即将开发的技术将大大减少检测胚胎发育和评估卵子质量、生育能力和孵化能力的主观性，并将取代目前用于确定各种卵子质量属性的耗时且繁琐的蜡烛法。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。将培养高素质人才。工业合作伙伴在促进将技术引入行业方面处于独特的地位。

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.