Accurate diagnosis of both the stage of embryonic mortality and the hatch process in incubated eggs is a
fundamental component in troubleshooting and hatchery management. However, traditional methods disturb incubation, destroy egg samples, risk contamination, are time and labour-intensive and require specialist
knowledge and training. Therefore, a new method to accurately detect embryonic mortality and hatching
time would be of significant interest for the poultry industry if it could be done quickly, cheaply and be fully
integrated into the process. In this study we have continuously measured individual eggshell temperatures
and the corresponding micro-environmental air temperatures throughout the 21 days of incubation using
standard low-cost temperature sensors. Moreover, we have quantified the thermal interaction between eggs
and air by calculating thermal profile changes (temperature drop time, drop length and drop magnitude) that
allowed us to detect four categories of egg status (infertile/early death, middle death, late death and hatch)
during incubation. A decision tree induction classification model accurately (93.3%) predicted the status of
105 sampled eggs in comparison to the classical hatch residue breakout analyses. With this study we have
provided a major contribution to the optimisation of incubation processes by introducing an alternative
method for the currently practiced hatch residue breakout analyses.