An automated weight collection and prediction tool for poultry was successfully tested in conjunction with pan-Asia industrial agricultural and processed food company, Japfa.
“We would weigh the chickens manually and then compare the mean weight from the manual data set to the mean weight from the automated data set to establish baseline accuracy,” Giovanni dos Reis Montefiori, founder of Beakbook, explained. “We’ve now proven that the technology can work.”
The poultry weight monitoring technology integrates machine learning and algorithms to help track and analyze feed conversion and efficiency. It could increase the amount of weight-data collected by 33-fold over manual methods, resulting in up to a 7% decrease in production costs through more efficient farm management, Montefiori added.
More accurate weight monitoring helps farmers make decisions using the most up-to-date information, helping producers find the most efficient and environmentally sustainable ways to feed poultry to optimize growth.
Forecasting can also allow them to react more quickly to red flags, such as illness, through forecasting algorithms. This helps maintain high flock health and well-being by reducing periods of suboptimal growth patterns.
Results from the trial
According to a press release, the weighing technology demonstrated 97% accuracy for mean weight over a nine-month period in the validation phase and predicted the mean weight of chickens seven days in advance with a prediction accuracy of +97%.
The pilot study lasted nine months and involved approximately 280,000 broilers. Japfa, which produces upwards of 600-million-day-old chicks per year, and Beakbook are currently exploring future collaborations.
“The pilot study met all expectations and now we’re riding this momentum,” Montefiori said.
The technology was also recently named a finalist in the Nutreco 2022 Feed & Food Tech Challenge/ Beakbook expects a wider rollout of the automated weight collection and prediction technology in early 2023, followed by an industry-wide rollout in mid-2023.
