Believe it or not, China is the world’s biggest pork producer. The country’s 700 million pigs represent half of the world’s population. Many of those live and are bred on pig farms.
To keep an ever-steady supply of pigs, a pig farmer must keep track of which pigs are pregnant and which have the best chance at getting pregnant. Normally a pig’s behavior, appearance, and health have to be observed for at least 21 days to see if they are carrying another little piggie. To add to the challenge, a pig farmer usually has to monitor up to 100 pigs at one time. Of course, this often leads to errors and misdiagnosis.
Alibaba Using AI and Facial Recognition to Help Pig Farmers
To help pig farmers save time and become more efficient, Alibaba’s Cloud computing arm is looking to use artificial intelligence to help determine which pigs are pregnant and which have the best chance at becoming pregnant. It’s also cutting the monitoring process from 21 days down to only 3 saving valuable time and resources.
Using facial-recognition cameras, Alibaba’s Cloud group has developed an algorithm that can tell if pigs are pregnant by analyzing their appearance, behavior and eating patterns. The algorithm can also detect if a pregnancy has failed as early as three days after mating. Normally this could also take up to 21 days.
AI Cuts Monitoring Period From 21 Days to 3
“Apart from the observation of the behavioral changes, a more thorough check is quite time and energy consuming, requiring help from two to three people. Even if the solution can only tell which pig is more likely to be pregnant, that is a big help for us,” said Deng Qiang, a pig farm worker at Alibaba’s smart agricultural project in Sichuan.
According to Alibaba, the tech will help to increase the number of newborns on a pig farm. In early-stage experiments, the AI was able to help each sow give birth to 3 more piglets a year and reduce the mortality rate by 3%. The program is set to be implemented in some of Alibaba Cloud’s partner farms in China’s Sichuan province early next year.