There is already a boom of artificial intelligence technologies for pig farming in China after the concept was introduced early last year. In the China Animal Husbandry Expo held in May 2019, more than 10 tech firms launched AI technologies or products. Many traditional equipment manufacturers also highlighted a concept of intelligence in their showcases.
As the pioneer of AI farming, Tequ, a major pig producer in Sichuan province, and its partner Aliyun, the cloud computing unit of Alibaba, launched their first generation of a ‘material’ product. This product is a road-rail robot that can automatically monitor the image, voice and temperature of sows in an equipped house.
The system has been patented for its intelligent data collection methods. Zhenghu is the newly-established tech arm for Tequ’s AI pig farming program. Zhang Yong, Vice President of the company, said the road-rail robot has taken advantage of many hardware suppliers to optimize its cost, and it has been commercially applied in some of Tequ’s own farms and contract farms. Tequ and Aliyun planned total 11 application scenarios or critical control points when they launched the AI program a year ago, such as intelligent mating for breeders and automatic weighing for commercial stocks. According to Mr Zhang, their AI program is now mainly focused on the ‘pain points’ of the industry. They are automatic data collection, precise feeding and disease warning.
Take the precise feeding for example, the two cameras on both ends of the road-rail robot can collect and monitor the data of litter size, feed intake and body condition of sows, before deciding on a feed input that is precise to the animals’ demand, so as to ensure the sows have enough milk, thereby increasing the survival rate of newborn and weaning piglets.
“There has been a very remarkable decline in the death rate of piglets,” Mr Zhang said, “The PSY level needs a longer period to calculate, and I think it will be not bad.” The goal of Tequ and Aliyun’s AI program is to raise the PSY by 3 heads. Currently, the average level stands at about 18 in China.
Actually, most tech firms started their AI applications with road-rail robots, though the industry had paid much attention to the so-called pig facial recognition technology that requires a multi-point layout of cameras in the pig house.
For Tequ, they now use a tattoo code on the bodies of sows for recognition. “Compared with the cost of rail tracks, the reduced cameras make the business more economical and efficient, and also reduce the trouble of remodeling pig houses,” Mr Zhang explained. According to a white book on intelligent animal husbandry issued by the China Animal Agriculture Association, there are still three major hurdles in the development China’s AI pig farming: a low standardization level of agricultural IoT, the lack of practicability and cost efficiency of intelligent sensors, as well as the underutilized data.