In this farm, the use of artificial intelligence technology has reduced its manpower by two-thirds. At the tool level, intelligence can help solve human problems; At the quality level, problems can be solved intelligently.
The wind of AI has blown to animal husbandry. In the past two years, the news of AI raising cattle, pigs and geese has appeared one after another. Recently, it was reported that the China Animal Husbandry Association has officially established an intelligent animal husbandry branch. When the branch was established, Wang junxun, deputy director of the animal husbandry and Veterinary Bureau of the Ministry of agriculture and rural areas, said, “in the era of digital economy, Chinese agricultural enterprises are facing great opportunities and challenges. It is an industry problem for animal husbandry enterprises to stop or boldly try digital transformation.”
Low production efficiency, high cost, small profit space and low mechanization level. These negative labels have been on the domestic animal husbandry industry for a long time. However, with the addition of AI, the primitive and extensive industrial form of domestic aquaculture may usher in butterfly change.
From home to abroad, AI breeding is not fresh
Many people were still impressed by the news of April Fool’s day last year. Tencent, known as the “goose factory”, announced the launch of the goose raising plan, which will also carry out goose face recognition and goose language translation. Although it turned out that this was just an April Fool’s joke, it technology providers and terminal equipment manufacturers have indeed begun to show their skills in vast rural areas.
At the beginning of last year, Alibaba cloud started the business of intelligent pig breeding. Technicians marked each pig in the cooperative pig farm with a digital ID tag and established a life-cycle data file including age, weight, eating, exercise frequency, abnormal signs, pregnancy to delivery, etc. around the digital ID. its et agricultural brain can not only help pigs see doctors and health care, but also help sows have more babies.
In 2017, the company launched the “iwa-t” to measure the temperature of dairy cows in a timely manner, based on the specific practice of “iwa-t” and “iwa-t”, which can be used to control the pulse of dairy cows during the annual production period in China.
Not only at home, AI technology also has a precedent in animal husbandry and breeding abroad. Moreover, due to its high degree of large-scale animal husbandry and good digital foundation, the landing of AI technology is relatively faster.
Connecterra, an agricultural technology company based in the Netherlands, has developed the “intelligent dairy farmer’s assistant” (IDA) system in combination with AI technology. Wearable devices are worn on the cow’s neck. These devices have multiple sensors built in, and the supporting analysis software uses machine learning technology. The software and hardware cooperate to monitor the health of cows in real time. According to the introduction, IDA can know whether a cow is ruminating, lying down, walking, drinking water, etc. through the data, judge whether the cow is sick and ready to breed, and inform the rancher of relevant behavior changes, which greatly liberates people. A ranch manager in Georgia who uses the IDA system says productivity can be increased by 10% through IDA.
Agriwebb, founded in 2013 and headquartered in Australia, is also a company that helps animal husbandry enterprises improve production efficiency by providing intelligent services. It systematically collects and records the number, activity status, weight gain and loss, pregnancy rate and other data of livestock in the pasture, as well as the spraying, fertilization and sowing of the pasture.
From 20 to 23, the breeding dividend is very real
The Chinese eat more than half of the world’s pigs almost every year. However, according to the national pig production development plan (2016-2020), the cost of pig breeding in China is about 40% higher than that in the United States, and the feed consumption per kilogram of weight gain is about 0.5kg higher than that in the EU. The annual supply of commercial pigs for sows is 8 to 10 fewer than that in foreign advanced levels, and the comprehensive competitiveness is significantly lower than that in developed countries.
Last February, Alibaba cloud officially announced the cooperation with Sichuan Tequ group and Dekang group to realize artificial intelligence pig raising through et agricultural brain and improve the survival rate and litter rate of pigs. The project investment is as high as hundreds of millions. According to Zheng bin, general manager of Alibaba cloud smart agriculture, the most important purpose of the project is to improve psy (pigs weaned per sow per year), that is, the number of weaned piglets per sow per year, which is one of the most important indicators to measure the level of pig industry. “The ‘pregnancy diagnosis algorithm’ developed by the team can judge whether the sow is pregnant, and then AI can analyze whether the sow is successful in breeding.” Nian Jun, an Alibaba cloud algorithm engineer who was joked about the “pig man”, explained in detail how the algorithm works: “The algorithm we developed to determine whether a sow is pregnant uses multiple automatic tracking robots and cameras to identify the behavior characteristics of the sow after breeding. These characteristics include the depth of the sow’s sleep, the frequency of its standing, the change of food intake, and whether its eyes are blurred. These are the key characteristics.”
After the piglet is born, the ET agricultural brain will monitor the health status of each piglet through voice recognition technology and infrared temperature measurement technology. In case of abnormality, it can give an early warning in time to ensure the healthy growth of the piglet.
Zheng Bin said that the project applied artificial intelligence technologies such as video image analysis and recognition, living body recognition and speech recognition. “From the beginning of breeding, to sow pregnancy and reproduction, and then to the health monitoring of piglets, the * * * level of AI pig breeding is to do the intelligent optimization of the whole process. Through the application of our various technologies, the number of healthy piglets born to each sow in the pig farm has increased from 20 to 23 every year.” Although it is still far from the advanced level of more than 30, the energy of AI pig breeding has been slowly released.
From 0 to 1, intelligent transformation has just started
According to cunzhai, the big data architect of Alibaba cloud, the core of applying AI to agriculture and animal husbandry is three parts: “first, breeding enterprises have always been in fault with AI, which is a very advanced technology. Compared with other industries, agriculture and animal husbandry is an industry with backward informatization.”
Zheng bin also said that at present, the problem of applying AI in animal husbandry is the lack of data. This problem is not only faced by pig breeding, but a common problem in animal husbandry. “The research we are carrying out now includes two levels: tool intelligence and decision intelligence. The former mainly refers to the behavior monitoring and abnormal pig monitoring of pigs in pig farms through various cameras, thermometers, heartbeat monitors, audio processing and algorithms instead of human monitoring; the latter includes the judgment of diseases and the recommendation of treatment schemes, the capture of estrous behavior before reproduction, and the analysis and judgment.” According to him, most attempts to apply digital technology in domestic animal husbandry are still at the level of tool intelligence such as new technology breeding, and the direction of decision-making intelligence is still in the initial stage from 0 to 1.
Cunzhai said that the second problem is that in the traditional animal husbandry model, especially in planting, its own profits are not enough to support the intelligent transformation. “If you only make one algorithm and say you want productivity, this business model itself is not tenable. People’s profits are nine points, and it is unrealistic to spend millions on algorithms. Governments and scientific research institutions all face this problem when promoting intelligent technology.” Cunzhai said.
However, their attempts have also received a positive response. According to Zheng bin, at present, many pig breeding enterprises in China have actively contacted them in the hope of using intelligent pig breeding technology to improve productivity and profits and liberate the labor force. He said that in the cooperative pig farms, the use of artificial intelligence technology has reduced their manpower by two-thirds. “At the tool level, intelligence can help solve human problems; at the analysis level, intelligence can solve quality problems.” Zheng bin concluded.
Cunzhai cited an example. In the pig farm where he provided technical services, in the past, everyone might have to go to the site every two days to collect all the data, or take paper records, and then guide the paper into the form and then into the system. It is very complex. But now there are intelligent devices and all data are collected and entered automatically. “With the improvement of general living standards, fewer and fewer young people are willing to work in farms. AI assistants are bound to play a greater role in the breeding industry in the future.” Cunzhai said.
Responsible editor: CT