Facial recognition technology is being used to identify what type and how many pests there are in the fields of East China's Jiangsu province.

When Zhu Youli, deputy head of Zhenjiang Plant Protection and Inspection Station in the province, opens a special app on his mobile phone, an interface with text and images immediately pops up showing not only the real-time photos of insects, but also detailed data about them.

It is one of the achievements of an intelligent monitoring system for crop diseases and pests developed by the station, and plays a big part in ensuring a good harvest.

"To make disease and pest control more effective, it is necessary to provide more accurate forecasts," he said.

Facial recognition technologies have widely been applied in people's lives on the Chinese mainland, especially for mobile phone unlocking and payment.

"Applying this facial recognition technology to 'insect and worm facial recognition' is a powerful exploration for us to develop an intelligent monitoring system for crop diseases and pests," Zhu said.

Insect facial recognition is a pest detection method based on image recognition and detection technology, which enables machines to automatically identify the types and quantities of pests.

Through filming, uploading, analysis, feedback and other processes, plant protection personnel and farmers can quickly grasp the situation of pests and diseases in farmlands, according to Zhu.

Agricultural experts said that through intelligent induction, the moth population of rice-leaf rollers in the fields in local townships and villages can be monitored in real time.

Meanwhile, the peak of egg hatching and the population of the next generation can be predicted based on the number of insect eggs in the field, rice growth period and meteorological conditions.

The plant protection and inspection station is mainly responsible for monitoring plant diseases and pests within its area.

"Previously, we had to pick up tools and go deep into the fields every morning at about 5 am, shake plants on the edge of the field to startle the insects, then estimate the number of pests through observation," Zhu said.

"This method has significant errors in identifying the number and variety of pests."

A rural township crop protection and inspection station often has only three to four local staff available for testing the types and quantities of insects across hundreds of square kilometers of farmland, he said.

Zhenjiang city began to introduce intelligent detection equipment for crop disease and pest control in 2019.

"The intelligent detection equipment has greatly improved our efficiency," he said.

"Now, instead of picking up tools and going out to the fields, our farmers' daily habits are just opening mobile apps to view real-time data."

Intelligent equipment is not only highly efficiency, but can also refine data.

"Through the intelligent equipment, we can accurately provide pest warnings to specific towns and villages, which is more conducive to local pest control," Zhu said, adding that more than 30 townships in Zhenjiang have already used this type of equipment.

"At the moment the intelligent equipment and devices are mainly used for detecting rice related pests, but we plan to expand them to tea planting next year," he said.

However, some pests look highly similar, which makes it difficult to distinguish and classify them, and some are so small that they are difficult to spot. Moreover, backlighting and shadows caused by shooting techniques may increase the difficulty of recognition.

"In this case, manual monitoring is needed to make corrections," he said.

In order to safeguard food security, agricultural and related departments actively promote the intelligent monitoring of pests and diseases through administrative measures, coordinated funding and strengthened training programs, he added.

Zhuang Jintang, a farmer in Zhenjiang's Shangdang township, said the use of intelligent recognition has greatly improved his planting efficiency.

Zhuang and his wife grow rice and wheat on an 18.6 hectare field, with an annual output of about 196 metric tons.

Previously, he had to manually drive away insects from his fields, and the types and quantities of insects had to be determined based on observation, Zhuang said. During the busy farming season, two or three people must be hired to help do the work, he added.

"Last year, after we introduced intelligent equipment and devices that can recognize pests, we could accurately grasp both the variety and quantity of pests, which made the work much simpler," Zhuang said.

"With the accurate data, we can manage chemicals more precisely. And with the drones for spraying, the entire process from disease and pest monitoring to management can be completed on my mobile phone."