Indonesian Political, Business & Finance News

Xinjiang deploys 108-arm robot for cotton plant topping

| Source: ANTARA_ID Translated from Indonesian | Technology
Xinjiang deploys 108-arm robot for cotton plant topping
Image: ANTARA_ID

Urumqi (ANTARA) - The device has 108 mechanical arms, moves through cotton fields like a slow-moving tank, and can perform more than 300,000 cuts per hour.

It is not a new weapon from a superhero film, but a real robot built to carry out one of the most exhausting jobs in agriculture: pruning the tips of cotton plants.

The unmanned machine, developed by Chinese engineers, is now operating in the vast cotton fields of the Xinjiang Uighur Autonomous Region in north-western China, the country’s largest cotton-producing area.

Standing 3.8 metres tall and weighing 8 tonnes, the cotton-tipping robot is equipped with advanced sensors and a machine vision system.

According to its manufacturer, Viewer Tech, an agricultural robotics company based in Xinjiang, the machine can work an area of up to 2 hectares per hour, with efficiency roughly 120 times that of manual labour.

Trials have shown that this robotic approach can raise average yields by nearly 600 kg per hectare, with net gains of more than 3,000 yuan (1 yuan = Rp2,643) per hectare, said Liu Hangxing, general manager of Viewer Tech.

For generations, pruning the tips of cotton plants has been synonymous with back-breaking work under the scorching summer sun. Farmers had to walk from row to row, pinching off the tip of each cotton plant by hand to control plant height and suppress the growth of unproductive branches.

The task demands careful judgement: weaker plants require less pruning, whilst stronger plants need more. Manual pruning, previously the only option, often produced uneven results due to differences in workers’ skill and speed, directly affecting the harvest.

The new robot, now in its fourth generation, uses binocular vision cameras and artificial intelligence (AI) algorithms to scan each plant, determine the cutting point within milliseconds, and direct its 108 mechanical arms to make the cut.

From identification to action, the entire process can mimic the hand techniques of experienced cotton farmers.

Xinjiang’s cotton sector is already among the most mechanised in the world, with an overall mechanisation rate for cultivation and harvesting exceeding 97 per cent.

“With this robot, we can achieve full mechanisation and intelligence across the entire cotton growth cycle,” said Zhou Shuai, general manager of Xinmiao Agricultural Technology, a company based in Shihezi, one of the main cotton-producing areas in northern Xinjiang. Founded by a group of entrepreneurs born in the 1990s, the company manages around 133 hectares of cotton fields.

Zhou said he was impressed by the robot’s efficiency, noting that its daily output matched that of 50 to 60 workers. He added that the success rate for tip pruning had exceeded 90 per cent.

Before adopting the robot, Zhou’s company had already automated irrigation, fertilisation, spraying and weeding using the BeiDou satellite positioning system and domestically developed AI. The entire field operation is now managed by just five people.

The cotton-tipping robot is just one example of a broader drive towards agricultural automation across Xinjiang. In recent years, companies, universities and research institutes have stepped up innovation and research and development (R&D) in cotton machinery, focusing on precision, efficiency and intelligence.

Last year, a collaborative effort between Xinjiang University and EAVISION, a robotics technology company based in Xinjiang, successfully developed a laser-based cotton-tipping robot.

“Robotics technology is becoming increasingly mature,” said Pei Xinmin, a researcher at Xinjiang’s agricultural and rural mechanisation development centre.

“Robotics is not just about cotton. This technology provides a replicable model for smart farming of other crops, accelerating the digital transformation of the agricultural sector across Xinjiang,” Pei added.

View JSON | Print