In the vast wheat fields of Punjab, drones are patrolling the air and intelligent systems have replaced manual labour to take care of crops. Free from the pain of field labour during the boiling summer season, local farmers could accurately monitor fertiliser application and irrigation with just a few clicks on their mobile phones or tablets.
"Previously, I could only see which fields performed well when it was almost harvest time. Now, I check crop health regularly on my mobile and can fix problems early, which has helped improve yields and reduce losses," said Muhammad Taj from Gujranwala as he marvelled at the new agricultural artificial intelligence app – Kisan360, newly installed on his phone.
The China-Pakistan cooperation project using computer vision to realise high-throughput crop monitoring for climate-smart agriculture has injected sci-tech element into Punjab's farmland, with Kisan360 being the result of their painstaking efforts.
"We use drones equipped with visual recognition equipment for intelligent management, crop monitoring, cultivation management, pest and disease analysis, and agricultural production, including growth determination, health monitoring, water requirement analysis, fertiliser and pesticide application," said Dr Wu Jun, the Chinese project leader.
Initially, local farmers expressed doubts about smart technology, saying, "It's too complicated for us and its effectiveness and practicality have not been proved." A Pakistani team addressed their doubts by producing text guides and audio-visual materials on vegetation health maps in local dialect and sharing them with farmers in real time.
"I come from a farming family, understanding very well the problems faced by farmers today, with severe heat waves and water shortages becoming our main challenges. Providing farmers with real-time data to help them optimise planting decisions and ensure food security for their communities is urgent," said Saqib Ali, a Pakistani researcher at China-Pakistan Joint Laboratory for Agriculture.
"Our AI app is now live on both Apple App Store and Android Play Store. We have conducted training sessions in Punjab and Islamabad, with a user base of approximately 1,000 farmers," he said.
"Taking fertilisation as an example, Kisan360 uses satellite imagery and artificial intelligence to provide farmers with real-time insights into moisture and nitrogen levels, helping to identify areas that require precise fertilisation. Then farmers can listen to these reports in Urdu language," the researcher added.
Saqib Ali highlighted that a bilateral team had deployed a ChatBot in the app using DeepSeek. As an open-source model, it could be used to fine-tune by using Pakistan-related crop data.
"It is in the testing phase. We are fine-tuning DeepSeek to better understand and address the specific farming needs of Pakistan, incorporating local agricultural knowledge to provide more tailored insights for farmers. DeepSeek has proved to be extremely valuable for us in understanding and meeting the unique challenges of farming."
"Pakistan is currently in urgent need of agricultural practices driven by advanced technology. This project is an opportunity for Pakistan. After promotion, more and more farmers will be interested in it, which can bring substantial changes to local agriculture," said Dr Sultan Habibullah Khan, Director of the Centre for Agricultural and Food Security Research at the University of Agriculture, Faisalabad. Wu said that the World Bank and the Asian Disaster Management Centre had provided funding support for the project and emphasised the importance of the innovative solution in addressing climate challenges and enhancing agricultural sustainability.
"China-Pakistan Agricultural Laboratory plans to promote this technology throughout Pakistan within two years to comprehensively promote sustainable agricultural development through AI+, as a new tool under the new era."
COMMENTS
Comments are moderated and generally will be posted if they are on-topic and not abusive.
For more information, please see our Comments FAQ