It was the third day of a very different type of class. Students gathered outside, opened up their laptops, and chatted about artificial intelligence (AI), object recognition, and aerial robotics. Then their drones took to the air.
More than 60 budding software engineers and technology enthusiasts from Chinese University of Hong Kong (CUHK), the University of Hong Kong (HKU) and Hong Kong University of Science and Technology (HKUST) came together for the territory’s inaugural AI x Drones Joint University Competition – an event organized by leading drone technology maker DJI and Microsoft Hong Kong.
Their collective aim was to explore a high-flying future by combining the versatility of drone technology with the computational power of AI.
Earlier, the students spent two days in workshops where they were versed in computer learning and object recognition with Microsoft’s Custom Vision solution and Azure cloud computing platform.
After forming 15 teams, they were assigned a DJI Phantom 4 Pro drone. They were given flying lessons and also tasked with training it to recognize different objects using AI.
On competition day, each team was set a common challenge: maneuver their pre-trained drone over an area where different pieces of fruit had been placed, then operate the craft’s camera so that AI technology could recognize each as an apple, a pear, a pomelo, a banana, or a mango.
The teams were ranked and scored on the number accurately identified as well as physical drone flying skills and completion time.
“The whole experience was eye-opening,” said Liang Zhijian, the captain of the winning team from the Hong Kong University of Science and Technology. “I’d heard the term ‘AI’ in other classes before. But while I understood it was an important technology for the future, it only seemed like a distant topic for us. With this competition, we were able to have hands-on experience. It was fascinating to learn about object recognition and be part of the real-world developer community, such as using GitHub, cloud, and tinkering with a drone.”
Susanto Christopher Alvin from The University of Hong Kong said: “This was our first time flying a drone and we also needed to train our drone to recognize objects. The process was fun, and at the same time challenging because there were quite a number of different fruits and they were placed closely with each other. This can make it confusing for the AI system. I’m glad that we trained our system thoroughly by inputting a lot of related images to the AI system ahead of the competition – this helped us to achieve a good result.”
Last year, DJI and Microsoft formed a global strategic partnership to bring advanced AI and machine-learning capabilities to DJI drones. As AI and robotics become critical skills and fundamentals for the next digital era, the two global technology leaders have joined hands to nurture the brightest local engineers and roboticists of tomorrow via the provision of learning platforms and access to these cutting-edge technologies.
“We are delighted to partner with Microsoft on this meaningful project. Not only is this competition an opportunity for young engineers to learn about our latest aerial and AI technologies, it’s also a chance for them to create something real and explore applications that could be used in the real world,” said Christina Zhang, Senior Director, Corporate Strategy & Communication at DJI.
“What the students have demonstrated is just one of the many possibilities. We look forward to seeing how our combined technologies can help simplify workflows and bring tangible benefits to people and businesses across different industries.”
Winnie Chu, Business Group Lead, Cloud, and Enterprise at Microsoft Hong Kong described the competition as fun, engaging and experiential.
“Combining AI with drones provides a unique opportunity to educate students about new technologies like never before. As computing has immersed in our everyday life, the intelligent edge is emerging as the next technology frontier. Together with DJI, we are combining intelligent cloud, best-in-class aerial technology, and Azure IoT capabilities to devices on the edge, potentially changing the future of various industries such as construction, public safety and more.”
For more highlights about the final competition, visit or .