China News Service, Shanghai, October 12 (Reporter Chen Jing) The 26th China International Industrial Fair (hereinafter referred to as: Industry Fair) opened in Shanghai on the 12th. The "Qingshi Intelligent Drone Inspection Software System" (hereinafter referred to as the "Qingshi System") developed by the university team and the autonomous desert planting robot that coordinates air, space and ground have attracted a lot of attention.
Traditional drones have mature functions such as automatic takeoff, return, high-definition shooting, and real-time image transmission. However, when it comes to urban inspection sites, "taking photos" is still several steps away from "solving the problem." It is understood that the "Qingshi System" developed by the scientific research team of East China Normal University allows multiple drones to plan their own routes, coordinate "team work", and integrate visible light and thermal infrared to "understand" issues such as crowd anomalies, water surface garbage, and river discharges, and accurately place the recognition results on map coordinates, with a positioning accuracy of sub-meter level. Accordingly, the team deeply integrated the spatio-temporal analysis capabilities of GIS (Geographical Information System) with the perception capabilities of AI to form a full-chain technology closed loop of "flight control-positioning-identification-analysis".
Li Xiang, the leader of the R&D team and a professor at the School of Geographical Sciences of East China Normal University, said that what this system needs to do is to equip traditional drones with brains, eyes and maps. Facing buildings, parks or rivers, the system can select routes based on tasks, arrange who goes where and from what angle, reducing repeated flights. The supporting route optimization algorithm library supports single-point orbiting, building orbiting, multi-view surround scanning, vertical sweeping, etc., so that drones no longer "fly in their own way".
According to reports, the "Qingshi System" has been implemented in music festivals attended by nearly 30,000 people, Shanghai's key waters and other scenes, allowing drone inspections to move from "photographable" to "understandable, accurate finding, and capable of handling", and also allowing urban governance to shift from "finding problems everywhere" to "accurate handling according to coordinates."
Currently, the team has formed 4 software copyrights and 4 invention patents around key technologies such as flight control, intelligent identification, and precise positioning. It is like a set of "intelligent building blocks" that can be assembled and upgraded, and can customize different functions according to different tasks.
At the Fudan University booth, the reporter learned that the autonomous desert planting robot that cooperates with air, space and ground can effectively solve the outstanding problems of traditional desert management operations such as difficulty, high labor costs, low operating efficiency, and difficulty in large-scale promotion. Fudan University's Han Desert Management Robot Team innovatively builds an "air-space-ground" integrated collaborative operation system to provide intelligent, unmanned, and scalable systematic solutions for comprehensive desertification management across the country.
It is reported that the desert planting robot system integrates the core functions of autonomous perception, precise navigation, and integrated planting. It relies on Beidou, laser radar, and stereo vision multi-source fusion positioning technology to overcome key technologies such as complex terrain path planning, harsh environment image enhancement, and sand anti-sag control. It is equipped with a fully automatic planting operation mechanism and can adapt to extreme working conditions such as desert sand, low light, and soft terrain. Its operation stability and adaptability are industry-leading.
According to reports, the relevant equipment has completed field operation verification in the Dalate Banner Desert of Ordos, Inner Mongolia, and can achieve normalized and stable operations in the desert. In the future, the bicycle platform will be integrated with four sets of planting mechanisms, and the operating efficiency will be more than 10 times that of traditional manual operations, greatly improving the large-scale operation capability of desert ecological restoration.
