Customers: Singapore National Parks Board (NParks) Product: Artificial intelligence (AI, Artificial intelligence, AI) Second product: Mobile application Third product: IoT Projects Project date: 2022/04 - 2022/10
|
2022: Development of a remote plant care system
About 7 million trees grow on the island of Singapore, and 6 million of them are managed using a monitoring system. This became known on November 28, 2022. The Singapore National Parks Authority (NParks) monitors about 6 million trees so that foresters can manage them using a mobile application as soon as the trees reach a certain size.
The remote tree management system, combined with other digital assessments, allows many tree management tasks to be performed directly from home or office. The initiative will protect staff from working in the sweltering heat as Singapore is constantly in summer and daytime temperatures hover around 33C. In addition, it rains an average of 167 days a year on the island.
Singapore manages trees by creating a digital twin - LiDAR point clouds gather, and then using artificial intelligence, the geolocation of each plant is determined. NParks then applies finite element techniques to the digital twin to study the tree's overall stability in various weather conditions (Singapore has severe tropical storms), given factors such as wood structure, wood strength and root space.
In addition, the company uses satellite sensing for multispectral analysis to determine chlorophyll levels and make sure the tree is still growing and blooming.
The system also includes cameras that take panoramic photos for remote visual checks, as well as physical tilt sensors in older trees to detect any sudden movement that could be potentially dangerous.
If a problem is detected, staff can take measures to improve the structure of the tree or conduct further assessments to determine the life span of the plant.
The development of the high-tech analysis system began 20 years ago with tree geotags and progressed as new technologies emerged. About 5 years ago, tree geotagging began to be done using machine learning. After that, the inventory of trees began to be performed automatically[1].