Why is Shanghai using surveillance cameras and AI to monitor its autumn trees?

Fallen leaves have been left on selected landscape roads from November through December in Shanghai since 2014, and have become an iconic autumnal attraction, with residents and visitors savouring the season’s visual poetry.

But the campaign also attracted criticism, with the burden for managing the leaf litter falling on the city’s sanitation workers.

A Sixth Tone report in 2020 said the ambiguous cleaning standards had added to the workload of sanitary personnel who had to “spend more time sprucing up the fallen leaves”.

Shanghai’s low-level government officials were in a dilemma: to sweep or not to sweep?

The answer lay in refined urban management – a sophisticated policy adopted by the city in 2018 that seeks to solve complex issues by breaking them into smaller problems and dealing with them in detail.

Enterprising city managers turned to the hi-tech solution, harnessing more than 20 highly sensitive cameras since August to capture regular images of the changing canopies.

The resulting “defoliation index model” not only more accurately predicts the volume of leaf fall, it also helps the city’s sanitation workers maintain the number of fallen leaves at a manageable level.

According to Communist Party mouthpiece People’s Daily, the index can also support autumnal street cleaning arrangements by suggesting personnel and vehicle deployments, although no details were provided about how this works in practice.

Liu Jialin, a professor with the Shanghai Academy of Landscape Architecture Science and Planning, told Xinmin Evening News that an AI algorithm calculated the index based on data collected through the cameras and manual on-site observations.

‘Chinese-style modernisation’: tech innovation in focus as Xi visits Shanghai

Changes in the greenness of the canopies were used by the algorithm to calculate the volume of leaves to expect, with the larger relative value of the index indicating that more leaves had fallen, he said.

So far, the defoliation index is applied on only the nine designated roads, where double bougainvillea is the main species. The index is not one-size-fits-all, with fallen leaf volumes dependent on temperatures as well as the characteristics of different trees.

“In the future, this technology will be given more applications beyond the expansion on tree species, but also the measurement of other meteorological changes, such as tracking willow catkins or flower blossoms,” Liu said.

Before the canopy monitoring initiative was adopted, anticipating the volume of leaf litter and how best to deploy sanitation resources largely depended on the experience of the department, with no uniform rules.

According to China Meteorological News Press, Xiao Ming, vice-director of the city transport centre in Shanghai’s Jingan district, said the index “is tailored to meet the needs of refined urban management and fulfil the desire of citizens to enjoy fallen leaves landscapes”.