9:45 AM - 10:00 AM
[MIS21-04] Observation of litter in city area using dashcam and mobile app
Keywords:AI, Dashcams, Marine Litter, Plastic
Marine litter is a global environmental issue. Most of it is generated in urban areas and carried to the ocean through rivers. Yokosuka City has introduced littering regulations and supports cleaning activities, but detailed data on the distribution and amount of litter is lacking. Effective measures require analysis based on data. The purpose of this study is to visualize the distribution of litter generated within Yokosuka City.
Because it is difficult to observe litter by simply conducting visual inspections, we used video recordings from dashboard cameras on buses running in Yokosuka City. Over about seven months, we collected videos from three buses and used AI to analyze the distribution of litter. As a complementary approach, we analyzed the types and characteristics of litter by using posts from the mobile app 'Pirika,' analyzing videos of riverbanks taken with smartphones, and conducting visual inspections.
The survey found that there was a lot of litter in areas with many people, such as big train stations and bus stops. There was also a lot of litter in places like tunnels, mountain areas, and wide roads with many lanes, where it is harder for people to clean up. Also, the amount of litter increased on days when a typhoon was approaching, suggesting that weather conditions might affect the amount of litter. According to posts on "Pirika," cigarette butts made up 87% of the litter found around the station. According to the survey along the river, it was found that litter tends to be abundant in places that are difficult to access, such as around vacant lots, abandoned buildings, and along railroads. It was also found that the walls of the sidewalks along the river have a certain effect in stopping litter from falling into the river.
These results provide useful information for improving litter collection management and raising awareness among citizens. They also show the potential to improve urban environments and reduce marine litter. Efficient monitoring of litter distribution can be expected by using dash cams and mobile apps. Furthermore, by accumulating data over the long term, the impact of human flow and weather conditions on litter generation will become clearer. In the future, we hope to analyze video recordings from monitoring cameras set up along the coast to study how litter levels change after community cleaning activities. This will help us understand the effect of litter runoff and reduction.
Because it is difficult to observe litter by simply conducting visual inspections, we used video recordings from dashboard cameras on buses running in Yokosuka City. Over about seven months, we collected videos from three buses and used AI to analyze the distribution of litter. As a complementary approach, we analyzed the types and characteristics of litter by using posts from the mobile app 'Pirika,' analyzing videos of riverbanks taken with smartphones, and conducting visual inspections.
The survey found that there was a lot of litter in areas with many people, such as big train stations and bus stops. There was also a lot of litter in places like tunnels, mountain areas, and wide roads with many lanes, where it is harder for people to clean up. Also, the amount of litter increased on days when a typhoon was approaching, suggesting that weather conditions might affect the amount of litter. According to posts on "Pirika," cigarette butts made up 87% of the litter found around the station. According to the survey along the river, it was found that litter tends to be abundant in places that are difficult to access, such as around vacant lots, abandoned buildings, and along railroads. It was also found that the walls of the sidewalks along the river have a certain effect in stopping litter from falling into the river.
These results provide useful information for improving litter collection management and raising awareness among citizens. They also show the potential to improve urban environments and reduce marine litter. Efficient monitoring of litter distribution can be expected by using dash cams and mobile apps. Furthermore, by accumulating data over the long term, the impact of human flow and weather conditions on litter generation will become clearer. In the future, we hope to analyze video recordings from monitoring cameras set up along the coast to study how litter levels change after community cleaning activities. This will help us understand the effect of litter runoff and reduction.