Related Experiment Video
Updated: Jul 17, 2026
Automatic Image Processing to Determine the Community Size Structure of Riverine Macroinvertebrates
Published on: January 13, 2023
Quantification of litter in cities using a smartphone application and citizen science in conjunction with deep
Shin'ichiro Kako1, Ryunosuke Muroya2, Daisuke Matsuoka1
1Graduate School of Science and Engineering, Department of Engineering, Ocean Civil Engineering Program, Kagoshima University, 1-21-40 Korimoto, Kagoshima, Kagoshima 890-0065, Japan; Center for Earth Information Science and Technology, Research Institute for Value-Added-Information Generation, Japan Agency for Marine-Earth Science and Technology (JAMSTEC), 3173-25, Showa-machi, Kanazawa-ku, Yokohama, Kanagawa 236-0001, Japan.
Citizen science via the Pirika smartphone app and deep learning effectively quantifies urban street litter. This approach visualizes and characterizes plastic pollution, aiding environmental management.
Area of Science:
- Environmental Science
- Urban Ecology
- Data Science
Background:
- Urban litter, particularly plastic, poses a significant environmental challenge.
- Effective management requires accurate data on litter abundance and composition.
- Current methods for litter assessment can be labor-intensive and limited in scope.
Purpose of the Study:
- To propose and validate a multidisciplinary approach for quantifying and classifying urban litter.
- To leverage citizen science and deep learning for efficient street litter analysis.
- To provide a scalable method for monitoring urban pollution.
Main Methods:
- Utilized the Pirika smartphone application for citizen-sourced litter image collection.
- Employed deep learning algorithms for automated image analysis and litter categorization.
- Integrated geospatial data from the application for litter mapping.
Main Results:
- Collected approximately one million litter images, identifying cans, plastic bags, and bottles as the most prevalent items.
- Developed a deep learning model with >75% precision and recall for identifying top litter categories.
- Visualized litter distribution across urban areas using mapped data.
Conclusions:
- Citizen science, combined with smartphone technology and deep learning, offers a powerful tool for urban litter assessment.
- This integrated approach enables effective visualization, quantification, and characterization of street litter.
- The methodology supports enhanced pollution management and sustainable urban development.
