pith. sign in

arxiv: 2605.02316 · v1 · submitted 2026-05-04 · 💻 cs.CV · cs.LG

Open-access model for detecting openly dumped dispersed municipal solid waste from crowdsourced UAV imagery in Sub-Saharan Africa

Steffen Knoblauch , Ram Kumar Muthusamy , Luis M. A. Bettencourt , Costas Velis , Pierre Chrzanowski , Edward Charles Anderson , Pete Masters , Innocent Maholi
show 3 more authors
Antonio Inguane Levi Szamek Alexander Zipf
This is my paper

Pith reviewed 2026-05-08 19:35 UTC · model grok-4.3

classification 💻 cs.CV cs.LG
keywords wastedispersedsolidacrossdumpedimagerymodelopenly
0
0 comments X p. Extension

The pith

An open-access deep learning model detects openly dumped dispersed municipal solid waste in crowdsourced UAV imagery from 29 regions in 10 Sub-Saharan countries and shows strong performance with patterns tied to local population and infrastructure.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Sub-Saharan African cities are growing quickly but often lack enough waste collection services, so people dump trash in scattered spots that are hard to track. Drones flown by local communities can photograph these areas, but looking through thousands of pictures by hand takes too much time. The authors collected drone images from many different places, had people mark where the waste was in small sections of the photos, and used those marks to train a deep learning model. Deep learning is a computer method that learns to recognize patterns like piles of trash from examples. When the model was tested on new photos from the same regions, it found the waste accurately in varied settings such as near rivers and in crowded neighborhoods. The detected waste often clustered along waterways, which can worsen flooding and spread disease, and it appeared more in areas with high population but poor local services than in places with better overall development. The team made the finished model freely available so city workers or volunteer mappers can run it on their own drone photos without needing to build or train anything themselves. This turns raw drone pictures into clear maps of problem spots that can help decide where to send cleanup crews.

Core claim

A deep learning model trained on manually annotated image tiles achieved excellent performance in detecting openly dumped dispersed solid waste across all study regions.

Load-bearing premise

The manually annotated image tiles from 29 regions accurately capture the visual appearance of dispersed waste under the diverse environmental conditions present across the 10 countries.

Figures

Figures reproduced from arXiv: 2605.02316 by Alexander Zipf, Antonio Inguane, Costas Velis, Edward Charles Anderson, Innocent Maholi, Levi Szamek, Luis M. A. Bettencourt, Pete Masters, Pierre Chrzanowski, Ram Kumar Muthusamy, Steffen Knoblauch.

Figure 1
Figure 1. Figure 1: Overview of the end-to-end workflow for large-scale UAV-based detection of openly dumped dispersed MSW and socio-spatial assessment across Sub-Saharan Africa. The pipeline integrates multi￾source geospatial data acquisition (A-B), pre-processing (C-D), supervised training (E), regional prediction and validation (F-G), and bivariate correlation analysis (H). UAV imagery was curated from OpenAerialMap coveri… view at source ↗
Figure 2
Figure 2. Figure 2: Stability of evaluation metrics under stratified bootstrap subsampling of the test set. Perfor￾mance metrics were evaluated across increasing test-set sizes using stratified bootstrap subsampling that preserved the class balance between openly dumped dispersed MSW and background tiles. Curves show the mean value of each metric over bootstrap replicates, with shaded regions indicating variability across res… view at source ↗
Figure 3
Figure 3. Figure 3: Visual inspection of model predictions. Representative examples of true positives, true negatives, false positives, and false negatives are shown. We applied the trained model across all study regions to generate spatially explicit predictions of openly dumped dispersed MSW distribution. Predictions were generated for more than 13 mil￾lion 5 m × 5 m imagery tiles, enabling high-resolution characterization … view at source ↗
Figure 4
Figure 4. Figure 4: Predicted openly dumped dispersed MSW distribution across study sites. Each panel represents one of 29 regions. Red 5 m × 5 m grid cells indicate model-predicted openly dumped dispersed MSW locations, highlighting fine-scale spatial patterns including hotspots of dumping. Accompanying bar charts show calculated ODDMSWC values, which were used to rank the study sites. These maps provide i) a comprehensive o… view at source ↗
Figure 5
Figure 5. Figure 5: Openly dumped dispersed MSW contamination in Dar es Salaam, Tanzania detected from UAV imagery. Red tiles indicate detections of openly dumped dispersed MSW at 5 m × 5 m resolution. Panels A and B show zoomed-in views; panel C highlights concentrations along pathways and waterways. Reflecting these observations, ODDMSWC values reveal substantial variation across study sites. The highest ODDMSWC was observe… view at source ↗
Figure 6
Figure 6. Figure 6: Bivariate associations between relative solid waste contamination and socio-spatial indicators. Scatterplots show study-region–level relationships between the ODDMSWC and three contextual indicators: SHDI (A), a infrastructure deficit index capturing block street access (B), and gridded population density (C). Spearman’s rank correlations indicate no significant association with SHDI (ρ = 0.215), but stron… view at source ↗
Figure 7
Figure 7. Figure 7: Openly dumped dispersed MSW in Dar es Salaam, Tanzania, detected from openly available satellite imagery. Red tiles indicate detections of openly dumped dispersed MSW. Panels A and B show zoomed-in views, while panel C provides a large-scale overview. All panels illustrate the limited spatial resolution and detection performance of openly available satellite imagery compared with UAV imagery ( view at source ↗
read the original abstract

Managing municipal solid waste in rapidly urbanizing Sub-Saharan Africa remains challenging due to dispersed informal dumping and limited high-resolution datasets for spatial monitoring. We present an open-access deep learning model for automated detection of openly dumped dispersed solid waste via crowdsourced UAV imagery, trained and evaluated across 29 regions in 10 countries, encompassing diverse environmental contexts. A deep learning model trained on manually annotated image tiles achieved excellent performance in detecting openly dumped dispersed solid waste across all study regions. Predicted distributions reveal heterogeneous accumulation patterns, ranging from localized hotspots - often along waterways, where waste can exacerbate flood and public health risks - to more dispersed litter across urban areas. Waste accumulation is most strongly associated with population density and indicators of lack of local infrastructure access, whereas its relationship with broader measures of regional development is weaker, highlighting the importance of fine-scale data for understanding localized waste dynamics. By releasing the model, this study provides a ready-to-use tool for UAV imagery collected by municipalities and local mapping communities, enabling openly dumped dispersed solid waste monitoring without extensive technical expertise. This approach empowers local practitioners to convert UAV imagery into actionable insights, supporting targeted interventions and improved municipal solid waste management across Sub-Saharan Africa.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the representativeness of the crowdsourced annotations and the assumption that standard deep-learning generalization holds across the sampled environmental diversity; no new physical entities or ad-hoc constants are introduced.

axioms (1)
  • domain assumption Manually annotated UAV image tiles provide ground-truth labels that generalize to unseen tiles from the same regions
    Invoked when claiming excellent performance on the full set of study regions after training on annotated tiles.

pith-pipeline@v0.9.0 · 5557 in / 1388 out tokens · 54160 ms · 2026-05-08T19:35:12.246805+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

52 extracted references · 5 canonical work pages · 1 internal anchor

  1. [1]

    URL:https://digitallibrary.un.org/record/3923923?ln=en&v=pdf

    United Nations, Transforming our world: The 2030 agenda for sustainable development, 2015. URL:https://digitallibrary.un.org/record/3923923?ln=en&v=pdf

    work page arXiv 2030

  2. [2]

    S. Kaza, L. Yao, P. Bhada-Tata, F. van Woerden, What a waste 2.0: a global snap- shot of solid waste management to 2050, World Bank Publications, 2018. doi:10.1596/ 978-1-4648-1329-0

    2050

  3. [3]

    U. N. Environment, Global waste management outlook 2024 unep - un environment programme, 2024. URL:https://www.unep.org/resources/ global-waste-management-outlook-2024

    2024

  4. [4]

    E. Cook, K. Ionkova, P. Bhada-Tata, S. Yadav, F. van Woerden, What a Waste 3.0: Global Snapshot of Solid Waste Management Toward Circularity until 2050, Washington, DC: World Bank, 2026. doi:10.1596/978-1-4648-2309-1

    work page doi:10.1596/978-1-4648-2309-1 2050

  5. [5]

    Francesco Di Maria, Caterina Micale, Alessio Sordi, Giuseppe Cirulli, Moreno Marionni, Ur- ban mining: Quality and quantity of recyclable and recoverable material mechanically and physically extractable from residual waste, Waste Management 33 (2013) 2594–2599

    2013

  6. [6]

    Lambert, M

    S. Lambert, M. Wagner, Microplastics are contaminants of emerging concern in freshwater environments: An overview (2017) 1–23

    2017

  7. [7]

    Charles Kihampa, William J. S. Mwegoha, Heavy metals accumulation in vegetables grown along the msimbazi river in dar es salaam, tanzania, International Journal of Biological and Chemical Sciences 4 (2011)

    2011

  8. [8]

    Rouhani, M

    A. Rouhani, M. Hejcman, A review of soil pollution around municipal solid waste landfills in iran and comparable instances from other parts of the world, International Journal of Environmental Science and Technology 22 (2025) 9711–9728

    2025

  9. [9]

    Moritz, S

    M. Moritz, S. Knoblauch, Geospatial innovation’s potential for addressing mosquito- borne diseases, 2023. URL:https://unstats.un.org/unsd/undataforum/blog/ geospatial-addressing-mosquito-borne-diseases/. 23

    2023

  10. [10]

    Carbery, W

    M. Carbery, W. O’Connor, T. Palanisami, Trophic transfer of microplastics and mixed contam- inants in the marine food web and implications for human health, Environment international 115 (2018) 400–409

    2018

  11. [11]

    O. A. Mokuolu, A. K. Odunaike, J. O. Iji, A. S. Aremu, Assessing the effects of solid wastes on urban flooding: A case study of isale koko, Lautech Journal of Civil and Environmental Studies 9 (2022) 22–30

    2022

  12. [12]

    Knoblauch, M

    S. Knoblauch, M. Su Yin, K. Chatrinan, A. A. de Aragão Rocha, P. Haddawy, F. Biljecki, S. Lautenbach, B. Resch, D. Arifi, T. Jänisch, et al., High-resolution mapping of urban aedes aegypti immature abundance through breeding site detection based on satellite and street view imagery, Scientific Reports 14 (2024) 18227

    2024

  13. [13]

    Gkoutselis, S

    G. Gkoutselis, S. Rohrbach, J. Harjes, et al., Microplastics accumulate fungal pathogens in terrestrial ecosystems, Scientific Reports 11 (2021) 13214

    2021

  14. [14]

    Leslie, Martin J.M

    Heather A. Leslie, Martin J.M. van Velzen, Sicco H. Brandsma, A. Dick Vethaak, Juan J. Garcia-Vallejo, Marja H. Lamoree, Discovery and quantification of plastic particle pollution in human blood, Environment international 163 (2022) 107199

    2022

  15. [15]

    A. I. Osman, M. Hosny, A. S. Eltaweil, S. Omar, A. M. Elgarahy, M. Farghali, P.-S. Yap, Y.-S. Wu, S. Nagandran, K. Batumalaie, et al., Microplastic sources, formation, toxicity and remediation: a review, Environmental Chemistry Letters 21 (2023) 2129–2169

    2023

  16. [16]

    E. Cook, C. A. Velis, L. Black, Construction and demolition waste management: A systematic scoping review of risks to occupational and public health, Frontiers in Sustainability 3 (2022)

    2022

  17. [17]

    URL:https://aqmx.org/resources/ waste-wise-cities-tool

    UN Habitat, Waste wise cities tool, 2021. URL:https://aqmx.org/resources/ waste-wise-cities-tool

    2021

  18. [18]

    URL:https://www.c40.org/ accelerators/pathway-towards-zero-waste/

    C40 Cities, C40 sustainable waste systems accelerator, 2025. URL:https://www.c40.org/ accelerators/pathway-towards-zero-waste/

    2025

  19. [19]

    URL:https://www.c40.org/wp-content/uploads/2025/ 12/C40-Sustainable-Waste-Systems-Accelerator-Report.pdf

    C40 Cities, How cities are creating cleaner, equitable and climate-resilient cities through sus- tainable waste management, 2025. URL:https://www.c40.org/wp-content/uploads/2025/ 12/C40-Sustainable-Waste-Systems-Accelerator-Report.pdf. 24

    2025

  20. [20]

    URL:https://dortibox.com

    DortiBox, Dortibox website, 2026. URL:https://dortibox.com

    2026

  21. [21]

    URL:https://yowasteapp.com/

    Yo-Waste, Yo-waste website, 2026. URL:https://yowasteapp.com/

    2026

  22. [22]

    URL:https://baustaka.co.ke/

    Baus Taka, Baus taka website, 2026. URL:https://baustaka.co.ke/

    2026

  23. [23]

    URL:https://www.coliba.com.gh/

    Coliba, Coliba website, 2026. URL:https://www.coliba.com.gh/

    2026

  24. [24]

    URL:https://www.wrap.ngo/

    WRAP, Wrap project website, 2026. URL:https://www.wrap.ngo/

    2026

  25. [25]

    David C Wilson, Costas A Velis, Waste management – still a global challenge in the 21st century: An evidence-based call for action, Waste Management & Research 33 (2015) 1049– 1051

    2015

  26. [26]

    Lebreton, A

    L. Lebreton, A. Andrady, Future scenarios of global plastic waste generation and disposal, Palgrave Communications 5 (2019) 1–11

    2019

  27. [27]

    Gwada, G

    B. Gwada, G. Ogendi, S. M. Makindi, S. Trott, Composition of plastic waste discarded by households and its management approaches, Global Journal of Environmental Science and Management 5 (2019) 83–94

    2019

  28. [28]

    Combes, C

    P.-P. Combes, C. Gorin, S. Nakamura, M. Roberts, An anatomy of urbanization in sub-saharan africa, Regional Science and Urban Economics 115 (2025) 104152

    2025

  29. [29]

    Fraternali, L

    P. Fraternali, L. Morandini, S. L. Herrera González, Solid waste detection, monitoring and mapping in remote sensing images: A survey, Waste Management 189 (2024) 88–102

    2024

  30. [30]

    P. K. Kalonde, T. Mwapasa, R. Mthawanji, K. Chidziwisano, T. Morse, J. S. Torguson, C. M. Jones, R. S. Quilliam, N. A. Feasey, M. Y. R. Henrion, M. C. Stanton, M. S. Blinnikov, Mapping waste piles in an urban environment using ground surveys, manual digitization of drone imagery, and object based image classification approach, Environmental Monitoring and...

    2025

  31. [31]

    Seán Lynch, Openlittermap.com – open data on plastic pollution with blockchain rewards (littercoin), Open Geospatial Data, Software and Standards 3 (2018) 6

    2018

  32. [32]

    Tisserant, S

    A. Tisserant, S. Pauliuk, S. Merciai, J. Schmidt, J. Fry, R. Wood, A. Tukker, Solid waste and the circular economy: A global analysis of waste treatment and waste footprints, Journal of Industrial Ecology 21 (2017) 628–640. 25

    2017

  33. [33]

    J. R. Jambeck, K. Johnsen, Citizen-based litter and marine debris data collection and mapping, Computing in Science & Engineering 17 (2015) 20–26

    2015

  34. [34]

    Marthe Larsen Haarr, Michael Pantalos, Monica Kleffelgård Hartviksen, Marit Gressetvold, Citizen science data indicate a reduction in beach litter in the lofoten archipelago in the norwegian sea, Marine Pollution Bulletin 153 (2020) 111000

    2020

  35. [35]

    Irwin, Citizen science: A study of people, expertise and sustainable development, Environ- ment and Society, Routledge, 1995

    A. Irwin, Citizen science: A study of people, expertise and sustainable development, Environ- ment and Society, Routledge, 1995

    1995

  36. [36]

    H. Abdu, M. H. Mohd N., A survey on waste detection and classification using deep learning, IEEE Access 10 (2022) 128151–128165

    2022

  37. [37]

    X. Sun, D. Yin, F. Qin, H. Yu, W. Lu, F. Yao, Q. He, X. Huang, Z. Yan, P. Wang, C. Deng, N. Liu, Y. Yang, W. Liang, R. Wang, C. Wang, N. Yokoya, R. Hänsch, K. Fu, Revealing influencing factors on global waste distribution via deep-learning based dumpsite detection from satellite imagery, Nature communications 14 (2023) 1444

    2023

  38. [38]

    M. R. Devesa, A. V. Brust, Mapping illegal waste dumping sites with neural-network classifi- cation of satellite imagery (2021)

    2021

  39. [39]

    D. Zeng, S. Zhang, F. Chen, Y. Wang, Multi-scale cnn based garbage detection of airborne hyperspectral data, IEEE Access 7 (2019) 104514–104527

    2019

  40. [40]

    B. Wang, Y. Xing, N. Wang, C. L. P. Chen, Monitoring waste from uncrewed aerial vehicles and satellite imagery using deep learning techniques: A review, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 17 (2024) 20064–20079

    2024

  41. [41]

    L. G. Papale, G. Guerrisi, D. de Santis, G. Schiavon, F. Del Frate, Satellite data potentialities in solid waste landfill monitoring: Review and case studies, Sensors 23 (2023)

    2023

  42. [42]

    S. Pan, K. Yoshida, A. S. Boney, S. Nishiyama, The application of drone-assisted deep learning technology in riverbank garbage detection 78 (2022)

    2022

  43. [43]

    Wyard, B

    C. Wyard, B. Beaumont, T. Grippa, S. Georganos, E. Hallot, Uavs for fine-scale open-source landfill mapping, in: 2021 IEEE International Geoscience and Remote Sensing Symposium IGARSS, 2021, pp. 8217–8220. doi:10.1109/IGARSS47720.2021.9553815. 26

    work page doi:10.1109/igarss47720.2021.9553815 2021

  44. [44]

    Sliusar, T

    N. Sliusar, T. Filkin, M. Huber-Humer, M. Ritzkowski, Drone technology in municipal solid waste management and landfilling: A comprehensive review, Waste Management 139 (2022) 1–16

    2022

  45. [45]

    Knoblauch, L

    S. Knoblauch, L. Szamek, J. Wenk, I. Chazua, I. Maholi, M. Adamiak, S. Lautenbach, A. Zipf, Uav-assisted municipal solid waste monitoring for informed disposal decisions, in: Proceedings of the 2024 International Conference on Information Technology for Social Good, ACM, New York, NY, USA, 2024, pp. 105–113. doi:10.1145/3677525.3678649

    work page doi:10.1145/3677525.3678649 2024

  46. [46]

    AI-based Waste Mapping for Addressing Climate-Exacerbated Flood Risk

    S. Knoblauch, L. Szamek, I. Chazua, B. Adamu, I. Maholi, A. Zipf, Ai-based waste mapping for addressing climate-exacerbated flood risk, in: NeurIPS 2025 Workshop on Tackling Cli- mate Change with Machine Learning, 2025. URL:https://www.climatechange.ai/papers/ neurips2025/6. doi:10.48550/ARXIV.2604.18151

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2604.18151 2025

  47. [47]

    L. M. A. Bettencourt, N. Marchio, Infrastructure deficits and informal settlements in sub- saharan africa, Nature 645 (2025) 399–406

    2025

  48. [48]

    Majchrowska, A

    S. Majchrowska, A. Mikołajczyk, M. Ferlin, Z. Klawikowska, M. A. Plantykow, A. Kwasigroch, K. Majek, Deep learning-based waste detection in natural and urban environments, Waste Management 138 (2022) 274–284

    2022

  49. [49]

    S. Jin, Z. Yang, G. Królczykg, X. Liu, P. Gardoni, Z. Li, Garbage detection and classification using a new deep learning-based machine vision system as a tool for sustainable waste recycling, Waste management (New York, N.Y.) 162 (2023) 123–130

    2023

  50. [50]

    W. Chen, H. Wang, H. Li, Q. Li, Y. Yang, K. Yang, Real-time garbage object detection with data augmentation and feature fusion using suav low-altitude remote sensing images, IEEE Geoscience and Remote Sensing Letters 19 (2022) 1–5

    2022

  51. [51]

    URL:https://www.hotosm.org/en/ tools-resources/tech-product-suite/chat-map/

    Humanitarian OpenStreetMap Team, 2026. URL:https://www.hotosm.org/en/ tools-resources/tech-product-suite/chat-map/

    2026

  52. [52]

    Sebastian, Agro-ecological zones of africa (2013)

    K. Sebastian, Agro-ecological zones of africa (2013). 27

    2013