Remote sensing smart city solution for municipal waste management?

By Luong Nguyen and Tram-Anh Pham from Accelerator Lab Viet Nam

March 9, 2022



Environmental degradation and waste pollution have increasingly been among the most pressing problems in Viet Nam, mainly due to the rapid growth of industrial and economic activities, population growth, and a weak waste management system (SDGs 11, 12, 13).

According to a 2019 report by Minister of Natural Resources and Environment  MONRE ,the daily municipal waste increased from 32,000 tons in 2014 to 35,624 tons in 2019, accounting for more than 50% of the national solid waste. Meanwhile, as the waste collection system and urban infrastructure has not been able to catch up with the rising waste; open-pit landfills and temporary dumps lead to air pollution in the surrounding area -- negatively affecting health and upsetting the lives of the residents. During the pandemic, the amount of plastic waste skyrocketed, putting further pressure on Viet Nam’s poor waste management systems.


Beautiful beaches in Danang are being covered with waste (Photo: Zing)


In Viet Nam, municipal waste management is handled by a combination of actors, including the Peoples’ Committee (PC), the Department of Natural Resource and Environment (DoNRE), state-owned environmental companies such as Urban Environment Company (URENCO) and CITENCO, private enterprises, and the network of Đồng Nát/Ve chai (informal waste workers, also known as the informal waste sector). Waste management in large metropolitan cities such as Ha Noi and Da Nang is a wicked problem with systemic social, technological, and economic causes. One of the key issues from an urban and policy planning standpoint is the lack of predictive data on waste hotspots to aid the collection, evaluation, and monitoring of pollutants.

To bridge the gap in policymaking and to aid the municipal waste management effort, and as part of the participation in the Japan SDGs Innovation Challenge, the team at Accelerator Lab UNDP in Viet Nam, in partnership the Japan Manned Spaced Systems Corporation (JAMSS) and the Da Nang Institute for Socio-Economic Development (DISED), has been conducting a 2-phase feasibility study since February 2021 on the application of remote sensing (satellite imagery) technology to municipal waste management to inform Da Nang City’s Circular Economy Roadmap:

Phase 1: Remote sensing of solid waste (plastic) and marine plastic litter in Da Nang City to serve the municipal plan to use waste in developing a circular economy system.

Phase 2: Explore specific questions related to wastewater discharging from factories and residential areas to sea areas.

For Phase 1, our local partner, the Da Nang Institute for Socio-Economic Development (DISED) suggested four plastic waste hotspots for investigation (Tho Quang wharf; Khanh Son landfill; Son Tra peninsula beaches; sewer outfall locations without adequate treatment). Plastic waste hotspots are places with a persistent plastic concentration large enough for satellites such as the Sentinel-2 to detect (>10m2). Using satellite imaging data and spectrum analysis, JAMSS reported seven potential hotspots across the city, to be verified using remote sensing data (Âu Huyền Thọ Quang; the Green Island; Dong No Resettlement Area; FPT New Urban Area; Tu Cau Bridge; Khue Dong Bridge; the intersection area at Nguyen Hong Anh; the area at Cau Do River; the area of Quang Xuong), while false positives were detected in seven locations by on the ground verification with drone imagery.


Using a combination of satellite imagery data and predictive algorithms, the research team investigated where potential plastic waste hotspots are located



Phase 1 is completed, and results show the advantages of using satellites to detect plastic waste on a large scale, pointing out potential coastal and riverside waste hotspots for further verification by the local authority. We investigated a method to separate the plastic and background spectra. The attached figure shows the validation results, and we have detected plastics ranging from large-scale plastic facilities to those plastic piles accumulated outdoors in a craft village in Ha Noi. The effectiveness of the approach has been reported to have an accuracy of over 80% in the craft village – though more on ground research is needed to confirm this method reliability.



A sample of the accuracy verification and sensitivity assessment in a craft village in Ha Noi. The detection accuracy of the plastic accumulation was maintained at roughly 88% for two months. The sensitivity is stable at around 80-90% (Photo: JAMSS)


The real-world application of this technology needs further development and research. The limitation of current remote sensing and identification technology is related to resolution and spectrum analysis.

On-ground verification is still needed to confirm hotspot detection and avoid false positives. For example, in certain instances we’ve detected rock surfaces and FRP-equivalent plastics (a ship) as plastic waste due to a similarity in wavelengths. JAMSS found that the spectral profiles of the rock and the ship matched each other almost within the margin of error. The finding gave a better understanding of the whole picture:

- The algorithm derives the target spectrum almost correctly.

- Under certain conditions, it is difficult to distinguish between objects with spectral similarity. The identification of only 12 wavelengths, which is a feature of the spectrum from the Sentinel satellite, is insufficient. 


A comparison of spectra between rock and ship (FRP-equivalent). The correlation coefficient between the two spectra was calculated to be 0.993, and the two profiles, in the error range, were almost identical.


Another key finding of Phase 1 is the possibility of detecting other substances that have similar spectral profiles to plastics. Hence, more research is needed to increase accuracy and reliability. Possible methods include using ground truth data verification from local hotspots or making better use of color information from satellite data (using hyperspectral satellites). Under the second method, clustering with more color information instead of 12 points could potentially increase the accuracy of JAMSS. The aim at this phase is to establish a way to eliminate false positive data in the future.

What’s next?

Our team is currently at the end of Phase 2, primarily focusing on water pollution, which can be more easily identified in a large area with remote sensing technology and spectral profile. JAMSS is to use remote sensing technology (and GIS) to build a database to map wastewater pollution areas caused by sludge discharged into the sea along the Han River (East Coast Region/Da Nang Bay), after which DISED can test samples in the same location to initially provide and confirm any pollutants. Testing is also applied to study how hyperspectral remote sensing can be used to measure turbidity and chloroph in rivers and reservoirs. 


Phase 2: Initial findings regarding turbidity, chlorophyll, normalized difference vegetation index (NDVI), and flood analysis by SAR


The intention is that the collaboration between DISED and JAMSS can contribute to the design of a smart city remote sensing monitoring plan suitable for the capacity of Danang City, which could then be developed for application in similar urban settings. The results of this collaboration will be the basis for recommendations for the Danang Circular City Roadmap currently being developed by DISED. 


Danang city is pioneering the development of a city-wide Circular Economy Roadmap, one of the first in Viet Nam


Satellite imagery can be a powerful tool for identifying and tracking socio-environmental changes on a large scale and aiding municipal monitoring and management of waste. With substantial progress in Phase 2, we look forward to exploring further unexpected findings and continuing to refine this approach. Drop us a message at if you would like to talk more about our work with JAMSS and DISED or if you have suggestions.

Acknowledgement: This work has been made possible through Japanese Innovation Challenge with funding from Japan Government and contributions from the following experts: Mr. Yuichi Ito (JAMSS), Ms. Nguyet Bui (DISED), Ms. Xuan Quach, and Mr. Huy Nguyen, and with support from the Japan Innovation Network.