Day 4

Detailed paper information

Back to list

Paper title Decision support tools for landfill management based on UAVs data
  1. Benjamin Beaumont ISSeP Speaker
  2. Coraline Wyard Institut Scientifique de Service Public (ISSeP)
  3. Taïs Grippa ULB
  4. Julien Dumont Institut Scientifique de Service Public - ISSeP
  5. Fabian Stassen Institut Scientifique de Service Public - ISSeP
  6. Emilie Navette Institut Scientifique de Service Public - ISSeP
  7. Éric Hallot Institut Scientifique de Service Public - ISSeP
Form of presentation Poster
  • C4. HAPs/UAVs
    • C4.01 Innovative UAV applications
Abstract text The ability of unmanned aerial vehicles (UAVs) to acquire data in a non-intrusive way is a definite advantage for the development of decision support tools for the monitoring of complex sites. For this purpose, active landfills, due to the continuous ground operations, such as earthmoving and depositing of miscellaneous waste with heavy machinery, as well as the pronounced topography, are sensitive sites where UAVs are relevant compared to traditional ground survey methods. Legislation requires the quarterly monitoring of the site’s topography, ground instability risk and landfill completion rate with respect to the environmental permits. Mapping of landfill infrastructure and monitoring of biogas and leachate leaks is also crucial for controlling authorities. This research led to the development of three cost effective solutions to support day-to-day activities and controlling campaigns over landfills by site managers and competent authorities: the monitoring of the land cover (LC) of the site, the monitoring of its topography, the detection of biogas emissive areas and leachate leaks.

First, visible orthomosaic with centimetric spatial resolution provides an unparalleled image for visual site and infrastructure inspection as well as for LC classification. A state-of-the-art object-oriented image analysis (OBIA) approach initially designed for the processing of very high-resolution satellite data was successfully applied to UAVs data to maps LC of a 30 ha landfill site. The optimization of this processing chain through texture computation and feature selection has made it possible to achieve an overall accuracy higher than 80% for a nine-LC-categories classification. These classes include various types of waste, bare soils, tarps, green and dry vegetation, road and built-up infrastructures. Active landfill LC is usually very fragmented and evolves significantly from day to day. Therefore, such an automated method is useful for spatial and temporal monitoring of dynamic LC changes.

Second, digital surface model (DSM) is a classic by-product of photogrammetric processing. In addition to its use in draping the orthophotos mosaic for a tridimensional visualization of the site, DSM allows for a precise monitoring of topography, slopes, volumetric change and volumetric estimation of deposits. In this study, the comparison between UAVs DSM and ground topographic surveys shows that UAVs DSM models completely and finely all topographical features in a short space of time (less than a day) while a ground-based topographic survey could take several days. This completeness of the measure and its non-intrusive character are a clear advantage according to the site managers. Still, well-known limitations are that it does not allow reaching the same standards of quality at the level of the ground survey points taken by GPS/GNSS solutions (precision in Z around 10 cm for UAVs DSM versus 2 cm) and is impacted by the presence of vegetation.

Thirdly, thermal data are used as a proxy for the detection of emissive areas and leachate leaks. Indeed, the degradation processes of the waste lead to a heating of the buried bodies up to 60°C. Although this heating is reduced at the surface, this research confirms again the hypothesis that the temperature differential allows the detection of weakness areas and warmed liquids. Flying in optimal conditions (at dawn, cold, dry and not windy weather), our thermal mosaic dataset allowed the detection of three leachate leaks and one biogas emitting area. Such tool speeds up the control procedure in the field and allows the rapid implementation of corrective measures to avoid greenhouse gas emissions, optimize biogas collection for energy production, and reduce odors and risks of explosion or internal fire.

The three decision support tools developed will now be operationally integrated into the administration's landfill control activities. Data acquisition and processing can theoretically be done in less than a day, but is still highly dependent on flight clearances and weather conditions. Several derived applications are envisaged, in particular for the follow-up of other sites at risk or the qualification and quantification of illegal activities such as clandestine deposits. The whole should contribute to a more efficient and less costly monitoring of our environment.