Plastic pollution is a critical global issue. Increases in plastic consumption have triggered increased production, which in turn has led to increased plastic disposal. In situ observation of plastic litter is tedious and cumbersome, especially in rural areas and around transboundary rivers. We therefore propose automatic mapping of plastic in rivers using unmanned aerial vehicles (UAVs) and deep learning (DL) models that require modest compute resources. We evaluate the method at two different sites: the Houay Mak Hiao River, a tributary of the Mekong River in Vientiane, Laos, and Khlong Nueng canal in Talad Thai, Khlong Luang, Pathum Thani, Thailand. Detection models in the You Only Look Once (YOLO) family are evaluated in terms of runtime resources and mean average Precision (mAP) at an Intersection over Union (IoU) threshold of 0.5. YOLOv5s is found to be the most effective model, with low computational cost and a very high mAP of 0.81 without transfer learning for the Houay Mak Hiao dataset. The performance of all models is improved by transfer learning from Talad Thai to Houay Mak Hiao. Pre-trained YOLOv4 with transfer learning obtains the overall highest accuracy, with a 3.0% increase in mAP to 0.83, compared to the marginal increase of 2% in mAP for pre-trained YOLOv5s. YOLOv3, when trained from scratch, shows the greatest benefit from transfer learning, with an increase in mAP from 0.59 to 0.81 after transfer learning from Talad Thai to Houay Mak Hiao. The pre-trained YOLOv5s model using the Houay Mak Hiao dataset is found to provide the best tradeoff between accuracy and computational complexity, requiring model resources yet providing reliable plastic detection with or without transfer learning. Various stakeholders in the effort to monitor and reduce plastic waste in our waterways can utilize the resulting deep learning approach irrespective of location.

/pdf/detection-of-river-plastic-using-uav-sensor-data-and-deep-c0i24wco.pdf

Detection of River Plastic Using UAV Sensor Data and Deep Learning

Plastic pollution in coastal and marine areas is an ongoing environmental concern in the world. Despite its growing concern worldwide, there is a knowledge gap in terms of its detection and quantification of spatial distribution patterns. In this review paper, we investigated the existing trends in monitoring marine and beached (macro and meso) plastic debris using remote sensing techniques. Marine plastic debris monitoring using remote sensing data is highly complex, partially due to the constraints in spatial and spectral resolutions as well as due to the differences in surface reflectance properties of different polymers. Modern remote sensing with high-quality data, including airborne, spaceborne, and data taken from an unmanned aerial vehicle (UAV), have revolutionized marine litter mapping. UAV platform offers a promising and cost-effective way for mapping both floating and beached marine plastic debris compared to data from other platforms. Automatic and manual delineation for mapping marine plastic litters can be applied depending on the objective of the study. For assisting beach clean-up processes, automatic detection of marine plastic debris is enough for locating the worksite. For detailed characterization of debris, including type, size and material, manual delineation of high resolution images is necessary. More accurate mapping of marine plastics can be achieved by developing detailed spectral libraries with material level details and launching dedicated sensors for marine litter mapping in the future. Challenges exist in mapping submerged marine plastics, plastic materials conglomerated with biomass and unaccounted for beached marine litter buried along the coasts due to accretion and erosion. 

Coastal and marine plastic litter monitoring using remote sensing: A review

The detection of waste plastics in the marine and terrestrial environment using satellite Earth Observation data offers the possibility of large-scale mapping and reducing on-the-ground manual investigation. In addition, costs are kept to a minimum by utilizing free-to-access Copernicus data. A Machine Learning-based classifier was developed to run on Sentinel-1 and -2 data. In support of the training and validation, a dataset was created with terrestrial and aquatic cases by manually digitizing varying landcover classes alongside plastics under the sub-categories of greenhouses, plastic, tyres and waste sites. The trained classifier, including an Artificial Neural Network and post-processing decision tree, was verified using five locations encompassing these different forms of plastic. Although exact matchups are challenging to digitize, the performance has generated high accuracy statistics, and the resulting land cover classifications have been used to map the occurrence of plastic waste in aquatic and terrestrial environments.

/pdf/detection-of-waste-plastics-in-the-environment-application-2gfc5v3c.pdf

Detection of Waste Plastics in the Environment: Application of Copernicus Earth Observation Data

Plastic debris has a long-term and quite severe impact on the marine ecosystem. Population growth in the developing area, Makassar coastal zone, will contribute significantly to the land’s waste. Consequently, floating plastic debris comes directly from the land’s plastic leaks to rapidly increase marine plastic debris. Due to being a major environmental issue in the coastal zone and reducing options for removal are very limited, this study becomes to detect the potential marine plastic debris in Makassar by Sentinel-2 Multi-Spectral Instrument (MSI) using Kernel Normalization Vegetation Index (KNDVI) and Floating Debris Index (FDI) for satellite image processing. The plastic debris was corrected atmospheric effect by Sen2Cor and processed by Sentinel Application Platform (SNAP). The resolution of images was set to 10 x 10 meters. The target area was calculated as 173.47 Ha. The results show that KNDVI and FDI combined are a developed and powerful approach to detecting plastic debris. KNDVI has more representative indices (SE 0.38) than FDI indices (SE 0.47). Through the discriminant analysis were verified pixels of KNDVI (3506 pixels) and FDI indices (701 pixels). Meanwhile, the Sensitivity Analysis Value (SAV) was well performed to detect the potential of marine plastic debris by KNDVI with SAV 7.5 than FDI indices with SAV 2.9.

https://iopscience.iop.org/article/10.1088/1755-1315/1117/1/012054/pdf

Potential Marine Plastic Debris Detection using Sentinel-2 Multi-Spectral Instrument (MSI)

Marine plastic debris (MPD) is a globally relevant environmental challenge, with an estimated 8 million tons of synthetic debris entering the marine environment each year. Plastic has been found in all parts of the marine environment, including the surface layers of the ocean, within the water column, in coastal waters, on the benthic layer and on beaches. While research on detecting MPD using remote sensing is increasing, most of it focuses on detecting floating debris in open waters, rather than detecting MPD on beaches. However, beaches present challenges that are unique from other parts of the marine environment. In order to better understand the spectral properties of beached MPD, we present the SWIR reflectance of weathered MPD and virgin plastics over a sandy substrate. We conducted spectral feature analysis on the different plastic groups to better understand the impact that polymers have on our ability to detect synthetic debris at sub-pixel surface covers that occur on beaches. Our results show that the minimum surface cover required to detect MPD on a sandy surface varies between 2–8% for different polymer types. Furthermore, plastic composition affects the magnitude of spectral absorption. This suggests that variation in both surface cover and polymer type will inform the efficacy of beach litter detection methods.

Samantha M. Blades

Papers

Quantifying Marine Plastic Debris in a Beach Environment Using Spectral Analysis