Quantification of litter in cities using a smartphone application and citizen science in conjunction with deep learning-based image processing.

Cities are a major source of litter pollution. Determination of the abundance and composition of plastic litter in cities is imperative for effective pollution management, environmental protection, and sustainable urban development. Therefore, here, a multidisciplinary approach to quantify and classify the abundance of litter in urban environments is proposed. In the present study, litter data collection was integrated via the Pirika smartphone application and conducted image analysis based on deep learning. Pirika was launched in May 2018 and, to date, has collected approximately one million images. Visual classification revealed that the most common types of litter were cans, plastic bags, plastic bottles, cigarette butts, cigarette boxes, and sanitary masks, in that order. The top six categories accounted for approximately 80 % of the total, whereas the top three categories accounted for more than 60 % of the total imaged litter. A deep-learning image processing algorithm was developed to automatically identify the top six litter categories. Both precision and recall derived from the model were higher than 75 %, enabling proper litter categorization. The quantity of litter derived from automated image processing was also plotted on a map using location data acquired concurrently with the images by the smartphone application. Conclusively, this study demonstrates that citizen science supported by smartphone applications and deep learning-based image processing can enable the visualization, quantification, and characterization of street litter in cities.

Marine litter weight estimation from UAV imagery: Three potential methodologies to advance macrolitter reports.

In the context of marine litter monitoring, reporting the weight of beached litter can contribute to a better understanding of pollution sources and support clean-up activities. However, the litter scaling task requires considerable effort and specific equipment. This experimental study proposes and evaluates three methods to estimate beached litter weight from aerial images, employing different levels of litter categorization. The most promising approach (accuracy of 80 %) combined the outcomes of manual image screening with a generalized litter mean weight (14 g) derived from studies in the literature. Although the other two methods returned values of the same magnitude as the ground-truth, they were found less feasible for the aim. This study represents the first attempt to assess marine litter weight using remote sensing technology. Considering the exploratory nature of this study, further research is needed to enhance the reliability and robustness of the methods.

Drones for litter monitoring on coasts and rivers: suitable flight altitude and image resolution.

Multirotor drones can be efficiently used to monitor macro-litter in coastal and riverine environments. Litter on beaches, dunes and riverbanks, along with floating litter on coastal and river waters, can be spotted and mapped from aerial drone images. Items detection and classification are prone to image resolution, which is expressed in terms of Ground Sampling Distance (GSD). The GSD is determined by drone flight altitude and camera properties. This paper investigates what is a suitable GSD value for litter survey. Drone flight altitude and camera setup should be chosen to obtain a GSD between 0.5 cm/px and 1.25 cm/px. Within this range, the lowest GSD allows litter categorization and classification, whereas the highest value should be adopted for a coarser litter census. In the vision of drawing up a global protocol for drone-based litter surveys, this work sets the ground for homogenizing data collection and litter assessments.

BePLi Dataset v1: Beach Plastic Litter Dataset version 1 for instance segmentation of beach plastic litter.

Marine plastic pollution is a pressing global issue nowadays. To address this problem, automated image analysis techniques that can identify plastic litter are necessary for scientific research and coastal management purposes. The Beach Plastic Litter Dataset version 1 (BePLi Dataset v1) comprises 3709 original images taken in various coastal environments, along with instance-based and pixel-level annotations for all plastic litter objects visible in the images. The annotations were compiled in the Microsoft Common Objects in Context (MS COCO) format, which was partially modified from the original format. The dataset enables the development of machine-learning models for instance-level and/or pixel-wise identification of beach plastic litter. All original images in the dataset were extracted from beach litter monitoring records operated by the local government of Yamagata Prefecture in Japan. Litter images were taken in different backgrounds, such as sand beaches, rocky beaches, and tetrapods. The annotations for instance segmentation of beach plastic litter were made manually, and were given for all plastics objects, including PET bottles, containers, fishing gear, and styrene foams,all of which were categorized in a single class "plastic litter". Technologies developed using this dataset have the potential to enable further scalability for the estimation of plastic litter volume. This would help researchers, including individuals, and the the government to monitor or analyze beach litter and the corresponding pollution levels.

The BeachLitter dataset for image segmentation of beach litter.

This dataset consists of 3500 images of beach litter and 3500 corresponding pixel-wise labelled images. Although performing such pixel-by-pixel semantic masking is expensive, it allows us to build machine-learning models that can perform more sophisticated automated visual processing. We believe this dataset may be of significance to the scientific communities concerned with marine pollution and computer vision, as this dataset can be used for benchmarking in the tasks involving the evaluation of marine pollution with various machine learning models. The beach litter images were obtained from coastal environment surveys conducted between 2011 and 2019 by the Yamagata Prefectural Government, Japan. These images were originally obtained owing to the reporting guidelines concerning regular coastal-environmental-cleanup and beach-litter-monitoring surveys. Based on these images, the Japan Agency for Marine-Earth Science and Technology created 3500 images comprising eight classes of semantic masks for beach litter detection [1].

Pixel-level image classification for detecting beach litter using a deep learning approach.

Mitigating and preventing beach litter from entering the ocean is urgently required. Monitoring beach litter solely through human effort is cumbersome, with respect to both time and cost. To address this problem, an artificial intelligence technique that can automatically identify different-sized beach litter is proposed. The technique was established by training a deep learning model that enables pixel-wise classification (semantic segmentation) using beach images taken by an observer on the beach. Eight segmentation classes that include two beach litter classes were defined, and the results were qualitatively and quantitatively verified. Segmentation performance was adequately high based on three metrics: Intersection over Union (IoU), precision, and recall, although there is room for further improvement. The potency of the method was demonstrated when it was applied to images taken in different places from training data images, and the coverage of artificial litter calculated and discussed using drone images provided ground truth.

Estimation of plastic marine debris volumes on beaches using unmanned aerial vehicles and image processing based on deep learning.

Plastic marine debris (PMD) is of global concern. To help address this problem, a novel approach for estimating PMD volumes using a combination of unmanned aerial vehicle (UAV) surveys and image processing based on deep learning is proposed. A three-dimensional model and orthoscopic image of a beach, constructed via Structure from Motion software using UAV-derived data, enabled PMD volumes to be computed by edge detection through image processing. The accuracy of the method was verified by estimating the volumes of test debris placed on a beach in known sizes and shapes. The proposed approach shows potential for estimating PMD volumes with an error of <5%. Compared with subjective methods based on beach surveys, this approach can accurately, rapidly, and objectively calculate the PMD volume on a beach and can be used to improve the efficiency of beach surveys and identify beaches that need preferential cleaning.

Sequential webcam monitoring and modeling of marine debris abundance.

The amount of marine debris washed ashore on a beach in Newport, Oregon, USA was observed automatically and sequentially using a webcam system. To investigate potential causes of the temporal variability of marine debris abundance, its time series was compared with those of satellite-derived wind speeds and sea surface height off the Oregon coast. Shoreward flow induced by downwelling-favorable southerly winds increases marine debris washed ashore on the beach in winter. We also found that local sea-level rise caused by westerly winds, especially at spring tide, moved the high-tide line toward the land, so that marine debris littered on the beach was likely to re-drift into the ocean. Seasonal and sub-monthly fluctuations of debris abundance were well reproduced using a simple numerical model driven by satellite-derived wind data, with significant correlation at 95% confidence level.

Fate of microplastics and mesoplastics carried by surface currents and wind waves: A numerical model approach in the Sea of Japan.

A numerical model was established to reproduce the oceanic transport processes of microplastics and mesoplastics in the Sea of Japan. A particle tracking model, where surface ocean currents were given by a combination of a reanalysis ocean current product and Stokes drift computed separately by a wave model, simulated particle movement. The model results corresponded with the field survey. Modeled results indicated the micro- and mesoplastics are moved northeastward by the Tsushima Current. Subsequently, Stokes drift selectively moves mesoplastics during winter toward the Japanese coast, resulting in increased contributions of mesoplastics south of 39°N. Additionally, Stokes drift also transports micro- and mesoplastics out to the sea area south of the subpolar front where the northeastward Tsushima Current carries them into the open ocean via the Tsugaru and Soya straits. Average transit time of modeled particles in the Sea of Japan is drastically reduced when including Stokes drift in the model.

The potential of oceanic transport and onshore leaching of additive-derived lead by marine macro-plastic debris.

The long-distance transport potential of toxic lead (Pb) by plastic marine debris was examined by pure water leaching experiments using plastic fishery floats containing high level of additive-Pb such as 5100±74.3mgkg(-1). The leaching of Pb ended after sequential 480-h leaching experiments, and the total leaching amount is equivalent to approximately 0.1% of total Pb in a float. But it recovered when the float was scratched using sandpaper. We propose that a "low-Pb layer," in which Pb concentration is negligibly small, be generated on the float surface by the initial leaching process. Thickness of the layer is estimated at 2.5±1.2µm, much shallower than flaws on floats scratched by sandpaper and floats littering beaches. The result suggests that the low-Pb layer is broken by physical abrasion when floats are washed ashore, and that Pb inside the floats can thereafter leach into beaches.

A decadal prediction of the quantity of plastic marine debris littered on beaches of the East Asian marginal seas.

Large quantities of plastic litter are expected to wash ashore along the beaches of the East Asian marginal seas in the coming decade. Litter quantities were predicted using three techniques: a particle tracking model (PTM) used in conjunction with two-way PTM experiments designed to reveal litter sources, an inverse method used to compute litter outflows at each source, and a sequential monitoring system designed to monitor existing beach litter using webcams. Modeled year-to-year variation in litter quantities indicated that the amount of litter would continue to increase in the East Asian marginal seas if the level of outflow remains constant in the coming decade. The study confirms that about 3% of all East Asian beaches may potentially experience a 250-fold increase in the amount of plastic beach litter washed ashore in the next 10 years.

Quantification of toxic metals derived from macroplastic litter on Ookushi Beach, Japan.

The potential risk of toxic metals that could leach into a beach environment from plastic litter washed ashore on Ookushi Beach, Goto Islands, Japan was estimated by balloon aerial photography, in situ beach surveys, and leaching experiments in conjunction with a Fickian diffusion model analysis. Chromium (Cr), cadmium (Cd), tin (Sn), antimony (Sb), and lead (Pb) were detected in plastic litter collected during the beach surveys. Polyvinyl chloride (PVC) fishing floats contained the highest quantity of Pb. Balloon aerial photography in conjunction with a beach survey gave an estimated mass of Pb derived from plastic litter of 313 ± 247 g. Lead leaching experiments on collected PVC floats showed that Pb in the plastic litter could leach into surrounding water on the actual beach, and that plastic litter may act as a "transport vector" of toxic metals to the beach environment. Using the experimental data, the total mass of Pb that could leach from PVC plastic litter over a year onto Ookushi Beach was estimated as 0.6 ± 0.6 g/year, suggesting that toxic metals derived from plastic beach litter are a potential "pathway" to contamination of the beach environment due to their accumulation in beach soil over time.

A new technique for detecting colored macro plastic debris on beaches using webcam images and CIELUV.

We have developed a technique for detecting the pixels of colored macro plastic debris (plastic pixels) using photographs taken by a webcam installed on Sodenohama beach, Tobishima Island, Japan. The technique involves generating color references using a uniform color space (CIELUV) to detect plastic pixels and removing misdetected pixels by applying a composite image method. This technique demonstrated superior performance in terms of detecting plastic pixels of various colors compared to the previous method which used the lightness values in the CIELUV color space. We also obtained a 10-month time series of the quantity of plastic debris by combining a projective transformation with this technique. By sequential monitoring of plastic debris quantity using webcams, it is possible to clean up beaches systematically, to clarify the transportation processes of plastic debris in oceans and coastal seas and to estimate accumulation rates on beaches.

Low altitude remote-sensing method to monitor marine and beach litter of various colors using a balloon equipped with a digital camera.

This study aims to establish a low-altitude remote sensing system for surveying litter on a beach or the ocean using a remote-controlled digital camera suspended from a balloon filled with helium gas. The resultant images are processed to identify the litter using projective transformation method and color difference in the CIELUV color space. Low-altitude remote sensing experimental observations were conducted on two locations in Japan. Although the sizes of the litter and the areas covered are distorted in the original photographs taken at various angles and heights, the proposed image process system is capable of identifying object positions with a high degree of accuracy (1-3 m). Furthermore, the color difference approach in the CIELUV color space used in this study is well capable of extracting pixels of litter objects of various colors allowing us to estimate the number of objects from the photographs.

Using aerial photography and in situ measurements to estimate the quantity of macro-litter on beaches.

This study has demonstrated a reliable method of quantifying the total mass of litter on a beach. It was conducted on Ookushi beach, Goto-Islands, Japan, and uses a combination of balloon-assisted aerial photography and in situ mass measurements. The total mass of litter over the beach was calculated to be 716±259kg. This figure was derived by multiplying the litter-covered area (calculated using balloon-assisted aerial photography) by the mass of litter per unit area. Light plastics such as polyethylene made up 55% of all plastic litter on the beach, although more work is needed to determine whether lighter plastics are transported to beaches more readily by winds and ocean currents compared with heavier plastics, or whether lighter plastics comprise a greater percentage of marine litter. Finally, the above estimates were used to calculate the total mass of metals released into coastal ecosystems via plastic litter on beaches.

Establishment of numerical beach-litter hindcast/forecast models: an application to Goto Islands, Japan.

This study attempts to establish a system for hindcasting/forecasting the quantity of litter reaching a beach using an ocean circulation model, a two-way particle tracking model (PTM) to find litter sources, and an inverse method to compute litter outflows at each source. Twelve actual beach survey results, and satellite and forecasted wind data were also used. The quantity of beach litter was hindcasted/forecasted using a forward in-time PTM with the surface currents computed in the ocean circulation model driven by satellite-derived/forecasted wind data. Outflows obtained using the inverse method was given for each source in the model. The time series of the hindcasted/forecasted quantity of beach litter were found consistent with the quantity of beach litter determined from sequential webcam images of the actual beach. The accuracy of the model, however, is reduced drastically by intense winds such as typhoons which disturb drifting litter motion.

Sequential monitoring of beach litter using webcams.

This study attempts to establish a system for the sequential monitoring of beach litter using webcams placed at the Ookushi beach, Goto Islands, Japan, to establish the temporal variability in the quantities of beach litter every 90 min over a one and a half year period. The time series of the quantities of beach litter, computed by counting pixels with a greater lightness than a threshold value in photographs, shows that litter does not increase monotonically on the beach, but fluctuates mainly on a monthly time scale or less. To investigate what factors influence this variability, the time derivative of the quantity of beach litter is compared with satellite-derived wind speeds. It is found that the beach litter quantities vary largely with winds, but there may be other influencing factors.