Underwater Marine Life and Plastic Waste Detection Using Deep Learning and Raspberry Pi

Hegde, Rahul; Patel, Sanobar; Naik, Rosha G; Nayak, Sagar N.; Shivaprakasha, K. S.; Bhandarkar, Rekha

doi:10.1007/978-981-16-0443-0_22

Cited by 8 publications

(6 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…2 shows the distribution of water bodies and dataset sources in the reviewed literature. Some papers are counted multiple times since they either consider multiple water bodies (Jakovljevic et al, 2020;Kylili et al, 2020;Panwar et al, 2020;Watanabe et al, 2019;Wolf et al, 2020) or employ multiple dataset sources (Hegde et al, 2021;Watanabe et al, 2019;Wu et al, 2020). Most studies dealt with macroplastic pollution in real settings, with the exception of two studies that considered a controlled artificial environment (AE).…”

Section: Water Bodies Polluted By Macroplastic Littermentioning

confidence: 99%

“…Battula et al (2020) extracted images from a Kaggle dataset 1 and labeled images with bounding boxes to train and test OD model. Hegde et al (2021) retrieved unlabelled images from Google and manually created the annotations to develop and test their models.…”

Section: Employed Dataset Sourcesmentioning

confidence: 99%

“…Five of these studies (Bajaj et al, 2021;Fulton et al, 2019;Marin et al, 2021;Xue et al, 2021bXue et al, , 2021a extracted data from the deep-sea debris database 2 provided by Japan Agency for Marine-earth Science and Technology. Four studies (Hegde et al, 2021;Marin et al, 2021;Musić et al, 2020;Wu et al, 2020) retrieved images from internet. While authors of these studies had to manually produce the annotations, one study (Deng et al, 2021) directly utilized the images with annotations from the TrashCan dataset (Hong et al, 2020).…”

Section: Employed Dataset Sourcesmentioning

confidence: 99%

“…Two studies deployed DL models, Resnet-50 neural network (Battula et al, 2020) and SSD MobileNet V2 (Hegde et al, 2021) to perform OD in edge computing devices, e.g., processing boards connected to fixed cameras or installed in ROVs. For example, Hegde et al (2021) stored a trained detection model in a Raspberry Pi board. The device used the model to detect macroplastic litter from the surrounding environment as sampled by the attached underwater camera.…”

Section: Model Architectures For Each CV Taskmentioning

confidence: 99%

See 3 more Smart Citations

Deep learning for detecting macroplastic litter in water bodies: A review

Jia

Kapelan

Vries³

et al. 2023

Water Research

View full text Add to dashboard Cite

Section: Water Bodies Polluted By Macroplastic Littermentioning

confidence: 99%

Section: Employed Dataset Sourcesmentioning

confidence: 99%

Section: Employed Dataset Sourcesmentioning

confidence: 99%

Section: Model Architectures For Each CV Taskmentioning

confidence: 99%

See 2 more Smart Citations

Deep learning for detecting macroplastic litter in water bodies: A review

Jia

Kapelan

Vries³

et al. 2023

Water Research

View full text Add to dashboard Cite

“…The model is highly influenced by the spectral signature of litter and water, as well as the geometrical optics of the sensor. (2) Deep learning models aim to automate the detection of litter through the application of computer vision tasks such as image classification (Jakovljević et al 2019 ), object detection (Hegde et al 2021 ), and image segmentation (Mifdal et al 2021 ). (3) Indices method aims to develop an index that discriminates between litter and water based on their spectral signatures.…”

Section: Introductionmentioning

confidence: 99%

Machine learning-based detection and mapping of riverine litter utilizing Sentinel-2 imagery

Mohsen

Kiss

Kovács

2023

Environ Sci Pollut Res

View full text Add to dashboard Cite

Despite the substantial impact of rivers on the global marine litter problem, riverine litter has been accorded inadequate consideration. Therefore, our objective was to detect riverine litter by utilizing middle-scale multispectral satellite images and machine learning (ML), with the Tisza River (Hungary) as a study area. The Very High Resolution (VHR) images obtained from the Google Earth database were employed to recognize some riverine litter spots (a blend of anthropogenic and natural substances). These litter spots served as the basis for training and validating five supervised machine-learning algorithms based on Sentinel-2 images [Artificial Neural Network (ANN), Support Vector Classifier (SVC), Random Forest (RF), Naïve Bays (NB) and Decision Tree (DT)]. To evaluate the generalization capability of the developed models, they were tested on larger unseen data under varying hydrological conditions and with different litter sizes. Besides the best-performing model was used to investigate the spatio-temporal variations of riverine litter in the Middel Tisza. According to the results, almost all the developed models showed favorable metrics based on the validation dataset (e.g., F1-score; SVC: 0.94, ANN: 0.93, RF: 0.91, DT: 0.90, and NB: 0.83); however, during the testing process, they showed medium (e.g., F1-score; RF:0.69, SVC: 0.62; ANN: 0.62) to poor performance (e.g., F1-score; NB: 0.48; DT: 0.45). The capability of all models to detect litter was bounded to the pixel size of the Sentinel-2 images. Based on the spatio-temporal investigation, hydraulic structures (e.g., Kisköre Dam) are the greatest litter accumulation spots. Although the highest transport rate of litter occurs during floods, the largest litter spot area upstream of the Kisköre Dam was observed at low stages in summer. This study represents a preliminary step in the automatic detection of riverine litter; therefore, additional research incorporating a larger dataset with more representative small litter spots, as well as finer spatial resolution images is necessary.

show abstract