Eelgrass Beds and Oyster Farming in a Lagoon Before and After The Great East Japan Earthquake of 2011: Potential for Applying Deep Learning at a Coastal Area

Abstract

There is a small number of case studies of automatic land cover classification on the coastal area. Here, I test extraction of seagrass beds, sandy area, oyster farming rafts at Mangoku-ura Lagoon, Miyagi, Japan by comparing manual tracing, simple image segmentation, and image transformation using deep learning. The result was used to extract the changes before and after the earthquake and tsunami. The output resolution was best in the image transformation method, which showed more than 69% accuracy for vegetation classification by an assessment using random points on independent test data. The distribution of oyster farming rafts was detected by the segmentation model. Assessment of the change before and after the earthquake by the manual tracing and image transformation result revealed increase of sand area and decrease of the vegetation. By the segmentation model only the decrease of the oyster farming was detected. These results demonstrate the potential to extract the spatial pattern of these elements after an earthquake and tsunami.

The Asia‐Pacific Biodiversity Observation Network: 10‐year achievements and new strategies to 2030

“…Conducting up‐to‐date statistical analyses of their changes from the past in marine biodiversity should be readily evaluated in this region where marine ecosystem is rapidly changing with economic developments and global climate changes. Moreover, AP‐MBON intends to promote following research and outreaching activities such as; (a) Continuing monitoring of marine biodiversity in the AP region (including citizen science programs) in a globally‐coordinated programs such as Reef Life Survey (Edgar et al, 2017), Reef Check (Chelliah et al, 2015), Seagrass Watch (McKenzie et al, 2000) and Global Ocean Observing System (GOOS), (b) Analyzing broad‐scale, long‐term changes in marine biodiversity in AP Region using global databases such as OBIS and GBIF; (c) Supporting design and adaptive governance of marine protected areas based on scientific data on marine biodiversity (Yamakita et al, 2015; Yamakita et al, 2017); (d) Analyzing values of ecosystem services, its trends, and its linkage to human society by social‐ecological studies (Nakaoka et al, 2018); (e) Developing and apply cross cutting‐edge technologies to marine biodiversity research such as meta‐barcoding/environmental DNA (eDNA) (Miya et al, 2015), and remote sensing/GIS/Deep learning technologies (Yamakita, 2019; Yamakita, Sodeyama, Whanpetch, Watanabe, & Nakaoka, 2019); (f) Outreach of scientific outputs to various types of stakeholders of marine biodiversity and developing sustainable management plans via codesign, coproduction and co‐delivery (Yamakita, 2019; Yamakita et al, 2019).…”

“…The 3D modeling and CT scanned data would also contribute to the development of morphology and taxonomy (see recent data in “ffish‐asia”; Kano et al, 2013, Table 1). Image recognition and analysis is also essential for biodiversity monitoring by deep/machine learning (e.g., satellite map, Samasse, Hanan, Anchang, & Diallo, 2020; seagrass, Yamakita, 2019). In addition, data representation techniques such as online dashboard are also expected for more efficient use of the data.…”

Mapping of Subtidal and Intertidal Seagrass Meadows via Application of the Feature Pyramid Network to Unmanned Aerial Vehicle Orthophotos

“…Deep learning methods have gained popularity for coping with the generalization problem owing to their advantages in digging semantic values after convolution. For example, Yamakita et al [21] confirmed that a deep convolutional generative adversarial network showed better results than the fully convolutional network for classifying different objects, including seagrass, by using aerial and satellite images (QuickBird). However, the deep convolutional generative adversarial network may exhibit limitations, such as collapsing generators, and is highly sensitive to hyper-parameter selections [22].…”

Application of deep learning techniques for determining the spatial extent and classification of seagrass beds, Trang, Thailand

“…the target image). Although the generator and discriminator use a kind of convolutional neural network that is the basic model of deep learning, cGAN shows higher accuracy in image recognition than other conventional machine learning methods, including supervised classification and other simpler deep learning methods (Goodfellow et al 2016, Creswell et al 2018, Yamakita 2018, 2019. cGAN was not specifically developed for remote sensing, but the model has been applied to automatic map production from satellite images (Isola et al 2016).…”