Red Tide Detection Method for HY−1D Coastal Zone Imager Based on U−Net Convolutional Neural Network

Abstract

Existing red tide detection methods have mainly been developed for ocean color satellite data with low spatial resolution and high spectral resolution. Higher spatial resolution satellite images are required for red tides with fine scale and scattered distribution. However, red tide detection methods for ocean color satellite data cannot be directly applied to medium–high spatial resolution satellite data owing to the shortage of red tide responsive bands. Therefore, a new red tide detection method for medium–high spatial resolution satellite data is required. This study proposes the red tide detection U−Net (RDU−Net) model by considering the HY−1D Coastal Zone Imager (HY−1D CZI) as an example. RDU−Net employs the channel attention model to derive the inter−channel relationship of red tide information in order to reduce the influence of the marine environment on red tide detection. Moreover, the boundary and binary cross entropy (BBCE) loss function, which incorporates the boundary loss, is used to obtain clear and accurate red tide boundaries. In addition, a multi−feature dataset including the HY−1D CZI radiance and Normalized Difference Vegetation Index (NDVI) is employed to enhance the spectral difference between red tides and seawater and thus improve the accuracy of red tide detection. Experimental results show that RDU−Net can detect red tides accurately without a precedent threshold. Precision and Recall of 87.47% and 86.62%, respectively, are achieved, while the F1−score and Kappa are 0.87. Compared with the existing method, the F1−score is improved by 0.07–0.21. Furthermore, the proposed method can detect red tides accurately even under interference from clouds and fog, and it shows good performance in the case of red tide edges and scattered distribution areas. Moreover, it shows good applicability and can be successfully applied to other satellite data with high spatial resolution and large bandwidth, such as GF−1 Wide Field of View 2 (WFV2) images.

Trend of harmful algal bloom dynamics from GOCI observed diurnal variation of chlorophyll a off Southeast coast of China

“…In recent years, neural network methods have been increasingly used in HAB prediction. The prediction mechanism is often formed by evolutionary computation (Recknagel et al, 2014), autoregressive fuzzy models (Kim et al, 2014), and artificial neural network models (Zhao et al, 2022). Nevertheless, these models are complex to establish and have limitations when dealing with nonstationary data (Cannas et al, 2006;Wang et al, 2011).…”

Red Tide Detection Method Based on Improved U-Net Model-Taking GOCI Data in East China Sea as an Example

“…In 2020, Lee et al developed a novel red tide detection scheme for the southern coast of the Korean Peninsula using multi-layer feedforward neural networks based on deep learning and high spatio-temporal resolution satellites [14]. In 2022, Zhao et al proposed the RDU-Net network based on the HY-1D coastal imager, which improved the detection accuracy of deep learning at the edge of red tides and scattered areas [15]. The above research verifies the feasibility of using deep learning for remote sensing red tide detection.…”

Red Tide Detection Based on Improved DenseNet Network—Example of Red Tide Detection from Geostationary Ocean Color Imager Data in Bohai Sea

“…Over the past few years, with the development of deep learning, an increasing number of researchers have begun to use deep learning to detect red tides. (22)(23)(24)(25) In 2019, Kim et al proposed a U-Net network for a study related to the detection of red tide using GOCI data from the waters surrounding the Korean Peninsula. U-Net is a U-shaped convolutional neural network (CNN) with an encoder and decoder structure, which is widely used for pixel classification or target segmentation.…”