WaterNet: A Convolutional Neural Network for Chlorophyll-a Concentration Retrieval

Abstract

The retrieval of chlorophyll-a (Chl-a) concentrations relies on empirical or analytical analyses, which generally experience difficulties from the diversity of inland waters in statistical analyses and the complexity of radiative transfer equations in analytical analyses, respectively. Previous studies proposed the utilization of artificial neural networks (ANNs) to alleviate these problems. However, ANNs do not consider the problem of insufficient in situ samples during model training, and they do not fully utilize the spatial and spectral information of remote sensing images in neural networks. In this study, a two-stage training is introduced to address the problem regarding sample insufficiency. The neural network is pretrained using the samples derived from an existing Chl-a concentration model in the first stage, and the pretrained model is refined with in situ samples in the second stage. A novel convolutional neural network for Chl-a concentration retrieval called WaterNet is proposed which utilizes both spectral and spatial information of remote sensing images. In addition, an end-to-end structure that integrates feature extraction, band expansion, and Chl-a estimation into the neural network leads to an efficient and effective Chl-a concentration retrieval. In experiments, Sentinel-3 images with the same acquisition days of in situ measurements over Laguna Lake in the Philippines were used to train and evaluate WaterNet. The quantitative analyses show that the two-stage training is more likely than the one-stage training to reach the global optimum in the optimization, and WaterNet with two-stage training outperforms, in terms of estimation accuracy, related ANN-based and band-combination-based Chl-a concentration models.

Deep Learning with WASI Simulation Data for Estimating Chlorophyll a Concentration of Inland Water Bodies

“…This finding is also met by Maier and Keller [37] on simulated Sentinel-2 resolution data. Other neural networks were applied by, e.g., Pahlevan et al [21], González Vilas et al [36], Chebud et al [38], whereas Syariz et al [40], Pu et al [41] applied CNNs on Landsat 8 and Sentinel-3 images to classify or estimate water quality.…”

“…Convolutional neural networks (CNNs) are one type of DL models. For example, they have been applied by Syariz et al [40], Pu et al [41] to either estimate chlorophyll a concentrations based on RS satellite data or to predict the water quality of different discrete classes. However, current studies only focus on the estimation and classification task for one single water body.…”

Retrieval of Water Quality from UAV-Borne Hyperspectral Imagery: A Comparative Study of Machine Learning Algorithms

“…Du et al investigated the tempo-spatial dynamics pattern of water quality in the Taihu Lake estuary using GOCI imagery [9]. Syariz et al used spectral and spatial information from Sentinel-3 images to retrieval the concentration of Chl-a [10]. Rajesh et al predicted the heavy metal concentration in water including Arsenic (As), cadmium (Cd), chromium (Cr), copper (Cu), iron (Fe), lead (Pb), nickel (Ni), zinc (Zn), aluminum (Al), cobalt (Co), manganese (Mg), beryllium (Be), boron (B), lithium (Li), molybdenum (Mo), selenium (Se), and vanadium (V), using Cartosat-2 data and measuring data [11].…”

Time Series Chlorophyll-A Concentration Data Analysis: A Novel Forecasting Model for Aquaculture Industry

“…Several studies have been conducted to establish a means of coping with water quality impairments caused by algal biomass using conventional numerical modelling methods, least squares support vector regression (LSSVR), neural networks methods such as Radial Basis Function neural network (RBFNN), Back Propagation neural network (BPNN) algorithms, and machine learning methods to predict chlorophyll-a concentrations as an indicator for future water quality changes [9][10][11][12]. However, the challenge with traditional numerical methods, LSSVR, and neural networks such as RBFNN and BPNN is the inherent weakness of the long-term dependency problem.…”