Monitoring the vertical distribution of HABs using hyperspectral imagery and deep learning models

Algal classification and Chlorophyll-a concentration determination using convolutional neural networks and three-dimensional fluorescence data matrices

In recent years, the frequency of harmful algal blooms has increased, leading to the release of large quantities of toxins and compounds that cause unpleasant odors and tastes, significantly compromising drinking water quality. Chlorophyll-a (Chl-a) is commonly used as a proxy for algal biomass. However, current methods for measuring Chl-a concentration face challenges in accurately quantifying algae by categories and effectively adapting to natural aquatic environments. This study combined convolutional neural networks (CNNs) and three-dimensional fluorescence data matrices to address these challenges. The algal classification model achieved over 99.5% accuracy in identifying thirteen types of algal samples, with class activation maps showing that the model primarily focused on algal pigment regions. In determining Chl-a concentrations of each algal species in mixed algae solutions (Microcystis aeruginosa, Cyclotella, and Chlorella), the Chl-a models demonstrated Mean Absolute Percentage Errors (MAPEs) ranging from 6.55% to 10.56% in the ultrapure water background, 11.57%-14.12% in the Qingcaosha Reservoir raw water background, and 21.46%-123.37% in the Lake Taihu raw water background. After calibration, the models were significantly improved, achieving MAPEs ranging from 11.86% to 14.18% in the Lake Taihu raw water background. Discrepancies in determination performance indicated that the intensity and locations of characteristic algal pigment fluorescence peaks greatly influenced the Chl-a models' accuracy. This research introduces a novel approach for algal classification and Chl-a concentration determination in water bodies, with significant potential for practical applications.

Estimating the vertical profile of water quality variables in reservoirs: Application of remotely sensed data and machine learning techniques

Water quality assessment and management of reservoirs depend on accurate, large-scale, and continuous monitoring of the vertical profile of Water Quality Variables (WQVs). Remote sensing data have been widely used to retrieve high spatiotemporal water quality data; however, their application has practically been limited to evaluating surface WQVs. In this paper, a novel and efficient approach is introduced for assessing the profile of WQVs in reservoirs that depend on stratification, by taking into account the shape of profile as prior knowledge. First, an appropriate function is fitted to the WQVs vertical profile, and second, the parameters of that function as representative of the WQVs vertical profile are estimated using machine learning techniques. The model's inputs are day, maximum depth of point, and remote sensing data. Finally, PAWN sensitivity analysis is applied to show the extent to which each input influences different parts of the vertical profile. This method is applied in the Wadi Dayqah Reservoir, the largest dam in Oman, to evaluate water temperature, dissolved oxygen, pH, and chlorophyll-a profile. The results show that the predicted profiles are properly representative of in situ measurements, with a mean absolute error of 0.28 °C, 0.25 mg/L, 0.052, and 0.33 μg/L on test data sets of water temperature, dissolved oxygen, pH, and chlorophyll-a, respectively. Finally, PAWN sensitivity analysis illustrates that satellite data not only influence the parameters representing surface WQVs but also contribute to the estimation of other curve parameters.

A new strategy based on multi-source remote sensing data for improving the accuracy of land use/cover change classification

Land Use/Cover Change (LUCC) plays a crucial role in sustainable land management and regional planning. However, contemporary feature extraction approaches often prove inefficient at capturing critical data features, thereby complicating land cover categorization. In this research, we introduce a new feature extraction algorithm alongside a Segmented and Stratified Principal Component Analysis (SS-PCA) dimensionality reduction method based on correlation grouping. These methods are applied to UAV LiDAR and UAV HSI data collected from land use types (e.g., residential areas, agricultural lands) and specific species (e.g., tree species) in urban, agricultural, and natural environments to reflect the diversity of the study area and to demonstrate the ability of our methods to be applied in different classification scenarios. We utilize LiDAR and HSI data to extract 157 features, including intensity, height, Normalized Digital Surface Model (nDSM), spectral, texture, and index features, to identify the optimal feature subset. Subsequently, the best feature subset is inputted into a random forest classifier to classify the features. Our findings demonstrate that the SS-PCA method successfully enhances downscaled feature bands, reduces hyperspectral data noise, and improves classification accuracy (Overall Accuracy = 91.17%). Additionally, the CFW method effectively screens appropriate features, thereby increasing classification accuracy for LiDAR (Overall Accuracy = 78.10%), HIS (Overall Accuracy = 89.87%), and LiDAR + HIS (Overall Accuracy = 97.17%) data across various areas. Moreover, the integration of LiDAR and HSI data holds promise for significantly improving ground fine classification accuracy while mitigating issues such as the 'salt and pepper noise'. Furthermore, among individual features, the LiDAR intensity feature emerges as critical for enhancing classification accuracy, while among single-class features, the HSI feature proves most influential in improving classification accuracy.

Simultaneous determination of pigments of spinach (Spinacia oleracea L.) leaf for quality inspection using hyperspectral imaging and multi-task deep learning regression approaches

Rapid and accurate determination of pigment content is important for quality inspection of spinach leaves during storage. This study aimed to use hyperspectral imaging at two spectral ranges (visible/near-infrared, VNIR: 400-1000 nm; NIR: 900-1700 nm) to simultaneously determine the pigment (chlorophyll a, chlorophyll b, total chlorophyll, and carotenoids) content in spinach stored at different durations and conditions (unpackaged and packaged). Partial least squares (PLS), back propagation neural network (BPNN) and convolutional neural network (CNN) were used to establish single-task and multi-task regression models. Single-task CNN (STCNN) models and multi-task CNN (MTCNN) models obtained better performances than the other models. The models using VNIR spectra were superior to those using NIR spectra. The overall results indicated that hyperspectral imaging with multi-task learning could predict the quality attributes of spinach simultaneously for spinach quality inspection under various storage conditions. This research will guide food quality inspection by simultaneously inspecting multiple quality attributes.

Three-dimensional augmentation for hyperspectral image data of water quality: An Integrated approach using machine learning and numerical models

This research introduces a comprehensive methodology to enhance hyperspectral image data (HSD) utility, specifically focusing on the three-dimensional (3-D) augmentation of Chlorophyll-a (Chl-a). This study comprises three significant steps: (1) the augmentation of limited field water quality data in terms of time interval and number of variables using neural network models, (2) the generation of 3-D data using numerical models, and (3) the extension of the hyperspectral image data into 3-D data using machine learning models. In the first phase, Multilayer Perceptron (MLP) models were developed to train water quality interactions and successfully generated high-frequency water quality data by adjusting biased measurements and predicting detailed water quality variables. In the second phase, high-frequency data generated by MLP models were applied to develop two numerical models. These numerical models successfully generated 3-D data, thereby demonstrating the effectiveness of integrating numerical modeling with neural networks. In the final phase, ten machine learning models were trained to generate 3-D Chl-a data from HSD. Notably, the Gaussian Process Regression model exhibited superior performance, effectively estimating 3-D Chl-a data with robust accuracy, as evidenced by an R-square value of 0.99. The findings align with theories of algal bloom dynamics, further validating the effectiveness of the approach. This study demonstrated the successful integrated development for HSD extension using machine learning models, numerical models, and original HSD, highlighting the potential of such integrated methodologies in advancing water quality monitoring and estimation. Notably, the approach leverages readily accessible data, allowing for the swift generation of results and bypassing time-consuming data collection processes. This research marks a significant step towards more robust, comprehensive water quality monitoring and prediction, thereby facilitating better management of aquatic ecosystems.

Deep learning-based efficient drone-borne sensing of cyanobacterial blooms using a clique-based feature extraction approach

Recent advances in remote sensing techniques provide a new horizon for monitoring the spatiotemporal variations of harmful algal blooms (HABs) using hyperspectral data in inland water. In this study, a hierarchical concatenated variational autoencoder (HCVAE) is proposed as an efficient and accurate deep learning (DL) based bio-optical model. To demonstrate its usefulness in retrieving algal pigments, the HCVAE is applied to bloom-prone regions in Daecheong Lake, South Korea. By abstracting the similarity between highly related features using layer-wise clique-based latent-feature extraction, HCVAE reduces the computational loads in deriving outputs while preventing performance degradation. Graph-based clique-detection uses information theory-based criteria to group the related reflectance spectra. Consequently, six latent features were extracted from 79 spectral bands to consist of a multilevel hierarchy of HCVAE that can simultaneously estimate concentrations of chlorophyll-a (Chl-a) and phycocyanin (PC). Despite the parsimonious model architecture, the Chl-a and PC concentrations estimated by HCVAE closely agree with the measured concentrations, with test R2 values of 0.76 and 0.82, respectively. In addition, spatial distribution maps of algal pigments obtained from HCVAE using drone-borne reflectance successfully capture the blooming spots. Based on its multilevel hierarchical architecture, HCVAE can provide the importance of latent features along with their individual wavelengths using Shapley additive explanations. The most important latent features covered the spectral regions associated with both Chl-a and PC. The lightweight neural network DNNsel, which uses only the spectral bands of highest importance in latent-feature extraction, performed comparably to HCVAE. The study results demonstrate the utility of the multilevel hierarchical architecture as a comprehensive assessment model for near-real-time drone-borne sensing of HABs. Moreover, HCVAE is applicable to a wide range of environmental big data, as it can handle numerous sets of features.

Deep learning-based classification of microalgae using light and scanning electron microscopy images

Microalgae possess diverse applications, such as food production, animal feed, cosmetics, plastics manufacturing, and renewable energy sources. However, uncontrolled proliferation, known as algal bloom, can detrimentally impact ecosystems. Therefore, the accurate detection, monitoring, identification, and tracking of algae are imperative, albeit demanding considerable time, effort, and expertise, as well as financial resources. Deep learning, employing image pattern recognition, emerges as a practical and promising approach for rapid and precise microalgae cell counting and identification. In this study, we processed light microscopy (LM) and scanning electron microscopy (SEM) images of two Cyanobacteria species and three Chlorophyta species to classify them, utilizing state-of-the-art Convolutional Neural Network (CNN) models, including VGG16, MobileNet V2, Xception, NasnetMobile, and EfficientNetV2. In contrast to prior deep learning based identification studies limited to LM images, we, for the first time, incorporated SEM images of microalgae in our analysis. Both LM and SEM microalgae images achieved an exceptional classification accuracy of 99%, representing the highest accuracy attained by the VGG16 and EfficientNetV2 models to date. While NasnetMobile exhibited the lowest accuracy of 87% with SEM images, the remaining models achieved classification accuracies surpassing 93%. Notably, the VGG16 and EfficientNetV2 models achieved the highest accuracy of 99%. Intriguingly, our findings indicate that algal identification using optical microscopes, which are more cost-effective, outperformed electron microscopy techniques.