Rapid COD Sensing in Complex Surface Water Using Physicochemical-Informed Spectral Transformer with UV–Vis-SWNIR Spectroscopy

Environmental Science & Technology, Год журнала: 2025, Номер unknown

Опубликована: Март 7, 2025

Water, as a finite and vital resource, necessitates water quality monitoring to ensure its sustainable use. A key aspect of this process is the accurate measurement critical parameters such chemical oxygen demand (COD). However, current spectroscopic methods struggle with accurately consistently measuring COD in large-scale, complex environments due an insufficient understanding spectra limited generalizability. To address these limitations, we introduce physicochemical-informed spectral Transformer (PIST) model, combined ultraviolet-visible-shortwave-near-infrared (UV-vis-SWNIR) spectroscopy for sensing. best our knowledge, first approach combine PIST integrates block incorporate existing physical information into encoding domain adaptation, along feature embedding comprehensive features extraction. We validated using actual surface data set extensive geographic coverage including Yangtze River Poyang Lake. demonstrated notable performance sensing within environments, achieving impressive R2 value 0.9008 reducing root mean squared error (RMSE) by 45.20% 29.38% compared benchmark models support vector regression (SVR) convolutional neural network (CNN). These results emphasize PIST's accuracy generalizability, marking significant advancement multidisciplinary approaches that deep learning rapid

Язык: Английский

---

Prediction of Water Chemical Oxygen Demand with Multi-Scale One-Dimensional Convolutional Neural Network Fusion and Ultraviolet–Visible Spectroscopy

Biomimetics, Год журнала: 2025, Номер 10(3), С. 191 - 191

Опубликована: Март 20, 2025

Chemical oxygen demand (COD) is a critical parameter employed to assess the level of organic pollution in water. Accurate COD detection essential for effective environmental monitoring and water quality assessment. Ultraviolet–visible (UV-Vis) spectroscopy has become widely applied method due its convenience absence need chemical reagents. This non-destructive reagent-free approach offers rapid reliable means analyzing Recently, deep learning emerged as powerful tool automating process spectral feature extraction improving prediction accuracy. In this paper, we propose novel multi-scale one-dimensional convolutional neural network (MS-1D-CNN) fusion model designed specifically prediction. The architecture proposed involves inputting raw UV-Vis spectra into three parallel sub-1D-CNNs, which independently data. outputs from final convolution pooling layers each sub-CNN are then fused single layer, capturing rich set features. output subsequently passed through Flatten layer followed by fully connected predict value. Experimental results demonstrate effectiveness method, it was compared with traditional methods on same dataset. MS-1D-CNN showed significant improvement accuracy prediction, highlighting potential more efficient monitoring.

Язык: Английский

---

Seawater nitrate assessment using a correction algorithm for temperature and pressure up to 100 MPa

Measurement, Год журнала: 2025, Номер unknown, С. 117526 - 117526

Опубликована: Апрель 1, 2025

Язык: Английский

---

Nondestructive detection of internal quality in multiple peach varieties by Vis/NIR spectroscopy with multi-task CNN method

Postharvest Biology and Technology, Год журнала: 2025, Номер 227, С. 113579 - 113579

Опубликована: Апрель 19, 2025

Язык: Английский

---