We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Hyperspectral detection of soil microplastics via multimodal feature fusion and a dual-path attention residual convolutional network
Summary
A hyperspectral imaging approach combined with multimodal deep learning was developed to detect microplastics in soil, achieving high accuracy in identifying plastic particles against complex soil backgrounds. The method offers a faster, less destructive alternative to traditional chemical extraction and spectroscopy for soil monitoring.
Rapid and precise detection of soil microplastics is crucial for environmental risk assessment but is challenged by complex matrices and low concentrations. To overcome the limitations of single-modal analysis, we developed a novel multimodal hyperspectral framework. This method integrates features from both one-dimensional (1D) spectral data and their two-dimensional (2D) image representations using a Multi-view Probabilistic Feature Fusion (MPFF) strategy, followed by classification with a Dual-path Attention Residual Convolutional Network (DAR-CNN). The integrated model achieved a classification accuracy of 96.75 %, outperforming conventional models. Notably, the framework maintained robust performance even at a low concentration of 0.5 %. This work provides an effective framework for monitoring soil microplastics and advances the application of multimodal information fusion in hyperspectral analysis.