DEA: Hyperspectral data high-throughput extraction and analysis software

• Data extraction Capability: DEA software supports high-throughput batch processing of hyperspectral data, enabling rapid data extraction. • Image segmentation Algorithms: DEA software incorporates various image segmentation methods, including traditional algorithms and techniques based on deep learning. • Regression Analysis: DEA software integrates particle swarm optimization and an improved sparrow search algorithm to optimize hyperparameters in regression models. • Performance Evaluation: the performance of DEA software was evaluated and compared with other software in terms of image segmentation, data extraction speed, and regression analysis accuracy. • Graphical interface and auxiliary Functions: DEA software includes a graphical interface and provides multiple auxiliary functions to facilitate hyperspectral data analysis. Hyperspectral imaging is a vital tool in phenomics, particularly for non-destructive testing. However, existing analysis methods often rely on commercial software or custom programming, which limits efficiency and accessibility. To address these challenges, we developed DEA (Data Extraction and Analysis), a Windows-based software for high-throughput hyperspectral data processing. DEA integrates modules for data extraction, preprocessing, and analysis to support batch processing of hyperspectral data, including grayscale, binary, and reflectance data. In terms of image segmentation, this study evaluated various methods on samples collected under different environmental conditions, with results demonstrated DEA's robustness in generating binary images. Building on this performance, comparative analyses with conventional tools suggest DEA's advantage in data extraction efficiency, defined in terms of faster data processing speed and improved model accuracy compared to existing methods such as ENVI, SPSS, etc. For regression modeling, DEA employs advanced optimization algorithms, including particle swarm optimization and improved sparrow search algorithms, achieving superior accuracy. On nine metabolites from rice datasets, DEA demonstrated R² improvements across all metabolites, with the highest improvement reaching 0.12, outperforming other tools. DEA provides an efficient and accessible solution for hyperspectral data processing, particularly benefiting researchers without programming expertise while addressing limitations of existing tools.