In the quality control of Chinese herbal medicines, hyperspectral characteristic chemical fingerprints serve as highly effective analytical parameters for authentication and adulteration screening. Due to intrinsic variations in tissue morphology and chemical compositions, distinct species and source-specific herbal samples exhibit characteristic absorption features across the visible to near-infrared (VNIR, $400\text{--}1000\text{ nm}$) and short-wave infrared (SWIR, $1000\text{--}2500\text{ nm}$) spectra. Hyperspectral imaging systems capture the reflectance spectra and spatial patterns of samples simultaneously. By extracting mean spectral profiles from specified Regions of Interest (ROIs), standard spectral libraries can be constructed to realize rapid classification of unknown specimens when paired with chemometric models. Addressing the pervasive market challenge of adulterated premium herbal materials—such as *Crocus sativus* (Saffron) mixed with dyed filaments, *Ophiocordyceps sinensis* weighted via metal splices, or *Panax ginseng* adulterated with *Platycodon grandiflorus*—hyperspectral techniques demonstrate significant performance advantages. Grounded in characteristic spectral band image texture analysis, chemometric chemometrics architectures including Principal Component Analysis (PCA), Partial Least Squares Discriminant Analysis (PLS-DA), and Support Vector Machines (SVM) can differentiate foreign zones unidentifiable by visual inspection and plot adulteration topology maps via pseudo-color visualizations. Compared to destructive chromatographic techniques (e.g., HPLC) or molecular protocols (e.g., DNA barcoding), HSI circumvents time-consuming sample pre-treatments, significantly curtails processing cycles, and is uniquely tailored for high-throughput screening of bulk commodities. For morphologically analogous and easily confused botanical pairs, such as *Bupleurum chinense* versus *Bupleurum scorzonerifolium*, *Ziziphus jujuba* var. *spinosa* versus *Ziziphus mauritiana*, or authentic *Rheum palmatum* versus counterfeit *Rheum franzenbachii*, hyperspectral imaging paired with deep learning classification networks enables precision discrimination of individual seeds or sliced decoction pieces within tight runtime bounds. This approach is widely adopted as an official supplementary verification protocol for Chinese herbal slices, playing a pivotal role in ensuring market distribution quality.

The geo-authenticity ("Daodi") of traditional herbal materials is explicitly governed by local growth habitats, yet classical organoleptic assessments remain hindered by subjective empirical variance. Hyperspectral imaging tracks non-targeted geographic origins by mapping intrinsic biochemical composition variations derived from regional climates, soil microbiomes, and environmental conditions. Quantitative evaluations show that concentrations of total flavonoids and astragalosides within *Astragalus membranaceus* profiles vary by habitat and present explicit signatures across NIR wavelengths. When integrated with Linear Discriminant Analysis (LDA) or Random Forest algorithms, samples harvested from primary production zones such as Gansu, Shanxi, and Inner Mongolia can be robustly classified. Within laboratory research infrastructures, the HG-HyperLab laboratory hyperspectral imaging system developed by Hagorun Technology Limited serves as a premium spectral data acquisition engine for training provenance validation models; utilizing its superior spectral resolution and structural spatial resolution, it maps minute chemical fingerprints corresponding to distinct spatial coordinates. Throughout botanical cultivation and harvesting sequences, HSI is strategically deployed for dynamic tracking of active constituent accumulation and optimal harvest window estimation. By scanning herbal specimens at consecutive phenological phases via laboratory spectrometers, calibration models such as Partial Least Squares Regression (PLSR) can be constructed to link raw spectral components with destination assay metrics, predicting chemical markers dynamically to direct automated harvesting. This system supersedes traditional destructive "sample-digging-chemical-assay" cycles, minimizing laboratory analytical overhead while enabling continuous, non-invasive lifecycle telemetry. For high-value powder matrices such as *Ganoderma lucidum* spore powders or *Cervus elaphus* (Velvet Antler) powders, hyperspectral chemical imaging computes chemical spatial uniformity and tracks adulteration topologies. By decomposing the spectral data of individual particles, the presence of localized starch, dextrin, or other low-cost bulking excipients can be mapped accurately, equipping herbal processing facilities with reliable, visible diagnostics for incoming raw material inspections and inline manufacturing quality assurance.

The manufacturing continuum transforming raw botanicals into finished pharmaceutical products involves multiple complex unit operations including cleaning, slicing, processing ("Paozhi"), extraction, concentration, dehydration, and blending. Quality discrepancies at any intermediate step compromise downstream consumer safety and clinical efficacy. Functioning as a core tool within the Process Analytical Technology (PAT) framework, hyperspectral imaging tracks Critical Quality Attributes (CQAs) during offline or at-line herbal processing sequences. For instance, across traditional processing methods involving rice wine, vinegar, or honey stir-frying, real-time HSI datasets monitor surface colorimetric variations, moisture dynamics, and excipient diffusion depth, guiding automated end-point determination and minimizing manual operator bias. In the production of solid oral dosage forms, hyperspectral imaging screens active ingredient blending uniformity across tablet matrices, identifying coating defects, capping phenomena, or uneven component loading. For herbal extract powder blend processes, the system evaluates composite powders during mixing phases, deploying homogeneity validation algorithms (e.g., correlation coefficients, relative standard deviation) to monitor blend end-points and improve batch-to-batch consistency. For TCM formula granule quality assessment, HSI combined with deep learning architectures classifies product grades and sorts structural properties, paving the technical way for miniaturized, handheld near-infrared hyperspectral screening devices suitable for point-of-care hospital pharmacies or retail settings. As regulatory mandates for pharmaceutical supply-chain traceability intensify, the high-dimensional data cubes generated by hyperspectral instruments can be cross-linked with localized data infrastructures to output a unique "spectral identity card". This methodology establishes a totally traceable, quantifiable, and visualized quality assurance paradigm stretching across agricultural cultivation, primary processing, and final pharmaceutical formulation.

Interested in Advancing Hyperspectral Solutions?

Our engineering group delivers advanced technical consulting and integrated solutions

Hagorun Technology Limited  |  Focus · Dedication · Exploration · Vision