Portable Hyperspectral Imaging from 400 to 1000 nm
The LiSen Optics iSpecHyper-VS1000 portable hyperspectral imaging system combines straightforward operation with flexible deployment. Its proprietary transmission-grating spectrograph, large-format detector, and optimized optics provide high sensitivity, a high signal-to-noise ratio, and consistent image quality across the detector.
Integrated data acquisition and analysis, with USB 3.0 data transfer, support an efficient spectral imaging workflow. Interchangeable lenses and a standard C-mount for manual-focus lenses support applications in forensics, precision agriculture and forestry, remote sensing, industrial inspection, biomedical research, and mineral exploration.
Key Features
- Fine spectral resolution. Resolve spectral features with a resolution of <2.8 nm.
- High optical throughput. The transmission-grating spectrograph provides high sensitivity and signal-to-noise performance.
- Automated operation. Motorized autofocus, automatic exposure, synchronized scanning, and laser ranging help streamline image acquisition.
- Real-time monitoring. High-frame-rate acquisition and a live auxiliary-camera view support imaging and scene monitoring.
- Portable power. An integrated lithium-ion battery supports operation in the field.
- Image quality across the field. The optical design maintains a spot diameter of <0.5 pixel across the full field.
- Flexible analysis and optics. ENVI-compatible data output, multiple regions of interest (ROIs), and interchangeable lenses support a range of imaging tasks.
Technical Specifications
| Parameter | iSpecHyper-VS1000 |
|---|---|
| Product Image |
 |
| Detector | CMOS |
| Spectral Range | 400–1000 nm |
| Spectrograph | Transmission grating |
| Spectral Resolution | Better than 2.8 nm |
| Spectral Channels | 300 |
| Slit Width | 25 µm |
| Transmission Efficiency | >60% |
| F-number | F/2.6 |
| Imaging Method | Built-in push-broom scanning |
| Spatial Pixels | 1920 (1×), 960 (2×), 480 (4×) |
| Pixel Size | 5.86 µm |
| Effective Bit Depth | 12-bit |
| Horizontal Scan Angle | >40° @ F = 35 mm |
| Vertical Scan Range | >15.6° @ F = 35 mm |
| Standard Lens Focal Length | 35 mm |
| Optional Focal Lengths | Manual-focus lenses: 12.5 / 16 / 25 / 35 mm; Autofocus lenses: 24 / 35 mm |
| Working Distance | 350 mm to infinity (with 35 mm lens) |
| Acquisition Time | ≤5 s |
| Lens Mount | C-mount (manual-focus lenses only) |
| Data Interface | GigE / USB |
| Auxiliary Camera | 5 MP |
| Laser Rangefinder | Yes |
| Focus | Manual / autofocus |
| Power Supply | Integrated battery; external power supported |
| Data Acquisition System | iSpecHyper-View hyperspectral acquisition software; CPU: Intel Core i5; RAM: 16 GB; storage: 512 GB |
| Operating Temperature | 0–40 °C |
| Storage Temperature | 0–50 °C |
| Weight | <7 kg |
| Included Accessories | Outdoor/indoor illumination source; reflectance reference panel; field tripod; data cable; power cable; external power module; custom rugged carrying case |
Example System Configuration
The example configuration below shows the imaging unit with controlled illumination, a support stand, plant samples, an external control unit, and a laptop running the imaging software.
Acquisition and Analysis Software
The iSpecHyper-View acquisition software supports hyperspectral image capture and analysis. The specified acquisition computer has an Intel Core i5 processor, 16 GB of RAM, and 512 GB of storage. The interfaces below show image-region selection, spectral profiles, acquisition controls, and live monitoring.
Hyperspectral Imaging Applications
The following examples illustrate how hyperspectral imaging combines spatial and spectral information across different applications. The cited studies describe their own instruments and experimental conditions.
Fruit Sorting
Nondestructive quality inspection and grading are increasingly important in agricultural processing and modern farming. Internal fruit quality reflects physiological, chemical, and physical properties. Quality-related spectral information is concentrated mainly in the 650–950 nm range. Sugar content is an important quality measure, with characteristic features reported in the 700–820 nm region, around 750 nm, and between 800 and 900 nm.
Industrial fruit-sorting systems typically use conveyor motion to provide the push-broom scan. A hyperspectral camera is mounted above the belt on a gantry and paired with a dedicated line light for controlled illumination. A complete system generally includes the camera and mount, line illumination, control module, positioning sensors, and a computer running control and acquisition software.
Blood Oxygenation Research
The 2015 study Hyperspectral optical tomography of intrinsic signals in the rat cortex used hyperspectral imaging to examine the rat cerebral cortex. The researchers reported wavelength-sensitive changes associated with oxyhemoglobin and deoxyhemoglobin at 529 and 630 nm, respectively. They developed a hyperspectral diffuse optical tomography (DOT) algorithm for rapid assessment of blood oxygenation.
Research example: Konecky et al., 2015.
Photosynthesis Research
The 2017 paper Kleptoplast photosynthesis is nutritionally relevant in the sea slug Elysia viridis investigated photosynthetic activity in this sea slug, which feeds on macroalgae and retains chloroplasts in cells of its digestive tubules. Hyperspectral imaging was used to examine chloroplast abundance, distribution, and photosynthetic function. Chloroplast abundance decreased markedly after 24 days of starvation in darkness, indicating that retained chloroplasts can be broken down to help meet energy demands under severe nutrient stress.
Research example: Cartaxana et al., 2017.
Biomedical Research
The 2012 paper Hyperspectral imaging and spectral-spatial classification for cancer detection developed and evaluated an image-processing and least-squares support vector machine (LS-SVM) method for hyperspectral prostate-cancer detection. Applied to tumor-bearing mice, it produced spatially resolved images highlighting reflectance differences between cancerous and normal tissue. Preliminary animal-model results indicated reliable tumor detection and demonstrated the potential of hyperspectral imaging for optical cancer diagnosis.
Research example: Fei, Akbari, and Halig, 2012.
Studies by Houzhu Ding and colleagues (2015) and Michael S. Chin and colleagues (2015) used hyperspectral imaging to assess burn depth and recovery in animal models. The left panel maps oxygen saturation and hemoglobin at 0, 1, 4, and 24 hours after injury; the right panel compares color imaging, an oxyhemoglobin map, and histology. The results demonstrate the potential for noninvasive, noncontact burn-depth assessment.
Biological Species Classification
The 2013 study Non-Invasive Measurement of Frog Skin Reflectivity in High Spatial Resolution Using a Dual Hyperspectral Approach used two push-broom imaging systems covering 400–2500 nm to acquire reflectance images of three arboreal frog species. Although all three appeared green to the human eye, their reflectance differed substantially between 700 and 1100 nm, enabling species discrimination.
Research example: Pinto et al., 2013.
Cultural Heritage and Archaeology
Discovered in 1974, the Terracotta Army attracted worldwide attention and was described by former French President Jacques Chirac as the “Eighth Wonder of the World.” After more than 2,000 years underground, the artifacts can deteriorate rapidly when exposed to air, making conservation difficult.
Hyperspectral imaging captures spatial and spectral information without contact, helping assess deterioration and identify pigments used on the figures. These measurements can also support restoration analysis and simulation.
Crop Classification and Identification
Hyperspectral data resolve subtle differences among crops over narrow wavelength intervals, supporting detailed classification and information extraction. Common methods include Spectral Angle Mapper (SAM) and hierarchical decision-tree classification.
Xiong Zhen of the Institute of Remote Sensing Applications, Chinese Academy of Sciences, used hyperspectral imagery to monitor rice growth in Changzhou. A hybrid decision-tree method classified six rice varieties with a reported accuracy of 94.9%. Zhang Bing applied an expert decision classification method using optimized spectral features to fine crop classification in Minamimaki, Japan. The results showed better pixel-level classification accuracy and substantially less misclassification noise.
Grain Inspection
China is a major grain producer, with crops including rice, wheat, corn, and peanuts. Hyperspectral imaging supports rapid rice inspection, variety identification, and quality assessment.
Principal component analysis (PCA) can reduce the image data and extract features such as chalkiness and grain shape. Grain-recognition models based on PCA and back-propagation neural networks (BPNN) achieved reported accuracies of 89.18% and 89.91%, respectively. Combining BPNN with data fusion increased accuracy to 94.45%, demonstrating the potential for rice variety identification and quality analysis.
Forest Species Identification
Species-specific spectral information supports forest-type delineation, stand mapping, and forest-resource inventories. Research also focuses on vegetation identification and mapping in wetlands, including rivers and lakes, salt marshes, and coastal flats, often at the community scale.
Ground surveys provide reference spectral signatures. Field hyperspectral data and spectral models can then help identify dominant species, forest types, and individual tree species. Vegetation mapping and change-monitoring studies have shown good agreement with ground observations. Hybrid and expert decision-tree methods also support crop classification, grassland biomass estimation, and retrieval of crop properties.
Explore the portable hyperspectral imaging system range for related instruments.
