When appearance is not enough, composition decides.

Conventional machine vision inspects what something looks like. Hyperspectral imaging reveals what something is made of - its chemical composition, moisture content, and material identity. The Teledyne DALSA Kaleido brings that capability to high-speed industrial sorting lines, making it a practical line scan camera for automated inspection at production scale. 

Throughput Without Compromise

A 1280-pixel spatial resolution at 2300 Hz line rate across the full spectral range removes the trade-off between spatial detail and acquisition speed that limits most hyperspectral platforms. High-throughput sorting lines no longer require a sacrifice in classification granularity.

Spectral Granularity for Reliable Classification

At 3,5 nm spectral resolution with up to 220 selectable bands, Kaleido provides the spectral detail required for reliable material differentiation. Band selection reduces data volume to only the wavebands relevant to the process, keeping data pipelines and compute requirements manageable.

Integrated Design, Simplified Deployment

Sensor and spectrograph are integrated into a single unit with built-in aberration correction and consistent factory spectral calibration between units. Multi-camera deployments become significantly more straightforward, reducing the setup effort that has historically made hyperspectral systems difficult to scale.

Accurate Co-Registration Across the Full Field

Minimal keystone and smile distortion means that spatial and spectral data remain accurately co-registered across the full field of view. Reliable co-registration is a prerequisite for classification in precision sorting and quality inspection applications where spatial position and spectral identity must be linked accurately.

Open Lens Interface

System builders are free to select optics suited to their working distance and field of view, without being constrained to a proprietary optical assembly. This flexibility is particularly valuable in OEM and custom system designs where physical installation geometry varies across projects.

Efficient Data Delivery

Kaleido uses a 10 GigE Vision interface fully compliant with GenICam® standards, delivering data in the ENVI hyperspectral file format. Programmable spectral band selection and spatial zone selection mean only the data relevant to the process is transmitted, reducing system load and keeping compute and storage requirements manageable. Frame and multi-frame triggers, meta-data support, and partial image modes (ROI, multi-ROI, binning) provide the acquisition flexibility needed for demanding production environments.

Food Inspection and Grading

SWIR hyperspectral imaging detects water absorption signatures at 970 nm, 1,200 nm, and 1,400 nm, enabling:

• Moisture level measurement and classification
• Sugar content and soluble solids analysis
• Foreign material detection where contaminants are visually indistinguishable from product

Recycling and Plastic Sorting

Up to 90% of plastics can be differentiated using SWIR wavebands to 1,700 nm, with further discrimination available beyond 1,900 nm. Kaleido enables identification and sorting of:

• PET, PE, PS, PA, ABS, PP, PC, and PVC plastic types
• Mixed polymer streams in automated inspection lines

Pharmaceutical Inspection

Pharmaceutical tablets with different dosages can be visually identical. Hyperspectral imaging reveals chemical composition differences, enabling dosage verification and contamination detection that visual or RGB-based inspection cannot reliably achieve.

Wood and Natural Materials

Moisture and chlorophyll content in wood and plant material are detectable from spectral signatures, supporting grading, drying process control, and quality classification in timber and agricultural processing lines.

Textiles

Fibre composition and dye homogeneity are measurable through SWIR spectral signatures, enabling quality grading and contamination detection in textile manufacturing and recycling.

Our Expertise is your benefit

Hyperspectral imaging systems require more than a camera. Illumination selection, spectral calibration, data processing pipelines, and application-specific band configuration all influence classification reliability. STEMMER IMAGING's application and engineering expertise supports the selection and implementation of Kaleido-based systems, from initial feasibility assessment through to deployment.

• Feasibility testing and application assessment through STEMMER IMAGING Engineering Services
• Support for illumination selection suited to SWIR hyperspectral acquisition
• Component configuration and system design advisory
• Hyperspectral data acquisition using Common Vision Blox with the CVB Spectral Library, including code, documentation, and sample programmes for acquiring and saving data as ENVI hyperspectral cubes

What is the difference between hyperspectral and multispectral imaging?

Multispectral systems capture a limited number of discrete, widely-spaced bands, typically fewer than 20. Hyperspectral imaging captures hundreds of contiguous, narrow bands, producing a full spectral curve for every pixel. This spectral density enables reliable material identification and subtle differentiation that multispectral systems cannot replicate.

Why is SWIR used for material discrimination rather than visible or NIR?

SWIR wavelengths (roughly 900-2,500 nm) correspond to the characteristic absorption and reflectance features of organic bonds, water molecules, and polymer structures. Many materials that appear identical in visible light have distinct, measurable SWIR spectral signatures, making SWIR the preferred spectral range for composition-based classification.

How does Kaleido’s integrated design reduce deployment complexity?

Conventional hyperspectral systems require separate sensor and spectrograph components that must be individually aligned and calibrated. Kaleido’s integrated design with built-in aberration correction and factory spectral calibration eliminates this alignment step and ensures consistent spectral response between units, which is particularly relevant when deploying multiple cameras across a single line.

Is Kaleido compatible with Common Vision Blox (CVB)?

Yes. Hyperspectral data from Kaleido can be acquired using Common Vision Blox with the CVB Spectral Library. STEMMER IMAGING provides code, documentation, and sample programmes for acquisition and for saving data directly as ENVI hyperspectral cubes. Kaleido is nativly compatible with the ENVI format. Contact our technical team for access to the CVB Spectral Library resources.

Can STEMMER IMAGING support an application feasibility assessment for Kaleido?

Yes. Through our Engineering Services, we offer feasibility testing and technical consultation to evaluate whether Kaleido is suited to your inspection requirements. This covers spectral band selection, illumination considerations, data volume, and classification pipeline design. Contact your local STEMMER IMAGING team to discuss your application.

What illumination is required for SWIR hyperspectral imaging?

SWIR hyperspectral imaging requires illumination with sufficient output in the 900–2,500 nm range, with selection depending on line speed, working distance, and the spectral bands of interest. For applications where spatial and spectral uniformity across the illuminated line is critical, broadband SWIR LED line lights such as the CCS LNSP are well suited - offering longer service life and lower radiant heat compared to halogen sources. Our engineering team can support source selection for your application