Improving image quality

Computational Imaging

MVTec HALCON supports computational imaging, where multiple images of the same object under varying lighting or capture conditions are combined. This improves image quality without relying on complex lighting or specialized optical setups. As a result, image features can be extracted in real time, enabling robust applications with reduced development time and lower costs.

Do you have an application in mind?

Get a free evaluation

various computational imaging techniques

Functions In HALCON

Hyperspectral imaging
High Dynamic Range (HDR) imaging
Photometric Stereo (PMS) method

Hyperspectral Imaging

Hyperspectral Imaging analyzes materials based on spectral signatures. Hyperspectral cameras provide a full spectrum per pixel, spanning dozens to hundreds of closely spaced wavelengths. This reveals differences that cannot be detected with RGB or grayscale cameras – such as distinguishing between chemically different but optically identical objects.

With MVTec HALCON, hyperspectral data can be used for classification, material identification, and quality control. HALCON offers a variety of directly applicable example programs. A particularly relevant example is the "Classify different pills based on hyperspectral images" program. It demonstrates a complete pipeline for spectral classification, enabling reliable differentiation of optically identical but chemically distinct pills.

The process includes:

Importing hyperspectral data in the widely used ENVI file format or from GigE Vision-compatible hyperspectral cameras.
Reorganizing the data structure from BIL (Band Interleaved by Line) to an image-based format.
Reflection transformation to calculate the actual intensity values. For this, in addition to the raw measurement data, dark and white reference data must also be available.

Classification in HALCON based on the emitted spectra

Hyperspectral imaging in HALCON reveals differences between optically identical but chemically distinct objects.

Hyperspectral classification in HALCON separates visually identical but chemically different pills.

Example of hyperspectral classification in HALCON: Optically identical pills are differentiated based on their spectra.

Pipeline in HALCON for hyperspectral imaging: Import, transformation, and classification.

Pipeline of hyperspectral imaging in HALCON: From data acquisition to classification.

Supported Hardware

GigE-Vision-Compatible Hyperspectral Cameras

The example is based on data captured with GigE-Vision-compatible hyperspectral cameras, such as the FX series from our technology partner Specim. Alternatively, offline stored ENVI datasets can be used.

Application Example: Pharmaceutical classification for counterfeit detection, sorting, or differentiation of generics and branded products.

Use Case

Classification Of Visually Identical Objects

A typical use case is distinguishing visually identical pharmaceuticals, such as for counterfeit detection, sorting different medications in production lines, or differentiating branded products from generics.

Although tablets may have the same shape and visible color, they can be reliably classified based on their infrared spectra – for example, by active ingredient or composition. HALCON enables the seamless integration of such processes into industrial applications.

High-Dynamic-Range Imaging (HDR)

HDR combines images with different exposures into a single image with uniform lighting, revealing details in both bright and dark areas. Typical use cases include defect detection and quality control.

Bore hole illuminated with short and long pulse.

HDR in HALCON: The hole is combined from bright and dark images into a detailed picture.

HDR combines bright and dark images into a uniformly lit picture.

Photometric Stereo Method (PMS)

PMS generates shape maps by capturing an object from multiple lighting directions. The computed shape maps make edges and textures clearly visible, such as Braille characters on pharmaceutical packaging. These maps can be further processed in machine vision tasks like OCR or object segmentation. Typical applications include detecting fine surface defects or reducing lighting effects in print inspections.

Photometric Stereo in HALCON displays an object with lighting from four directions for shape and texture analysis.

Object illuminated from 4 different directions.

PMS shape map in HALCON clearly displays Braille characters for OCR and quality control.

Photometric Stereo in HALCON generates a shape map with clearly recognizable Braille characters on a pharmaceutical packaging.

APPLICATIONS IN PRACTICE

Our Success Stories

3D vision enables robots to “see” in the carpentry industry

As in many industrial areas, automation is advancing in furniture production as well. One example of this is MAB Möbel AG from Switzerland: with the support of HOMAG Bohrsysteme GmbH and MVTec Software GmbH, this furniture manufacturer automated the…