Perception MODEL

The Perception MODEL Plugin allows chemometric modelling and to develop Chemical Color Imaging applications.

				It further allows to configure a Perception Core e.g. for stand-alone operation together with an interfaced hyperspectral camera. Especially in combination with a "Perception System", Perception MODEL enables the realization of so far unthinkable industrial real-time solutions, even for non-experts of spectroscopy and chemometrics.
	EUR 500,00			Perception MODEL offers various "Perception Methods", well suited to overcome most situations in industrial hyperspectral imaging

Perception Methods


Preview	Extract	Correlate	Constrain


Design	Coming soon	Coming soon	Coming soon

PREVIEW

The preview method provides an unsupervised access to hyperspectral data to get an impression about spectroscopic differences as well as about the quality of the measurement, measurement setup and applicability of pre-processing methods. The Preview Method has strong advantages concerning the selection of proper spectra.

The result of the preview method is a configuration data set, applicable for stand-alone operation together with a Perception Core connected to a hyperspectral camera.

KEY FEATURES:

Unsupervised access to hyperspectral data
Check quality of the measurement
Check the applicability of chosen of pre-processing
Check quality of measurement setup
Different Colors represent different spectroscopic information
Enables to select proper spectra. Highly understandable output format
Real-time capable
Utilizable by image processing

The preview method intuitively provides an unsupervised access to hyperspectral data. Our unique preview method is the base of working with hyperspectral data in the Perception Studio. Different colors in the preview are occurred by different spectroscopic information. Consequently the user can quickly see if there is differing spectroscopic information in the measurement. To get more detailed information the user can easily switch between various pre-processing methods like derivatives and normalizations.

The preview method gives direct feedback about the quality of the measurement too. Also reflections and morphologic information can be seen in the preview. Therefore the quality of the measurement setup can be evaluated.

Due that disturbing influences are also visualized the preview is used as basis for the selection process of spectra and spectra sets. This assures that the user is able to consider influences like reflections, over steering or shadows.

Extract

The Extract method is an unsupervised, method to extract principle components out from a hyperspectral data cube. From a gallery, showing the score-images of resulting components, a linking to the preferred color channels can be configured. Resulting in a perception based on principle components of employed data.

The result of the extract method is a configuration data set, applicable for stand-alone operation together with a Perception core connected to a hyperspectral camera.

KEY FEATURES:

Unsupervised access to principal components in hyperspectral data
Highly understandable output format
Real-time capable
Utilizable by image processing

Correlate

The Correlate method is a supervised way for correlating pure substances (constituents) to hyperspectral measurements. A perception is composed by selecting from a gallery of resulting score-images. This method is often used to correlate up to three substances in one color image and to classify the composition of a material by image processing. But also various monochrome streams representing the correlation of various substances can be processed in parallel.

The result of the “correlate” method is a configuration data set, applicable for stand-alone operation together with a Perception Core connected to a hyperspectral camera.

KEY FEATURES:

Correlate pure substances to hyperspectral data
Highly understandable output format
Real-time capable
Utilizable by image processing

Constrain

The Constrain method is a supervised method that allows generating a perception by constraining color information to hyperspectral data.

As a consequence, the user incorporates a priori knowledge to the measured data within the configuration process. This can result to well-suited perceptions, even for strong disturbed measurement data or for material with a changing chemical nature (like it occurs during the ripening e.g. of fruits).

The result of the Constrain Method is a configuration data set, applicable for stand-alone operation together with a Perception Core connected to a hyperspectral camera.

KEY FEATURES:

Highly understandable output format
Real-time capable
Utilizable by image processing

Design

The Design method is intended to give the user the ability to design a feature extraction configuration manually. In the current version the user can choose various filter curves being convolved with the spectral information.

KEY FEATURES: