Multivariate Machine Vision Systems

See Your Process More Clearly

Multivariate machine vision systems combine the power of multivariate analysis with the detailed information contained in high quality colour images to help you truly "see" what is going on inside your process and as product moves down the line. The extracted information enables you to make adjustments that improve product quality and reduce production costs.

Machine vision is popular because of its ability to perform fast and non-invasive low-cost analysis on products and processes. At ProSensus, we combine the power of traditional signal processing techniques like wavelet analysis with multivariate methods to extract both spatial and spectral features from colour images or hyperspectral near infrared (NIR) images. These features can either be used directly to evaluate/predict product quality, or used as a soft sensor in the scheme of statistical process control (SPC) for product quality monitoring, process fault detection and control.

Multivariate Machine Vision Gives You

  • Real-time estimation of the quality of your product and automatic adjustment so you'll reach quality targets
  • Real-time characterization of the state of your process
  • Immediate notification when a product deviates from the desired result
  • Improved process understanding—discover the true causes of product defects
  • Information on hard-to-measure but critical quality properties, such as the crunch of a chip or matching two pieces of marble based on their appearance

We've Helped Clients With

  • NIR chemical imaging of pharmaceutical tablets
  • Snack food seasoning level prediction
  • Characterizing and optimizing visual appearance, e.g. flotation froth
  • Real-time characterization of combustion processes
  • Remote sensing
  • Lumber grading
  • Seed disease classification
  • Grading steel surface defects