Scorpion Vision XV – 3D Defect Classification with Machine Learning

Thor

We are proud to announce that Scorpion Vision Software version XV is out.

Three years of hard works has come to an end.
We all look forward to take advantage of the new features and benefits.

IS-2019-0034-MLwithXV_zoom3
3D heightmap where defect are classified with machine learning on this image cuts and bulges

The 3D Image is captured by the Scorpion 3D Stinger Scanner featuring a peak scan-rate of 3500 laser lines per second and a 12 bit high dynamic range image sensor.

IS-2019-0034-Scanner

In Scorpion Vision XV we use Python 3 and the machine learning library skLearn to train a defect-classification with the following defects on a 3D surface scan:

  • Cut
  • Hole
  • Dent
  • Bulge

The Scorpion 3D Stinger classification is super fast

 We provide a sample profile showing how the Scorpion tool Blob4 can provide the features to the machine learning algorithms.

IS-2019-0034-MLwithXV_zoom4
3D Classification of Cuts, Holes, Bulges, Dents or nothing marked with ‘-‘

This provides the technology for a new generation 3D applications based upon the latest machine learning technology technology.

This sample using Linear Regression and a State Vector Machine

Scorpion 3D Stinger™ is a family of machine vision components and software products. They provide simple, state of the art and cost-effective building blocks for OEM and system integrators.

Register for Scorpion and Tordivel News