Quality control of fascia board defects

We design and deliver modern machine vision systems based on neural networks for manufacturers and factories of furniture and wood flooring. We support the automation of quality control processes on production lines from raw material to the finished board. We offer two groups of applications that can be distinguished depending on the measurement method.

  1. Detecting defects on raw boards – algorithms based on neural networks enable the detection of defects such as resin pockets, dead or loose knots, and mechanical damage. This type of quality control can be used to eliminate defects on pin joints of slats and to automate the repair of the top surface of products.
  2. Precise 3D geometry measurements - laser triangulation method enables accurate measurement of the board surface after mechanical processing. That type of quality control can be used to detect the volume of defects, among others, which enables the automation of the skimming process.

The algorithm for detecting defects in wood products based on neural networks achieves a great efficiency of 95%, with a low number of incorrectly classified samples. Our system is based on two measurement techniques: 3D scanning, based on the laser triangulation method, and an RGB colour line scan camera.

The system detects defects such as:

  1.  cracks,
  2.  black knots,
  3.  resin pockets,
  4.  lack of geometry (insufficient planing),
  5.  bark.
Scroll