VTT Electronics is developing a multi-channel machine vision system, which is able to generate information for grading dried fir boards. Grading is based on the detection, classification and locating of knots. The system can be modified for the grading of pine boards as well. A concise overview of the system is presented in Figure 1.
 

The heart of the system is matrix colour camera and the DSP processor card. As the board moves under the imaging system a photoelectric sensor triggers the imaging. The camera grabs images at constant intervals and pastes them together. The resulting image represents the whole board in which surface defects (knots and other defects) are marked. The next step is to classify the defects into right surface defect classes. An example of two separate images and the resulting board image is shown in Figure 2.

The inspection system consists of 450 MHz Pentium III, two lighting devices with 80 kHz fluorescent lamps, a 3CCD matrix camera and a TMS320C80 based processor board, which is used for analogous image grabbing and image analysis. Depending on the type of the 3CCD multispectral camera, the analysis can utilise greyscale, colour, NIR or polarisation information. 

At the moment the system has the following performance:
 

• image analysis can detect defects and their locations and calculate essential features in real time,
• imaging with > 0.5 mm resolution; with 1.0 mm resolution board can move up to 5 m/s,
• different classifiers included in the system can be applied to the classification of the defectous objects,
• digital multichannel images can be saved for further analysis.

Dr. Jyrki Laitinen
VTT Electronics
P.O.Box 1100
FIN-90571 Oulu
FINLAND
tel. +358 8 551 2234
fax. +358 8 551 2320
e-mail: Jyrki.Laitinen@vtt.fi