Skip to main content

DEFECT DETECTION BY DIGITAL IMAGE PROCESSING FOR COMPUTER ASSISTED MACHINING

Objective



Research into new techniques for automatic detection of features in sawn wood, these techniques to be adaptable to machines manufactured by small to medium size companies. Transfer of technique based on processing of images and of data from several sensors. Advanced data processing, with dynamic and adaptative analysis of signals. Applications to sawing operations, and notably the ripping and cross cutting of sawn hardwoods.

Raw material and labour costs, which can each represent up to 30 % of the cost of the finished product, imply that:

there should be optimization of the sawing, especially of hardwoods, and that;
there should also be a transfer of labour employed in repetitive and particularly dangerous tasks towards more noble and safer tasks.

The expected gains are:

an increase in raw material yield from 50% to 57%, ie a saving, across the EEC, of 48 MEcus;
the possibility of including 15% of second quality sawn timber;
the transfer of 50% of the personnel exposed, both to noise and to the amputation dangers inherent in saw blades, towards safer tasks;
improved technology for small to medium sized woodworking machinery manufacturers, with a resultant increase in export volume, both in terms of volume and of added value, (especially to North America).

In order to meet the specific needs of small to medium sized companies, the real problems which they face will be analysed and modelled, so that they may be handled by automatic form recognition and image analysis processes. These processes, which are derived, notably, from X-rays, from color vision and from ultrasounds must give response times of less than 200 ms, and must allow for feedspeeds of up to 200 m per minute and cutting rates of 120 strokes per minute.

The particular case of certain defects which are difficult to detect will be studied from the standpoint of the wood's physical and chemical make up, with the aim of developing specific solutions.

The integration of these technologies into existing machines will be studied both at communication protocol level and at the level of the nature and structure of the data for optimization.

The machining processes concerned are the cross cutting and the rip sawing operations in wood manufacturies. A coupling between a sensor, or a set of sensors, and a woodworking machine requires a new definition of the mechanical and electrical environment.

The mechanical part concerns the transfer of the pieces, the integration of the cover in the line, the lighting and power equipment, the physical implementation of the sensors and their associated electronic devices. The electrical part concerns the synchronization of the signals, the handshake between computers and the multitasking sequence during the automatic operating mode. These points will be started at the last task which provides a design of a new industrial automatic inspection line.

Funding Scheme

CSC - Cost-sharing contracts

Coordinator

Centre Technique du Bois et de l'Ameublement
Address
Rue De Blenod Maidières
54700 Pont-à-mousson
France

Participants (4)

Máquinas Pinheiro
Portugal
Address
D. Pedro V
4785 Trofa
Mécanique et Engrenage Modernes
France
Address
15 Avenue Barthélémy Thimonnier
69300 Caluire-et-cuire
NATIONAL RESEARCH COUNCIL OF ITALY
Italy
Address
Via A.barazzuoli 23
50136 Firenze
Omga SpA.
Italy
Address
Via Carpi-ravarino
41010 Modena - Soliera