The objective of this project was to develop an integrated, flexible and robust CIM vision system. It is optimised as an entire system rather than optimizing the individual system components alone. The vision system is designed for performing quality assurance inspections in an industrial concrete tile fabrication environment.
An integrated, flexible and robust computer integrated manufacture (CIM) vision system has been developed. It has been optimized as an entire system rather than optimizing the individual system components alone. The vision system is designed for performing quality assurance inspections in an industrial concrete tile fabrication environment.
Novel optical components, sensor architecture and processing electronics, incorporating recent progress in the field such as random access to the region of interest, have been constructed. The zoom lens is digitally controlled and has an appropriate tiltable mirror in front in order to move the optical axes. The computation time is reduced by an optical implementation of the Gaussian convolution with a diffractive element, ie a computer generated hologram. The latter is electrically switchable using liquid crystal technology. A novel dynamic image sensor offers random access of pixels, selectable exposure time and frame rate, as well as programmable image resolution through on-chip charge summation capabilities. Communication with the CIM environment is ensured with bidirectional lines with through the lens (TTL) levels, available on the CIVIS optocontroller card, and standard interfaces such as RS-232 and Ethernet adapters.
The integrated approach adopted involved a careful adaptation of the submodules to each other. Some of the image preprocessing steps are done in the optical domain and on the sensor chip in order to achieve optimum performance. The system achieves active vision by responding to changing situations by a laboratory test bed, simulating closely the industrial environment. Although the demonstration of the system was in the field of surface inspection of concrete tiles, its potential for robotics applications has also been investigated.