The goal of the VIDIMUS project was to demonstrate the feasibility of a Vision System Development Environment (VSDE) which would act as an "intelligent assistant" to a non-vision expert when prototyping vision systems for industrial applications (eg IG inspection or 3-D component assembly).
The goal of the project was to demonstrate the feasibility of a vision system development environment (VSDE) which will act as an intelligent assistant to a nonvision expert when prototyping vision systems for industrial applications.
The main features that the VSDE will provide are:
a query system for eliciting relevant application information;
automatic recommendation of image interpretation systems;
facilities for the developer experiment with algorithm chaining in an interactive environment;
a library of algorithm performance;
an environment for running the system developed;
a language in which image interpretation systems can be expressed prior to algorithm instantiation.
A first version of the interactive computing enivronment (ICE) has been completed and as part of the validation of the ICE, an exisiting nondestructive testing (NDT) has been redesigned with the ICE. A Questionnaire has been circulated to elicit development methods followed in past applications; the results are being complied for introduction in the VSDE. A definition of the Vision language has been issued and a first version of the algorithms performance library (APL) has been developed. The APL is being enhanced with a wider range of algorithms. A survey of matching algorithms has been performed and 4 mathcing techniques selected for the applications at hand. Also, a prototype task definition tool (TDT) has been developed for the specification of 3 dimensional measurement tasks based on computer aided design on (CAD) data. A mock-up of the VSDE has been demonstrated, showing the functionality of the system in the context of 2 applications, namely 3 dimensional measurement verfication and X-ray inspection of alloy wheels and a 3 dimensional inspection cell has been developed for the verification of measurements on machined workpieces.
The main features of such a VSDE include:
- A query system for eliciting relevant application information and then guiding the user through a reduction of his inspection problem to a set of well-defined image-mappable and hardware-based tasks( Functional Representation Query System - FRQS).
- Automatic recommendation of image interpretation systems using knowledge bases on both previous inspection systems as well as on algorithm performance (Intelligent Application Mapper - IAM).
- Facilities for the developer to experiment with algorithm chaining in an interactive environment (Interactive Computing Environment - ICE)
- A library of algorithm performance (Algorithm Performance Library - APL)
- An environment for running the system developed (the Image Interpretation Engine - IIE).
- A language in which image interpretation systems can be expressed prior to algorithm instantiation (Vision Language - VL).
20009 San Sebastian
5621 BA Eindhoven
G1 IXQ Glasgow