Completeness check and type recognition with VMT® IS
The core of the system is a neuron network that can be trained to recognize characteristics with the aid of models. This network is trained to recognized model characteristics and symbols, enabling the system to read a liberal number of characteristics.
By adding further appearance variants, the system achieves maximum recognition capability possible. Fluctuations in the surrounding conditions and varying image backgrounds may therefore be easily optimized. The system is an equally effective tool for the end user and the OEM customer in terms of optimized production procedures, process controlling and documenting, thus reducing need for additional and follow-up work.
The unit is operated with a modern user interface that allows intuitive working. No knowledge of programming at all is required to operate the unit.
Through simple movement of the mouse, the user may call up new models and test tasks, change testing plans, or follow trained recognition. Since operation is being kept so simple, one day of training is usually sufficient to be able to operate the system.
Integrated into an automatic sequence VMT IS fulfills its task reliably. In case of irregularities, it is possible, with the aid of statistics and service tools, to analyze the source of the problem and remove the cause.
- Position, type and completeness check, color verification and processing control
- Type differentiation through combination of several recognized characteristics
- Type recognition with subsequent type-specific inspection
- Suitable also for difficult application conditions, such as changing backgrounds and object properties
- Trainable for an unlimited range of characteristics or symbols
- Automatic image memorizing, therefore short operation startup and time optimizing, as well as error documenting
- Accurate verification of the position of characteristics in relation to the current object position
- Suitable also for quickly moving objects and high tact rate
- Highest recognition certainty possible through use of preset knowledge of character positions or expected characters, e.g., through use of negative examples and blanking out of irrelevant or disturbing areas
- Owing to swivel/tilting-head cameras, details maybe recognized even on longish test objects and in the largest variety of places.