Exploitable results Automated quality control system for end-products in a textiles industry: Final prototype The final prototype was the final result of the project. This machine has most of the features exposed in the proposal. Special kind of lights have been applied with a very low power consumption, very low heat and concentrated illumination under the line scan cameras. The software can also be considered a success. In fact it implements a special neural network system which is able to recognize a defected sample by comparison with the features of a good one from the same type of fabric. This is very important because normal neural network systems are able to recognize a particular detail only if they have received the instruction in order to identify that particular feature. Training time of the network is therefore extremely reduced. The best performance of the prototype is obtained on fabrics with uniformity of light colours. The machine, at the moment is not able to work on striped fabrics. Searching for OpenAIRE data... There was an error trying to search data from OpenAIRE No results available
