Neural Network Classifier and Robotic Manipulation for an Autonomous Industrial Cork Feeder

• Autor: Salgueiro, André ; Santos, Sofia ; Pereira, Artur ; Cunha, Bernardo ; Pedrosa, Eurico ; Azevedo, José Luis ; Lau, Nuno ; Lopes, Paulo ; Gomes, Tiago
• Assuntos: Computer vision ; Convolutional neural network ; Cork ; Deep learning ; Universal Robots
• É parte de: Progress in Artificial Intelligence, p.433-444
• Descrição: This paper presents a solution for an autonomous cork puncher feeder with a robotic arm using image processing techniques and a convolutional neural network. Due to the need for cork strips to be inserted into the puncher with a specific orientation, to produce high quality cork stoppers, the identification of the orientation of each cork strip on the conveyor belt is a necessity. In response to this problem a convolutional neural network is used to analyse images processed with subtracted background, to create a robust solution for cork strips classification. In the tests carried out, a classification accuracy of 100% was obtained in a test data set with 12 different cork strips.
  
• Editor: Cham: Springer International Publishing
  
• Idioma: Inglês