Shape Recognition of Plant Equipment from 3-D Scanned Point Cloud Data Using a Convolutional Neural Network

This paper proposes a method for recognizing the shape of equipment using a convolutional neural network with point cloud data for the equipment generated using a 3-D scanner. The architecture of the convolutional neural network, the method for generation of the training data, the training method, and the hyperparameter values for recognition are described in detail. Experiments on the test data using the proposed method demonstrated a recognition accuracy of 93.8%, which is higher than the accuracy of previous recognition methods using shape descriptors.