Intelligent Hand Posture Recognition System Integrated to Process Control

This paper presents the development of a hand posture-based interface applied to process control by means of OPC technology. The hand segmentation is based on skin color using HSV model. The hand posture descriptors are obtained from the combination of SIFT and PCA feature extraction techniques. The classifier algorithm is an artificial neural network, which uses MLP architecture, and different topologies have been evaluated. The hand posture recognition system is the OPC Client, which sends data to the OPC server connected to a PLC. The OPC configuration is described. Finally, results from the classifier implementation tests are presented and a control system operated by hand postures is demonstrated to validate this work.