Agricultural Worker Tracking and Gesture Recognition System for Remote Assistance

Abstract

We have developed a worker tracking system and a lifting-and-turning gesture recognition system using a single PTZ camera, our original working uniform marked with four different red-blue combinative colors, the twin particle filters, and the Hidden Markov Model. In our experiments, we demonstrated that the proposed system robustly detected the postures and the symbol sequences in the lifting-and-turning gestures of a single worker observing from the three different directions of the front side, the left side, and the right side in real time and achieved a recognition rate of 92.5% in 120 validation data. The experimental results indicate that our developed system could be used as an agricultural worker tracking and a gesture recognition system for the remote assistance.