Low-Cost, High-Fidelity Virtual Landmine Detection Training System

The development of a low-cost motion capture based landmine detection training system is presented in this paper. Two Nintendo Wii Remotes (Wiimotes) are used to form a stereo vision system to obtain the position and orientation of the simulated physical landmine detector. The position data is used to compute the distance between the detector and a virtual mine, and the distance is used to generate a synthetic sound based on a sound model. Metal detector sounds have been recorded for several types of landmines as well as for clutter. From the recorded sound data, an artificial neural network is implemented for sound modeling. During the training session, a trainee is asked to identify the landmine’s type and buried location. The virtual training system can help train a trainee’s ability to discriminate among different landmines and clutter, and help refine his/her sweeping skill. This system also provides additional visual cues based on a color scheme to improve the training effectiveness.