Demining Augmented Reality
NREC has developed an augmented reality training system for hand-held land mine detectors. It is implemented using a pair of custom-designed stereo cameras with onboard visual-inertial odometry tracking paired with a commercially available augmented reality headset.
The use of hand-held land mine detectors requires many hours of training for human operators to effectively learn the sweep motions required. It is important for operators to learn the proper detector sweep technique in order to reliably find potential land mines when operating these devices in the field.
The augmented aeality (AR) training system for hand-held land mine detectors provides an operator with direct visual feedback on the quality of their detector sweep technique.
A custom-designed stereo camera with onboard visual-inertial odometry algorithm tracks the hand-held detector motions. The swept path is then visualized through an augmented reality headset to show the traversed path and highlight gaps in coverage on the ground. Several motion parameters including speed, height, and tilt angle are measured and used to provide immediate feedback to the operator on their technique. If any parameters exceed pre-defined thresholds, the colors will show up in red to indicate the need for improvements.
The augmented reality interface may be used as a self-training system to allow operators to refresh their skills and practice while in transit to the field.
This project is funded by the US Army Research Laboratory (ARL) and Joint Improvised-Threat Defeat Organization (JIDO).