Depth Resolution of 3D Imaging Techniques for Target Detection in Mobile Work Machines

3D imaging is important for enabling autonomous operation of smart mobile work machines. Different sensing techniques have different characteristics, which affect the choice of sensors for each application. We study the depth resolution of three different active imaging technologies and stereoscopic depth estimation with multiple different baselines. We test the effect of depth measurement abilities in measuring average depth, angles between planar structures and diameter from logs. Active sensors show their robustness, while stereoscopic depth estimation follows the error behavior expected from the theory. However, in practice we find that the effect of poor depth resolution can be significantly reduced by averaging over multiple measurements from different points and using more sophisticated stereo processing.