Imaging configurations and partial occlusion mapping for multi-camera volumetric velocimetry around bodies

Moving bodies present imaging challenges for multi-camera volumetric velocimetry techniques. In addition to measurement volume regions which are fully obstructed by the body, partially-occluded regions arise, where some viewpoints are obstructed but a reduced set of cameras still view the particle field. These partially-occluded regions may have reduced resolution and accuracy compared to the unoccluded regions of the measurement volume. To characterize these limits, this work evaluates the effects of varying camera configurations and occlusion parameters (e.g., size, shape) on particle reconstruction and velocity measurement in partially-occluded regions. Of particular interest are trade-offs between optical access, depth resolution, and the fraction of the measurement volume which is partially- or fully- occluded. Using synthetic images, comparisons are made to a ground truth reference flow field and the same particle volume without any occlusions present. Results suggest that adding even 1-2 additional cameras to a typical four camera experiment can enhance measurement capabilities in partially-occluded regions, even when the additional camera does not extend the viewing angle.