P2.51 Thursday, Jan. 5 3D Image Correlation based reconstruction of fluid locomotor surfaces GUO, M*; HEDRICK, TL; Univ. of North Carolina at Chapel Hill; Univ. of North Carolina at Chapel Hill firstname.lastname@example.org
The wings and fins of a wide variety of swimming and flying animals are characterized by large, time varying deformations. These deformations arise from a variety of sources, including inertial effects, aeroelastic or fluid-structure interactions, or even active deformation controlled by the animal. Characterization of these deformations is vital to ongoing work on the fluid dynamics of animal locomotion, but presents a substantial challenge due to the large deformations, highly three dimensional nature of many animal movements, and difficulty of perfectly positioning a moving animal within experimental apparatus. Here we address these difficulties by adapting stereo digital image correlation techniques to rapidly extract time-deforming wing meshes from hovering hawkmoths. Our approach uses two closely positioned and stereo-calibrated high-speed cameras. The close alignment and resulting similar images from the two cameras facilitate automatic detection of dense matching feature sets among the two cameras. These feature pairs are then reconstructed in 3D using the camera calibration relationships and the resulting 3D point cloud smoothed and resampled over a regular grid with a thin plate spline function. We then trim the grid to overall wing shape outlines determined from contour based image segmentation with manual correction. The resulting wing grid points may then be meshed using a variety of surface meshing algorithms. Local velocity vectors are determined by an analogous operation from frame to frame on calibrated cameras. Additional camera pairs may be added as needed to measure deformations along complete locomotor cycles. We demonstrate operation of this technique on hawkmoth wings and plan to make the software implementation freely available for application to similar problems.