Using 3D Reconstruction Technology to Obtain Range and Motion Information of Tornado Debris Objects

Technologies of 3D reconstruction using images obtained by digital cameras have been developed and applied to meteorological observations in cloud development and tornado evolution, etc. In this paper, a 3D reconstruction technology was implemented to derive positions of objects in space. While the technique is not new, practical application and its efficacy for radar meteorological field experiment has not be quantified, which serve as the primary motivations and the key novel aspect of this work. In the Advanced Radar Research Center (ARRC) of the University of Oklahoma (OU), a relevant study to the proposed technique is to analyze tornado debris signature (TDS) of tornadoes to characterize the polarimetric signature observed by weather radar, i.e., negative differential reflectivity (ZDR) and low correlation coefficient (RhoHV), which have been commonly observed around tornadoes but not yet fully understood and characterized. One can imagine a series of stereographic photos of tornado could help provide complementary perspective to radar observations on understanding the TDS. Both simulation and experiment of object reconstruction will be introduced in this paper. For simplicity, the debris objects are represented using discrete points where their orientations are omitted at present stage. Simulated images of these objects are generated through camera projection, which is a function of Euclidean projection, location and orientation of camera. The 3D reconstructions of the objects are performed using Direct Linear Transformation (DLT) algorithm. In addition, Levenberg-Marquardt iteration algorithm is also applied to refine the measured projection matrices, which in practice would help correct position and orientation values recorded in the field experiment. Systematic errors were added into the images to reveal the effects and practical limitations of utilizing such techniques in the field, e.g., the accuracy of measuring orientation using a digital level and/or protractor and the camera position using a commercial GPS device.