大きな計測範囲のためのパンニングDLT法の開発

概要

論文の詳細を見る
The purpose of this study was to develop a DLT procedure using pan and tilt cameras (henceforth, panning DLT) for the collection of three-dimensional coordinates data in large measurement volumes. Two stationary VTR cameras and two pan and tilt VTR cameras were used to obtain images of two control volumes. Two control volumes were used to compare the accuracy of coordinates data taken from stationary DLT with that of panning DLT technique: control volume A (CV-A) was 1.25×10×5m^3, and control volume B (CV-B) was 1.25×5×3m^3 in dimension. Control points and a ball thrown in each control volume were videotaped. DLT parameters for panning technique were calculated as follows: (1) Relative coordinates of control points to reference points were calculated from the digitizer coordinates of the reference points and control points. (2) Digitizer coordinates of the reference points and target objects were digitized. (3) Virtual control points were reconstructed from the relative coordinates of the control points in each frame. (4) Virtual control points appeared in the view of each camera were selected and used for the calculation of DLT parameters. Three-dimensional coordinates of control points were computed, and RMS errors were compared between stationary and panning DLT techniques. In CV-A of large volume, RMS errors of control points in stationary and panning DLT techniques were 14.0 and 7.5mm, respectively, while in CV-B of small volume, there was no significant difference in RMS errors, 8.3 and 5.8mm, respectively. Vertical accelerations of a thrown ball computed with two techniques were 9.74 and 9.71m.s^<-2> in CV-A. These results reveal that panning DLT technique developed in this study can be adapted for the collection of three-dimensional coordinates data in large measurement volumes.
社団法人日本体育学会の論文
1997-05-10

大きな計測範囲のためのパンニングDLT法の開発

スポンサーリンク

概要

著者

関連論文

スポンサーリンク