Document Type
Thesis
Date of Award
10-31-1992
Degree Name
Master of Science in Mechanical Engineering - (M.S.)
Department
Mechanical and Industrial Engineering
First Advisor
Rajesh N. Dave
Second Advisor
Anthony D. Rosato
Third Advisor
Ian Sanford Fischer
Abstract
Collision parameter evaluation of rigid spherical particles requires estimation of pre- and post-impact position, velocity and angular velocity vectors of colliding particles. A three dimensional experimental technique is devised where the instantaneous position and orientation of the particles are determined by analyzing frames obtained through high speed video imaging system. Since the image obtained is two dimensional, a mirror setting is designed to capture two orthogonal views of collision in order to estimate motion in depth. The translational velocity vector is determined from the position information extracted through application of available image analysis techniques over two time frames of digitized image. The angle and the axis of rotation between two time frames is determined by tracking a few randomly marked feature points on the spheres. In order to reduce the effect of error involved in the coordinate measurement in rotation estimation, a least square fitting of the set of three dimensional feature points is employed. The technique performed satifactorily in rotation vector estimation. The method developed here is applied to a set of experimental data for computation of collision parameters. Preliminary results indicate a need for improving the accuracy of position measurement.
Recommended Citation
Sen, Sumit, "Computation of collision parameters of spheres by computer vision" (1992). Theses. 2341.
https://digitalcommons.njit.edu/theses/2341
