Recursive Hough Transform Tracking For Sonar Motion Estimation

Necessary and sufficient conditions for existence and uniqueness of a motion parameter solution are derived to establish the ground- work for estimating circular motion of a distributed object from matched feature sonar measurements. Conditions are derived first for an active sensor system first in and then out of the plane of the object motion assuming the plane of motion is known. The Hough transform is shown to solve the nonlinear motion equations of a feature match and a recursive tracking algorithm based on the Hough transform with 2-D kernel voting for noise reduction is proposed.