Object Tracking through Image Sequences
Song Wang, Hongkai Yu, Youjie Zhou, Jeff P. Simmons, Craig Przybyla · 2019
This chapter aims to study the use of tracking for the micro-structures from the material image sequence. It introduces the use of object tracking, particularly the recursive Rudolph Emil Kalman-filter tracking algorithm, to extract the three dimensional micro-structures from cross-sectioned material image sequences. The chapter focuses on addressing several specific issues in large-scale fiber tracking. Objective evaluation of the proposed tracking is a challenging issue given the large number of the objects and the difficulty to construct all the ground truth. The Kalman filter is one of the most widely used tracking models. It is a probabilistic model that recursively updates the tracking target/object’s state through the image sequence. The first step in using the Kalman filter for tracking is to choose appropriate model parameters, including the definition of the tracking states, the transition matrix, the observation matrix, and both the transition and the observation noise model parameters.