EXTENDING A BIOLOGICAL CELL TRACKING SYSTEM TO TRACK RANGE DATA OF PEDESTRIANS

Open Access
Author:
Poore, Ryan James
Graduate Program:
Electrical Engineering
Degree:
Master of Science
Document Type:
Master Thesis
Date of Defense:
December 08, 2016
Committee Members:
  • William Higgins, Thesis Advisor
  • Richard Tutwiler, Committee Member
  • Kultegin Aydin, Committee Member
Keywords:
  • LIDAR range data
  • topological constraints
  • geometric active contour
  • Interacting Multiple Model
Abstract:
The challenges faced when constructing a system to track biological cells coincides with the challenges when tracking pedestrians: they have complex topological shapes; wide range of behaviors; they move abruptly; they interact with other objects; their shape deforms; they leave the field of view (FOV) and enter the FOV; they split and merge. The similarities suggest that a system for tracking cells potentially works well for pedestrians. This work presents an automated tracking system that extends a framework, designed for tracking hundreds of biological cells in phase contrast microscopy, to tracking multiple human sized objects with LIDAR range data. It integrates various classic image-processing techniques with an adaptive interacting multiple model (IMM) estimator, a topologically constrained active contour, and spatiotemporal trajectory optimization. The framework processes the data with multiple independent collaborating modules. This module design facilitates a straight forward substitution of algorithms within a module without affecting the rest of the system.