Research Project P2:
Motion detection and tracking in geo-referenced panoramic video

Research Project Goal

Like project P1, we use a database of geo-referenced panoramic video acquired by a car. In this project, we seek to identify moving objects, such as people and cars, which may occur in the video stream.

Research Project Scope

Given a stream of geo-referenced panoramic images, identify and track moving objects. It may be assumed that moving objects have common appearances (e.g., walking people look similar), so that hypotheses may be identified using object-specific features. By tracking these objects over time, we seek to find evidence that they move.

Data

Sample image of a geo-referenced panorama, here overlayed with a map from the USGS. Click on the image above to see a video demonstrating the panoramic data acquisition process.

Tasks

• Train feature detector to find candidates of moving objects
• Implement tracking method for moving objects through multiple panoramic images
• Tricky: Implement method for deciding whether an object moves relative to its environment.

Research Project Status

student names here

Point of Contact

Sebastian Thrun

Midterm Report

not yet submitted

Final Report

not yet submitted