Date of Award
Master of Science
Eddie L Jacobs
This thesis presents implementation of data acquisition and object classification algorithms on a low-resource microcontroller for real-time, broad-scale object classification using a low-cost sparse detector imaging sensor. The sensor is designed to detect and classify objects into the broad categories of human, vehicle, or animal, making note of objects of high interest. This paper encompasses software for implementation onboard a low-resource microcontroller platform to acquire, process, and classify crude images of subjects for classification purposes. This paper also encompasses improvements made to a prototype hardware system to form a custom sensor array from commercially available, off-the-shelf hardware components. The sensor is designed for deployment scenearios to monitor vast geographic areas where broad monitoring is required with low false-alarm rates generated by objects of less interest.
Dissertation or thesis originally submitted to the local University of Memphis Electronic Theses & dissertation (ETD) Repository.
Reynolds, Robert Kenneth Jr., "Real-Time Object Classification Using a Custom Sparse Array Profile Sensor on an Embedded Microcontroller" (2011). Electronic Theses and Dissertations. 173.