| A multi-aperture optical sensor, known as a fly eye sensor, has been developed at the Wyoming Image and Signal Processing Research (WISPR) Laboratory based on the visual system of the common housefly Musca domestica. This biomimetic sensor shows promising edge detection capability, in varying contrast scenarios, with minimal processing overhead. In addition, this sensor provides a cheap, small size, and light weight package compared to the traditional motion detection systems. Thus, use of this sensor for fast motion detection, and object tracking is appealing, but optimizing the use of such a sensor requires detailed study.;In this research effort, a feasibility study into using the sensor for detailed target tracking in a specific application environment has been performed. In particular, the application is the precise measurement of wing deflection in a fixed wing aircraft under test using the sensor. Tracking of a moving target in real-time is the long term goal of this project. To do so, a simulation program has been developed first, which simulates the sensor output, for a moving target on the object plane at a specified distance from the sensor. This simulation program provides a flexible environment to study the application specific behavior of the sensor before installing expensive hardware setup. Several case studies have been performed using the simulation approach, and the results are confirmed from a real sensor data. The results show that the sensor-target-background distance, and sensor-target orientation can affect the tracking system. Other variables, such as non-uniformity of the target's speed can potentially degrade the performance of the system. Therefore, several characteristics of the sensor have been generalized to facilitate the scalability of using the sensor for different applications. Finally, as an effort to overcome these pitfalls, an estimator has been developed, which is expected to provide a simple, computationally inexpensive, and real-time target tracking solution. |
