Optical tracking has been an important subject of research since several decades. The utilization of optical tracking systems can be found in a wide range of areas, including military, medicine, industry, entertainment, etc. In this thesis a complete hardware platform that targets high-speed optical tracking applications is presented. The implemented hardware system contains three main components: a high-speed camera which is equipped with a 1.3M pixel image sensor capable of operating at 500 frames per second, a CameraLink grabber which is able to interface three cameras, and an FPGA+Dual-DSP based image processing platform. The hardware system is designed using a modular approach. The ﬂexible architecture enables to construct a scalable optical tracking system, which allows a large number of cameras to be used in the tracking environment. One of the greatest challenges in a multi-camera based optical tracking system is the huge amounts of image data that must be processed in real-time. In this thesis, the study on FPGA based high-speed image processing is performed. The FPGA implementation for a number of image processing operators is described. How to exploit diﬀerent levels of parallelisms in the algorithm to achieve high processing throughput is explained in detail. This thesis also presents a new single-pass blob analysis algorithm. With an optimized FPGA implementation, the geometrical features of a large number of blobs can be calculated in real-time. At the end of this thesis, a prototype design which integrates all the implemented hardware and software modules is demonstrated to prove the usability of the proposed optical tracking system.
|Date of thesis defense:||24 October 2012|
|Date:||30 October 2012|
|Faculties / Institutes:||The Faculty of Mathematics and Computer Science > Department of Computer Science|
|Subjects:||004 Data processing Computer science|