ETD

• computer vision
• FPGA
• Stereo-matching
• Stereoscopy

Stereoscopy is a computer vision technique that enables the estimation of scene depth. This technique has attracted significant attention due to its possible application in autonomous driving, 3D reconstruction, and robot navigation. The fundamental principle of stereoscopy is to capture images of the same scene using two cameras, positioned similarly to human eyes. In this work, after introducing the basic geometry underlying depth estimation, we provide an overview of the taxonomy and current state of the art of stereo matching algorithms. Then, the selected algorithm for FPGA implementation is presented, followed by a detailed description of the hardware design. Finally, the results obtained are discussed and potential directions for future improvements are outlined in the conclusion.