An autonomous agent designed to operate in an unknown environment must be able to determine quickly and accurately the locations of objects in that environment. Vision is the most appealing mode of sensing for this task, however existing vision-based approaches have proven inappropriate for real-world applications, largely due to the excessive computational effort required by these techniques. This thesis has presented a new technique for recovering depth information from grey-scale imagery that is suitable for real-world tasks. The Intensity Gradient Analysis (IGA) algorithm produces accurate range estimates, while using a fraction of the computational effort required by existing techniques.
KeywordsCamera Motion Image Brightness VLSI Implementation Imaging Geometry Subtraction Operation
Unable to display preview. Download preview PDF.