Aerorobot/robot vision-based navigation

Vision-base technique is an effective approach for aerobot/robot navigation in the close region. The aim of this thesis is to identify the position of a series of reference target autonomously from the image for close-range navigation. This project reviews some current work about videogrammetry in Cranfield University, develops an image processing algorithm to calculate the position of each label in the image. The main feature of this algorithm is to merge the fragments into clusters according their relevant property, and calculate the position of each unique label for robot/aerobot navigation. The experiments are performed on a single SONY digital camera and typical PC,to evaluate the feasibility and reliability of this method. By comparing the images in three conditions (ideal image draw by Windows Paint, real image in experiment and image with pepper noise), the results illustrate the performance and limitation of this algorithm. Finally, it also provides some ideas to improve this navigation technique.Validerat; 20101217 (root