An improved optical flow estimation method and its application to vehicle following /
| Main Author: | |
|---|---|
| Other Authors: | , , |
| Format: | Thesis Book |
| Language: | English |
| Published: |
1993.
|
| Subjects: | |
| Online Access: | ProQuest, Abstract Link to OAKTrust copy |
| Abstract: | The velocity field generated by motion in a scene is referred to as optical flow. Most methods used to estimate optical flow assume that the intensity variation in an image is caused by only the motion of objects in the scene. No consideration is given to intensity variations produced by light and background sources. Such noise sources limit the accuracy of the current optical flow estimation methods, in particular for images captured in outdoor realistic environments. In this dissertation, a method for the accurate estimation of optical flow is developed which copes with environmental noise sources. This method is an improved version of the correlation-based method. It relies on an adaptive response function which alters according to the local intensity distribution within a search window. Such an adaptive approach allows various environmental noise sources to be tolerated. This method is then used as part of a tracking algorithm to obtain the range and heading angle of a lead vehicle followed by an autonomous vehicle. In this research, no control of the following vehicle was demonstrated using the caculated parameters. Various examples are employed to illustrate the performance of this method for the vehicle-tracking application. |
|---|---|
| Item Description: | Vita. "Major subject: Electrical Engineering." |
| Physical Description: | xi, 94 leaves : illustrations ; 28 cm |
| Bibliography: | Includes bibliographical references. |