|
|
|
|
A Special Lecture Held by Prof. Hong Zhang, University of Alberta, Canada |
|
|
|
Published：2011-11-09 17:49 |
|
On November 9, Prof. Hong Zhang, from University of Alberta, Canada held a special lecture.
|
|
Title：「 Image Comparison and Matching for Robot Visual Navigation 」
|
|
Abstract： In this talk, I will discuss our recent research on two problems in appearance-based visual SLAM (simultaneous localization and mapping), namely, key frame detection (KFD) and loop closure detection (LCD). As a robot explores an environment and constructs its appearance map, visually distinct images or key frames are identified and included in the map. Our research investigates options for the critical step of measuring image similarity for KFD. LCD asks whether a robot has returned to a previously visited location and, in appearance-based visual SLAM, it is modeled as a problem of matching robot's current view with those images in the robot map. We specifically tackle the problem of perceptual aliasing, experienced by the predominant bag-of-words approach in state-of-the-art LCD algorithms, and present direct feature matching using hashing as an attractive and feasible alternative.
|
|
|
|
 |
|
A Special Lecture Held by Prof. Hong Zhang |
|
|
|
|
Prof. Hong Zhang Doing the Research Exchange |
|
|
|
|
Prof. Hong Zhang Doing the Research Exchange |
|