Gesture Recognition Based on LBP and WLD Fusion Features for Mobile Robot Control Systems
| Autor: | Chien-Han Hung, 洪健涵 |
|---|---|
| Rok vydání: | 2016 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 104 This thesis researches and implements the hand recognition based navigation mobile robot systems using color and depth cameras. Embedded devices recently play a more important role in some real-life applications, such as the entertainment, health care, and industrial applications, in particular, various types of mobile robot systems is booming faster. Furthermore, Human Robot Interaction (HRI) has drawn the great improvement in not only academic but also industrial applications. The embedded devices and all kinds of digital image processing methods have gained a good and rapid growth, resulting in this field in every stage to get a better play. In the navigation control, we create the present system in five different command through three different hand gesture and transfer these commands to ARM embedded devices for controlling the mobile robot's action. The hand detection based human’s face is implemented from skin segmentation with the RGB-H-CrCb color model. The deviation angle between left and right hands is choosing to control the rotation of robot. After the hand region has been fetched, the proposed method uses the Haar Wavelet Transform (HWT) to get the low-frequency part and high-frequency part, which contains most of the information and the texture information respectively. Then we combine it with the Weber Local Description (WLD) and Local Binary Pattern (LBP) to obtain the desired gesture features. In addition, the hand gesture classifier is Support Vector Machine (SVM) to control the linear motion. Implementation system also provides a more user-friendly graphical interface to improve the efficiency of the use of this device. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|