Image-Based Mobile Robot Guidance System by Using Artificial Ceiling Landmarks  

作　　者：Ching-Long Shih Yu-Te Ku Ching-Long Shih;Yu-Te Ku(Department of Electrical Engineering, National Taiwan University of Science and Technology, Taiwan)

机构地区：[1]Department of Electrical Engineering, National Taiwan University of Science and Technology, Taiwan

出　　处：《Journal of Computer and Communications》2016年第11期1-14,共14页电脑和通信（英文）

摘　　要：This paper presents an image-based mobile robot guidance system in an indoor space with installed artificial ceiling landmarks. The overall system, including an omni-directional mobile robot motion control, landmark image processing and image recognition, is implemented on a single FPGA chip with one CMOS image sensor. The proposed feature representation of the artificial ceiling landmarks is invariant with respect to rotation and translation. One unique feature of the proposed ceiling landmark recognition system is that the feature points of landmarks are determined by topological information from both the foreground and background. To enhance recognition accuracy, landmark classification is performed after the mobile robot is moved to a position such that the ceiling landmark is located in the upright- top corner position of the robot’s camera image. The accuracy of the proposed artificial ceiling landmark recognition system using the nearest neighbor classification is 100% in our experiments.This paper presents an image-based mobile robot guidance system in an indoor space with installed artificial ceiling landmarks. The overall system, including an omni-directional mobile robot motion control, landmark image processing and image recognition, is implemented on a single FPGA chip with one CMOS image sensor. The proposed feature representation of the artificial ceiling landmarks is invariant with respect to rotation and translation. One unique feature of the proposed ceiling landmark recognition system is that the feature points of landmarks are determined by topological information from both the foreground and background. To enhance recognition accuracy, landmark classification is performed after the mobile robot is moved to a position such that the ceiling landmark is located in the upright- top corner position of the robot’s camera image. The accuracy of the proposed artificial ceiling landmark recognition system using the nearest neighbor classification is 100% in our experiments.

关 键 词：Mobile Robot Guidance System Artificial Ceiling Landmark Image Processing NCC (Nearest Neighbor Classification) FPGA 

分 类 号：TP2[自动化与计算机技术—检测技术与自动化装置]