以視覺訊息驅動為基礎之機器人行為分析系統

A robot behavior analysis system base on the vision information

嚴宏成(Hong-Cheng Yan)

指導教授 ： 謝君偉

元智大學/電機通訊學院/電機工程學系/碩士(2009年)

摘要

本論文提出一套以視覺訊息驅動為基礎之機器人行為分析系統，包含一套影像處理、分析辨識軟體和一套智慧型馬達組成的小型雙足機器人，本系統先利用影像分析辨識軟體來分析與估算輸入影像內人體形態的特徵和變動情況，再透過通訊介面將辨識結果傳輸給機器人控制器，由機器人控制器依辨識結果取出預先編輯好的對應姿態資料，再操作機器人做出模仿人類行為動作以為回應，完成視覺訊息驅動與機器人互動; 有別於其他視覺訊息常用手指、手掌或臉部等區域影像,本系統視覺訊息包含人體行為動作全身影像, 且不需要於人體任何部位加作標記為輔助即可辨識出特定人類行為動作。由實驗結果得知，我們的系統確具有以視覺訊息來驅動機器人行為的分析系統功能，並且有良好的表現。

關鍵字

機器人,視覺訊息

並列摘要

ABSTRACT

This thesis proposes that one set is driven for the robot behavior analysis system base on the vision information, include a set of images process software and one robot that make up by a set of artificial intelligent motors ,this system utilizes the image to analyses and estimates characteristics and change situations of the human form first , and then distinguish the result to transmit to the robot controller through the communication interface, the robot controller according to recognize result take out the corresponding posture data that edited in advance and then operate the robot to imitate human behavior movements to reacted , finish vision information to drive and interact with robot procedure; different from some other vision information system focus on the regional images , such as finger , the palm or the face ,etc., our systematic vision information include human behavior movement whole body image ,and don’t need add any special mark on body for auxiliary analyses and recognize Experimental results prove that the proposed system is a robust and accurate for human behavior analysis base on the vision information.

並列關鍵字

Delaunay triangulation Mesh,Centroid context,Skeletal feature

參考文獻

[5]黃鈴凱 ，以手勢辨識進行人類與機器人之間的非言語互動，

[1]劉士彰，具視覺能力之鞦韆機器人，

[2]周立柏，具模仿能力之智慧型雙足機器人之設計與實現，

碩士論文，長庚大學　醫療機電工程研究所，桃園，民國95年。

[8] R. Cucchiara, C. Grana, A. Prati, and R. Vezzani, “Probabilities posture classification for human-behavior analysis,” IEEE Trans. Syst., Man, Cybern. A, vol. 35, no. 1, pp. 42–54, Jan. 2005.