利用有視覺自動車作室內安全巡邏與危險情況之偵測

Abstract

本研究主要是提出一套基於電腦視覺技術，讓自動車航行在室內環境中具有偵測危險情況與監控繪畫安全之能力。我們利用一台小型自動車作為實驗平台，並且利用無線操控的方式讓自動車航行在室內的環境中。我們提出了基於房屋中天花板角落的自動車定位方式，根據天花板角落的線條形成情況我們可以計算自動車在地圖上的實際位置。我們持續觀察環境影像並統計環境中的顏色特徵與連續影像的顏色變化，來判斷目前的環境情況是否安全。依據環境中的顏色統計資訊與變化情況，我們可以偵測出火災發生或環境停電。我們利用觀察影像中的限定區域，判斷是否遇到障礙物，以順利避開障礙物繼續巡邏。對於牆壁上的掛畫，我們提出一套將歪斜影像修復的方法來得到正確的掛畫影像，並基於此修正過的掛畫影像進行保全掛畫之安全巡邏。最後我們以成功的定位和偵測環境實驗結果證明本系統的完整性與可行性。

A vision-based vehicle system for security patrolling and danger condition monitoring in indoor environments using an autonomous vehicle is proposed. A small vehicle with wireless control and image grabbing capabilities is used as a test bed. A camera with panning, tilting, and zooming capabilities is used as the eye of the small vehicle. A vehicle location estimation method using house corners is proposed first. This is a vision-based estimation method to prevent the mechanic error accumulation in navigation. Next, several danger condition detection methods are proposed. We propose a method to change the path for obstacle avoidance in navigation using a map of nodes. To detect fire and lighting failure conditions, we monitor the image continuously and analyze the color feature in the HSI color model. We have also proposed a method for security monitoring of paintings on walls, including techniques of painting image rectification and matching. Good experimental results show flexibility and feasibility of the proposed methods for the application of indoor security patrolling.