bahadori-rescuews03 - Human Body Detection in the RoboCup...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon
Human Body Detection in the RoboCup Rescue Scenario Shahram Bahadori, Luca Iocchi Dipartimento di Informatica e Sistemistica Universit`a degli Studi di Roma ”La Sapienza” Via Salaria 113, 00198 Roma Italy E-mail: <lastname>@dis.uniroma1.it Abstract — This paper presents an analysis of techniques that have been studied in the recent years for human body detection (HBD) via visual information. The focus of this work is on developing image processing routines for autonomous robots operating for detecting victims in rescue environments. The paper both discusses problems arising in human body detection from visual information and describes the methods that are more adequate to be applied in a rescue scenario. Finally, some preliminary experiments for such methods in recognizing rescue victims are reported. I. INTRODUCTION One of the main objectives of a search and rescue mission is to rescue the disaster victims. The rescue environment may be very hazardous for the rescue human agents, because of the unstable situation of the camp, therefore robotic rescue agents can be important collaborators for people involved in a rescue mission. One of the most important requirements in such rescue scenarios is to identify the victims from all the other objects. This task is very simple for a human agent, while it is not for a robotic agent. However, there are situations in which human agents cannot (safely and quickly) intervene and thus the ability of detecting human victims is important also for rescue robots that can access such dangerous areas. In order to detect a human body, an autonomous robot must be equipped with a set of specific sensors, that provide information about the presence of a person in the environment around. Each of these sensors is suitable to detect a specific characteristic of human bodies, such as shape, color, motion, IR signals, temperature, voice signals, emissions, etc. Although sensor fusion is a fundamental issue in this context and an effective im- plementation of a human body detection system should integrate information coming from several kinds of sen- sors, in this paper we focus our attention on the problem of human body detection from visual information. The main reason is that, among all the sensors that may be used for detecting human bodies, most of the research work in the literature make use of vision systems, due to the fact that cameras are commonly used in robotics and of course in computer vision and images are rich sources of information, that are usually used also for other tasks, like localization, mapping, obstacle avoidance, etc. In fact, human body detection by visual information has been studied also in different application fields, that are for example autonomous vehicle navigation for detecting the people and pedestrians for autonomous navigation [1], surveillance [2], human motion capture for virtual reality [3], etc. A large amount of research on HBD has been carried out in the past, specially in the community of Computer Vision (see [1] for a
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 2
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 08/25/2011 for the course EGN 3060c taught by Professor Sukthankar,g during the Fall '08 term at University of Central Florida.

Page1 / 6

bahadori-rescuews03 - Human Body Detection in the RoboCup...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online