Object Detection and Distance Estimation of a Mobile Robot with Stereo Vision

Author Name: Susmita Guin, Dr.R.Hablani, Rahul Gupta
Author Email: susmita.guin198@gmail.com

Abstract

In day to day life, the studies on robotics are increasing very rapidly. Especially, for obstacle detection and estimating distance in dynamic and static environments. Therefore, researchers contribute more of their interest towards robotics as well as artificial intelligence. Hence, this paper mainly focused on detecting object and how we are estimating distance of a mobile robot. There are many solutions already given by the numerous researchers such as laser sensor, sonar sensor or camera sensor for estimating distance process. The main goal of our proposed approach is to fuse information gained from the stereo vision, infrared and ultrasonic sensors to enhance the accuracy. Thus, we proposed a system with a hybrid approach system.

Keywords

Obstacle Detection, Artificial Intelligence, Mobile Robot, Stereo Vision, Hybrid Approach.

Introduction

Being a humans, we see the 3D structure of the world around us. There are numerous studies encountered in this subject. Mobile robot is widely used in these kinds of fields. For any mobile robot, navigation and accurate information of its position is very important research area. The function of a mobile robot is commanded by the program which is running on the mobile robot (‘job’), the area (surrounding) and various sensors available on the mobile robot (hardware). These three essential elements are as displayed in Figure 1.

Robots are equipped with many sensors. Such as laser, sonar, radar, infrared, ultra-sonic and stereo camera etc. These sensors have their own benefits and features. In recent years, some researchers provide keen interest towards stereo vison. The Greek word “stereos” which means solid or firm, with stereo vision the object are seen solid in 3D range. The stereo vision contains two different systems that are: active system and passive system. The active stereo system provides depth directly by using ultra sound and laser whereas passive stereo system includes algorithms such as SAD (sum of absolute difference) algorithm and RANSAC (random sample consensus) algorithm used for images stereo system to estimate the accurate depth.

Conclusion

In this study, we have enhance the object detection and estimation of distance operation by fusing stereo-vision sensor, ultra-sonic sensor and infrared sensor. This fusion raise real-time mapping and accuracy. The proposed system that is hybrid approach system, acquire for mobile robot applications provide us with highly accurate results. Here, the open source software tools are utilized with intelligent hardware tools system. The speed, computational cost and efficiency were considered as the major design target. The main desire was to develop and design an intelligent visual supervision systems to help the human operators to detect unusual events and respond rapidly to them. Therefore, it can be employed in any kind of mobile robot applications, just as automated transportation systems, material handling, manufacturing, housework, search-and-rescue operations, or security patrolling. The proposed approach system has the succeeding attributes:

  • 3D vision.
  • Accurate point clouds.
  • Works under any unfamiliar light conditions.
  • Any of the sensor or stereo camera fails then other two sensor will continue to operate.

117 total views, 1 views today

Download File

About the author: admin