Saltar apartados
  • UA
  • i3a
  • Recognition of a marker and distance using stereo vision and segmentation techniques

Recognition of a marker and distance using stereo vision and segmentation techniques

Use of color information to recognize a marker and the distance which is located, by means of stereo vision and segmentation techniques

Introduction

We define Robotics mobile as the design and construction of mobile vehicles able to navigate autonomously or under the control of an operator. The rare time navigation is performed within known and static, which are usually unknown and potentially variable environments. We must, therefore, give the vehicle of certain autonomy through the incorporation of perceptual abilities and problem solving in real time. One of the problems that is working on robots navigation is the detection of obstacles through the use of sensors. These elements allow to detect obstacles that appear in his career, but are unable to discern between them.

Since we need to have the best possible information to, to the extent possible, sail without the supervision of a human, in this project, we propose a system of stereoscopic vision, allowing you to take advantage of the depth information from a pair of stereoscopic to determine at what distance is a marker which will indicate to a mobile platform that action should be carried out. It is essential to recognize the type of marker by which involved techniques of signal detection.

In this project we intend to provide information of different types to the system so that the method is flexible and can work with images of different types. The treatment that is made of color information by which we intend to test various models of color, to determine which of them provides best results is essential. Using segmentation techniques combined with stereo correspondence is a marker to recognize and locate the distance to that found in the camera.

Since we want the robot to respond to certain signals, it is essential to incorporate techniques of segmentation and recognition in order to determine the type of detected signal as well as the same distance to the camera. In our group we have obtained results with techniques of correspondence in stereo vision both segmentation and recognition methods. We aim to join these lines of research and progress in this context.

Taking as departure we ask images taken by cameras detect a marker, recognize its type and the distance that is so that the robot can perform the action indicated the marker.

Objectives

  • Improve the initial results in the problem of the correspondence and segmentation and image recognition by introducing new models of color. The latest research suggests that models based on human perception are more appropriate than the hardware models. Models based on the perception, however have a more complex structure and therefore require new metrics and conversion routines.
  • Comparisons between techniques of correspondence based on correlation-based features. Correlation techniques provide dense disparity maps with all the scene depth information which makes them more complete but also more complex. In terms of techniques based on features only provide certain regions depth information so they are simpler but they do not provide as much information. We intend to compare them and determine if there is enough for a robot navigation techniques of the second group information.
  • Implement the system with two conventional cameras instead of with a stereo camera to reduce costs. Stereo cameras have a high cost as well as high calibration requirements. The use of conventional cameras would mean a significant reduction. Using techniques based on features is predictably less calibration requirements.
  • Incorporation in robotic system and test. We intend to test the system built in a realistic environment, incorporating it to the mobile platform that we have.

Industrial Computing and Artificial Intelligence (i3a)


Universidad de Alicante
Dpto.Ciencia de la Computación e Inteligencia Artificial
Grupo Informática Industrial e Inteligencia Artificial

Carretera San Vicente s/n
03690 San Vicente del Raspeig
Alicante (Spain)

Tel: (+34) 96 590 3400

Fax: (+34) 96 590 3464

For further information: informacio@ua.es, and for enquiries on this web server: webmaster@ua.es

Carretera San Vicente del Raspeig s/n - 03690 San Vicente del Raspeig - Alicante - Tel. 96 590 3400 - Fax 96 590 3464