Although numerous attempts are being made to address the issue of robot vision and navigation, many inherent drawbacks are found in those approaches. “Simultaneous Localization And Mapping” (SLAM) is one of the major problems encountered by researchers building mobile robots, which is none other than dynamic map building. Further, irrespective of the type of vision sensor or its accuracy, there is always an element of non-linearity of acquired data, due to various environmental conditions. Although researchers have deployed several techniques other than vision to sense the environment, such as SONAR and LASER guidance, the application space becomes quite narrow due to implementation difficulties and cost-performance pitfalls. The aspiration of this paper is to discuss the development of a generic and extensible robot navigation framework making use of computer vision and advanced image processing techniques, with a view to construct robotic applications for industry, academia and entertainment. The prototype is constructed making use of two cameras, one mounted in an aerial position overlooking the terrain and the other on the robot itself. The intelligence gathered through the cameras would be communicated back to the robot to help it in autonomous navigation. By performing image processing and computation routines on a workstation, the Author expects to achieve two objectives; relieving the robot from excessive processing requirements and in turn minimizing the cost.
|Keywords:||Autonomous Navigation, Image Analysis, Computer Vision, Scene Understanding|
Research Student, Department of Computing, Informatics Institute of Technology, Wellawatta, Sri Lanka
Lecturer, Department of Computing, Informatics Institute of Technology, Colombo, Sri Lanka
There are currently no reviews of this product.Write a Review