Simultaneous real-time object recognition and pose estimation for artificial systems operating in dynamic environments
Year: 2013
Type: dissertation
Abstract: Recent advances in technology have increased awareness of the necessity for automated systems in \npeople’s everyday lives. Artificial systems are more frequently being introduced into environments \npreviously thought to be too perilous for humans to operate in. Some robots can be used to extract \npotentially hazardous materials from sites inaccessible to humans, while others are being developed \nto aid humans with laborious tasks. \nA crucial aspect of all artificial systems is the manner in which they interact with their immediate surroundings. \nDeveloping such a deceivingly simply aspect has proven to be significantly challenging, as \nit not only entails the methods through which the system perceives its environment, but also its ability \nto perform critical tasks. These undertakings often involve the coordination of numerous subsystems, \neach performing its own complex duty. To complicate matters further, it is nowadays becoming \nincreasingly important for these artificial systems to be able to perform their tasks in real-time. \nThe task of object recognition is typically described as the process of retrieving the object in a database \nthat is most similar to an unknown, or query, object. Pose estimation, on the other hand, involves \nestimating the position and orientation of an object in three-dimensional space, as seen from an observer’s \nviewpoint. These two tasks are regarded as vital to many computer vision techniques and regularly serve as input to more complex perception algorithms. \nAn approach is presented which regards the object recognition and pose estimation procedures as \nmutually dependent. The core idea is that dissimilar objects might appear similar when observed \nfrom certain viewpoints. A feature-based conceptualisation, which makes use of a database, is implemented \nand used to perform simultaneous object recognition and pose estimation. The design \nincorporates data compression techniques, originally suggested by the image-processing community, \nto facilitate fast processing of large databases. \nSystem performance is quantified primarily on object recognition, pose estimation and execution time \ncharacteristics. These aspects are investigated under ideal conditions by exploiting three-dimensional \nmodels of relevant objects. The performance of the system is also analysed for practical scenarios \nby acquiring input data from a structured light implementation, which resembles that obtained from \nmany commercial range scanners. \nPractical experiments indicate that the system was capable of performing simultaneous object recognition \nand pose estimation in approximately 230 ms once a novel object has been sensed. An average \nobject recognition accuracy of approximately 73% was achieved. The pose estimation results were \nreasonable but prompted further research. The results are comparable to what has been achieved using \nother suggested approaches such as Viewpoint Feature Histograms and Spin Images.
Authors: Van Wyk, Frans Pieter
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W1481175453