Indoor autonomous navigation using visual memory and pattern tracking

O. Ait Aider, T. Chateau, J.T. Lapresté
<span title="">2004</span> <i title="British Machine Vision Association"> <a target="_blank" rel="noopener" href="" style="color: black;">Procedings of the British Machine Vision Conference 2004</a> </i> &nbsp;
The paper deals with autonomous environment mapping, localisation and navigation using exclusively monocular vision and multiple 2D pattern tracking. The environment map is a mosaic of 2D patterns detected on the ceiling plane and used as natural landmarks. The robot is able to reproduce learned trajectories defined by key images representing the visual memory. The pattern tracker is based on particle filetring. It uses both image contours and gray scale level variations to track efficiently 2D
more &raquo; ... patterns on cluttered background. An original observation model used for filter state updating is presented. . This is not straightforward for any density function. It is then convenient to use a so called proposal density π¨S i¢ and from which samples can be generated. In this paper we use the Condensation algorithm which is a particle filter version where the proposal density is equal to the prior one π¨S i¢ Pattern modelling The pattern model must enable not only near real time tracking but also automatic generation and recognition. It must be complex enough to discard ambiguities due to the presence of objects in the background similar to parts of the model and simple enough to reduce the computational cost of tracking. The structure of a pattern is built in two levels: by Ω σ ¦ λẗ he one dimensional Gaussian function with standard deviation σ . Assuming that the probabilities p¨Z j¢ k ¥ S i¢ k © are mutually independent, it results that The characterization of the auto-correlation function by the parameter vector V j for each model point improve the precision of the estimate of p¨Z k ¥ S i¢ k © . Indeed, The intercorrelation may decrease faster for a highly textured contour point than for a point on a contour defined by two large and homogeneous regions (Figure 7 ). In addition, this allows to define a criterium to select or reject some contour points during the pattern model building phase. Moreover, note that this measure is robust to illumination changes thanks to normalization.
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="">doi:10.5244/c.18.68</a> <a target="_blank" rel="external noopener" href="">dblp:conf/bmvc/Ait-AiderCL04</a> <a target="_blank" rel="external noopener" href="">fatcat:627arwql75gb7kk6ogvcmq6gni</a> </span>
<a target="_blank" rel="noopener" href="" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href=""> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> Publisher / </button> </a>