Inter-Image Statistics for Mobile Robot EnvironmentModeling

This book presents a statistical learning frameworkfor inferring geometric structures from images.The proposed framework computes dense range maps oflocations in the environment using onlyintensity images and very limited amount of rangedata as an input. This is achieved by integrating andanalyzing the statistical relationships between thevisual data and the available depth on terms of smallpatches. The scientific issue is to represent thiscorrelation such that it can be used to recover rangedata where missing. Markov Random Fields are used asa model to capture the local statistics of theintensity and range.Experiments on real-world data are conducted underdifferent configurations to demonstrate thefeasibility of the method. In particular, theapplication is in mobile robotics, where inferringthe 3D layout of indoor environments is a criticalproblem for achieving exploration and navigationtasks.Additionally, the proposed method is used inthe color correction and augmentation problem withthe specific application to underwater images.

Autorentext

is a full-time researcher in the Robotics and Advanced Manufacturing Section at the Research and Advanced Studies Centre (CINVESTAV) Campus Saltillo, Mexico. Her research interests are in computational vision, sensor and data fusion, environment modeling, perception for mobile robots and probabilistic methods.

Klappentext
This book presents a statistical learning framework for inferring geometric structures from images. The proposed framework computes dense range maps of locations in the environment using only intensity images and very limited amount of range data as an input. This is achieved by integrating and analyzing the statistical relationships between the visual data and the available depth on terms of small patches. The scientific issue is to represent this correlation such that it can be used to recover range data where missing. Markov Random Fields are used as a model to capture the local statistics of the intensity and range. Experiments on real-world data are conducted under different configurations to demonstrate the feasibility of the method. In particular, the application is in mobile robotics, where inferring the 3D layout of indoor environments is a critical problem for achieving exploration and navigation tasks. Additionally, the proposed method is used in the color correction and augmentation problem with the specific application to underwater images.