In this paper we investigate SURF features for visual terrain classification for outdoor mobile robots. The image is divided into a grid and SURF features are calculated on the intersections of this grid. These features are then used to train a classifier that can differentiate between different terrain classes. Images of five different terrain types are taken using a single camera mounted on a mobile outdoor robot. We further introduce another descriptor, which is a modified form of the dense Daisy descriptor. Random forests are used for classification on each descriptor. Classification results of SURF and Daisy descriptors are compared with the results from traditional texture descriptors like LBP, LTP and LATP. It is shown that SURF features perform better than other descriptors at higher resolutions. Daisy features, although not better than SURF features, also perform better than the three texture descriptors at high resolution.
Abstract-In this paper we present a comparison of multiple approaches to visual terrain classification for outdoor mobile robots based on local features. We compare the more traditional texture classification approaches, such as Local Binary Patterns, Local Ternary Patterns and a newer extension Local Adaptive Ternary Patterns, and also modify and test three non-traditional approaches called SURF, DAISY and CCH. We drove our robot under different weather and ground conditions and captured images of five different terrain types for our experiments. We did not filter out blurred images which are due to robot motion and other artifacts caused by rain, etc. We used Random Forests for classification, and cross-validation for the verification of our results. The results show that most of the approaches work well for terrain classification in a fast moving mobile robot, despite image blur and other artifacts induced due to extremely variant weather conditions.
Abstract-Outdoor robots are faced with a variety of terrain types each possessing different characteristics. To ensure a safe traversal a robot has to infer the current ground surface from sensor readings. Recent techniques generate a model which predicts the terrain class from single vibration signals disregarding the temporal coherence between consecutive measurements. In this paper, we present a novel approach in which the final classification relies on the analysis of not only one, but several recent observations. Therefore, the probabilistic framework of the Bayes filter is adopted to the problem of terrain classification. We propose an adaptive approach which automatically adjusts its parameters according to the history of observations. To demonstrate the performance of our method we further describe and compare another technique based on temporal coherence. The evaluation using data collected from our RWI ATRV-Jr robot shows that our approach is both reactive and stable enough to detect fast terrain transitions and selective misclassifications.
Abstract-A safe traversal of a mobile robot in an unknown environment requires the determination of local ground surface properties. As a first step, a broad structure of the underlying environment can be established by clustering terrain sections which exhibit similar features. In this work, we focus on an unsupervised learning approach to segment different terrain types according to the clustering of acquired vibration signals. Therefore, we present a Markov random field-based clustering approach taking the inherent temporal dependencies between consecutive measurements into account. The applied generative model assumes that the class labels of neighboring vibration segments are generated by prior distributions with similar parameters. A temporally constrained expectation maximization algorithm enables the efficient estimation of its parameters considering a predefined set of neighboring vibration segments. Since the size of the neighbor set proves to be data-dependent, we derive a general means of estimating this set size from the observed data. We show that the Markov random field clustering approach generates valid models for a variety of driving speeds even in situations of frequent terrain changes.
Vibration signals acquired during robot traversal provide enough information to yield a reliable prediction of the current terrain type. In a recent approach, we combined a history of terrain class estimates into a final prediction. We therefore adopted a Bayes filter taking the posterior probability of each prediction into account. Posterior probability estimates, however, were derived from support vector machines only, disregarding the capability of other classification techniques to provide these estimates. This paper considers other classifiers to be embedded into our Bayes filter terrain prediction scheme, each featuring different characteristics. We show that the best classification results are obtained using a combined k-nearestneighbor and support vector machine approach which has not been considered for terrain classification so far. Furthermore, we demonstrate that other classification techniques also benefit from the temporal filtering of terrain class predictions.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.