Multiple-Model Based Verification of Road Data

Abstract

ABSTRACT:This paper describes a semi-automatic system for road verification based on high resolution imagery and 3D surface models. Potential update regions are identified by an object-wise verification of all existing database records. The proposed system combines several road detection and road verification approaches from current literature to form a more general solution. Each road detection / verification approach is realized as an independent module representing a unique road model combined with a corresponding processing strategy. The object-wise verification result of each module is formulated as a binary decision between the classes "correct road" and "incorrect road". These individual decisions are combined by Dempster-Shafer fusion, which provides tools for dealing with uncertain and incomplete knowledge about the statistical properties of the data. For each road detection / verification module a confidence function for the result is introduced that reflects the degree of correspondence of an actual test situation with an optimal situation according to the underlying road model of that module. A comparison with results from an EuroSDR test on road extraction demonstrate the strengths and limitations of the method.

SVM-based road verification with partly non-representative training data

“…The features were selected from each band such that changes in surface properties of the road object, shadow effects on the road and road markings are reflected in the feature space. For more comprehensive information concerning the feature selection, refer to [1].…”

“…For the purpose of automated road verification a modular system was presented in [1] that combines several stateof-the-art road detection modules to form a more general solution. The data fusion concept behind the approach requires a confidence or reliability metric for each module.…”

Multiple-Model Based Update of Belgian Reference Road Data

“…As the main contribution of this work is the definition of the model specific confidence functions this aspect is explained in more detail. Further information concerning the algorithms can be found in the denoted references and in (Ziems et al, 2010).…”

“…Then, the extracted lines are compared with the database and only if an object is not covered by appropriate linear objects it has to be inspected by a human operator. In (Ziems et al, 2010), this strategy is adapted to urban areas by introducing additional road models that were explicitly designed for the urban context. Similar to (Poulain et al, 2010) buildings and vegetation are explicitly considered as context objects and were combined by Dempster-Shafer theory.…”

Monitoring geometric properties of an existing forest road using airborne Lidar data

“…The auto-detection technique of a road on geospatial images was started more than thirty years ago (Mena, Malpica 2005;Amo et al 2006;Mayer et al 2006). This technique applies a high resolution spatial image (smaller than 1 m) in detecting geometry, topology, texture and especially colour (Lacoste et al 2005;Grote et al 2012;Mnih, Hinton 2012;Ziems et al 2012).…”