International audienceThe object of the TerraMobilita/iQmulus 3D urban analysis benchmark is to evaluate the current state of the art in urban scene analysis from mobile laser scanning (MLS) at large scale. A very detailed semantic tree for urban scenes is proposed. We call analysis the capacity of a method to separate the points of the scene into these categories (classification), and to separate the different objects of the same type for object classes (detection). A very large ground truth is produced manually in two steps using advanced editing tools developed especially for this benchmark. Based on this ground truth, the benchmark aims at evaluating both the classification, detection and segmentation quality of the submitted results

Abstract:In this paper, we present a novel framework for detecting individual trees in densely sampled 3D point cloud data acquired in urban areas. Given a 3D point cloud, the objective is to assign point-wise labels that are both class-aware and instance-aware, a task that is known as instance-level segmentation. To achieve this, our framework addresses two successive steps. The first step of our framework is given by the use of geometric features for a binary point-wise semantic classification with the objective of assigning semantic class labels to irregularly distributed 3D points, whereby the labels are defined as "tree points" and "other points". The second step of our framework is given by a semantic segmentation with the objective of separating individual trees within the "tree points". This is achieved by applying an efficient adaptation of the mean shift algorithm and a subsequent segment-based shape analysis relying on semantic rules to only retain plausible tree segments. We demonstrate the performance of our framework on a publicly available benchmark dataset, which has been acquired with a mobile mapping system in the city of Delft in the Netherlands. This dataset contains 10.13 M labeled 3D points among which 17.6% are labeled as "tree points". The derived results clearly reveal a semantic classification of high accuracy (up to 90.77%) and an instance-level segmentation of high plausibility, while the simplicity, applicability and efficiency of the involved methods even allow applying the complete framework on a standard laptop computer with a reasonable processing time (less than 2.5 h).

a b s t r a c tIn the past two decades, building detection and reconstruction from remotely sensed data has been an active research topic in the photogrammetric and remote sensing communities. Recently, effective high level approaches have been developed, i.e., the ones involving the minimization of an energetic formulation. Yet, their efficiency has to be balanced by the amount of processing power required to obtain good results.In this paper, we introduce an original energetic model for building footprint extraction from high resolution digital elevation models (≤1 m) in urban areas. Our goal is to formulate the energy in an efficient way, easy to parametrize and fast to compute, in order to get an effective process still providing good results.Our work is based on stochastic geometry, and in particular on marked point processes of rectangles. We therefore try to obtain a reliable object configuration described by a collection of rectangular building footprints. To do so, an energy function made up of two terms is defined: the first term measures the adequacy of the objects with respect to the data and the second one has the ability to favour or penalize some footprint configurations based on prior knowledge (alignment, overlapping, . . . ). To minimize the global energy, we use a Reversible Jump Monte Carlo Markov Chain (RJMCMC) sampler coupled with a simulated annealing algorithm, leading to an optimal configuration of objects. Various results from different areas and resolutions are presented and evaluated. Our work is also compared with an already existing methodology based on the same mathematical framework that uses a much more complex energy function. We show how we obtain similarly good results with a high computational efficiency (between 50 and 100 times faster) using a simplified energy that requires a single data-independent parameter, compared to more than 20 inter-related and hard-to-tune parameters.

No abstract

No abstract