Pierre Charbonnier

Detection and recognition of urban road markings using images

While road lane markings detection was extensively studied, in particular for intelligent vehicle applications, the detection and recognition of all kind of marking such as arrows, crosswalks, zebras, words, pictograms, continuous and discontinuous lane markings was drastically less studied. However, it has many potential applications in the design of advanced driver assistance systems, as well as for asset management along itineraries. An algorithm is proposed which is based on the following processing steps: marking pixel extraction, detection using connected components before Inverse Perspective Mapping and recognition based on the comparison with a single pattern or with repetitive rectangular patterns. The proposed algorithm is able to detect and recognize repetitive markings (such as crosswalks) as well as single patterns (such as arrows). We believe that the proposed algorithm can be extended easily to solve the problem of the identification of all types of markings.

Using Robust Estimation Algorithms for Tracking Explicit Curves

The context of this work is lateral vehicle control using a camera as a sensor. A natural tool for controlling a vehicle is recursive filtering. The well-known Kalman filtering theory relies on Gaussian assumptions on both the state and measure random variables. However, image processing algorithms yield measurements that, most of the time, are far from Gaussian, as experimentally shown on real data in our application. It is therefore necessary to make the approach more robust, leading to the so-called robust Kalman filtering. In this paper, we review this approach from a very global point of view, adopting a constrained least squares approach, which is very similar to the half-quadratic theory, and justifies the use of iterative reweighted least squares algorithms. A key issue in robust Kalman filtering is the choice of the prediction error covariance matrix. Unlike in the Gaussian case, its computation is not straightforward in the robust case, due to the nonlinearity of the involved expectation. We review the classical alternatives and propose new ones. A theoretical study of these approximations is out of the scope of this paper, however we do provide an experimental comparison on synthetic data perturbed with Cauchy-distributed noise.

Adjustment of Sonar and Laser Acquisition Data for Building the 3D Reference Model of a Canal Tunnel

In this paper, we focus on the construction of a full 3D model of a canal tunnel by combining terrestrial laser (for its above-water part) and sonar (for its underwater part) scans collected from static acquisitions. The modeling of such a structure is challenging because the sonar device is used in a narrow environment that induces many artifacts. Moreover, the location and the orientation of the sonar device are unknown. In our approach, sonar data are first simultaneously denoised and meshed. Then, above- and under-water point clouds are co-registered to generate directly the full 3D model of the canal tunnel. Faced with the lack of overlap between both models, we introduce a robust algorithm that relies on geometrical entities and partially-immersed targets, which are visible in both the laser and sonar point clouds. A full 3D model, visually promising, of the entrance of a canal tunnel is obtained. The analysis of the method raises several improvement directions that will help with obtaining more accurate models, in a more automated way, in the limits of the involved technology.

Evaluation of Robust Fitting Based Detection

Low-level image processing algorithms generally provide noisy features that are far from being Gaussian. Medium-level tasks such as object detection must therefore be robust to outliers. This can be achieved by means of the well-known M-estimators. However, higher-level systems do not only need robust detection, but also a confidence value associated to the detection. When the detection is cast into the fitting framework, the inverse of the covariance matrix of the fit provides a valuable confidence matrix.

A Revisited Half-Quadratic Approach for Simultaneous Robust Fitting of Multiple Curves

In this paper, we address the problem of robustly recovering several instances of a curve model from a single noisy data set with outliers. Using M-estimators revisited in a Lagrangian formalism, we derive an algorithm that we call Simultaneous Multiple Robust Fitting (SMRF), which extends the classical Iterative Reweighted Least Squares algorithm (IRLS). Compared to the IRLS, it features an extra probability ratio, which is classical in clustering algorithms, in the expression of the weights. Potential numerical issues are tackled by banning zero probabilities in the computation of the weights and by introducing a Gaussian prior on curves coefficients. Applications to camera calibration and lane-markings tracking show the effectiveness of the SMRF algorithm, which outperforms classical Gaussian mixture model algorithms in the presence of outliers.

3D Road Environment Modeling Applied to Visibility Mapping: An Experimental Comparison

Sight distance along the pathway plays a significant role in road safety and in particular, has a clear impact on the choice of speed limits. Mapping visibility distance is thus of importance for road engineers and authorities. While visibility distance criteria are routinely taken into account in road design, few systems exist for evaluating them on existing road networks. Most available systems comprise a target vehicle followed at a constant distance by an observer vehicle. This only allows to check if a given, fixed visibility distance is available: estimating the maximum visibility distance requires several passages, with increasing inter-vehicle intervals. We propose two alternative approaches for estimating the maximum available visibility distance, that exploit 3D models of the road and its close environment. These methods involve only one acquisition vehicle and use either active vision, more specifically 3D range sensing (LIDAR), or passive vision, namely, stereovision. The first approach is based on a Terrestrial LIDAR Mobile Mapping System. The triangulated 3D model of the road and its surroundings provided by the system is used to simulate targets at different distances, which allows for estimation of the maximum geometric visibility distance along the pathway in a quite flexible way. The second approach involves the processing of two views taken by digital cameras on-board an inspection vehicle. After road segmentation, the 3D road model is reconstructed which allows maximum roadway visibility distance estimation. Both approaches are described, evaluated and compared. Their pros and cons with respect to vehicle-following systems are also discussed.

An Image-Based Inspection System for Canal Tunnel Heritage

Nowadays, 33 canal tunnels, mainly bored during the 19th and 20th centuries, are still used in France for commercial navigation and boating. The preservation of this heritage is not only essential for cultural and historical reasons but also for economical necessities and security issues. A French partnership proposes to develop a prototype dedicated to the image-based inspection of canal tunnels. In this paper, we address the problems of image quality and location accuracy, which is particularly challenging since no conventional global referencing system is available in tunnels. We introduce the dynamic imaging system that has been tested during recording campaigns in Niderviller tunnel. Several solutions to locate the inspection barge have been also investigated. The experimental evaluation on the accuracy of the referencing methods is presented and analyzed.

Evaluating a Static Multibeam Sonar Scanner for 3D Surveys in Confined Underwater Environments

Mechanical Sonar Scanning (MSS) is a recent technology that allows sonar to be used for static measurements in the same way as Terrestrial Laser Scanners (TLS), which makes it an attractive tool for underwater infrastructure surveys. Nevertheless, the metrological capabilities of this type of device have been little explored in the literature, particularly in narrow and shallow environments. In this paper, we report on the experimental assessment of a recent MSS, the BlueView BV5000, in a lock. The 3D sonar scans performed with the system suspended from the surface are registered using an innovative algorithm that exploits external measurements from a total station and the symmetry of the structure. We review the different errors that impair sonar data, and compare the resulting point cloud to a TLS model that was acquired the day before, while the lock was completely emptied for maintenance. After correcting a tilt angle calibration error, the maximum difference is less than 10 cm, and the standard deviation is about 3 cm. Visual inspection shows that coarse defects of the masonry, such as stone lacks or cavities, can be detected in the MSS point cloud, while details smaller than 4 cm, e.g., damaged joints, are harder to notice.