Claude Laurgeau

Vehicle detection combining gradient analysis and AdaBoost classification

This paper presents a real-time vision-based vehicles rear detection system using gradient based methods and Adaboost classification, for ACC applications. Our detection algorithm consists of two main steps: gradient driven hypothesis generation and appearance based hypothesis verification. In the hypothesis generation step, possible target locations are hypothesized. This step uses an adaptive range-dependant threshold and symmetry for gradient maxima localization. Appearance-based hypothesis validation verifies those hypothesis using AdaBoost for classification with illumination independent classifiers. The monocular system was tested under different traffic scenarios (e.g., simply structured highway, complex urban environments, varying lightening conditions), illustrating good performance.

Comparative synthesis of the 3 main European projects dealing with Cooperative Systems (CVIS, SAFESPOT and COOPERS) and description of COOPERS Demonstration Site 4

Todays motorways utilize innovative technologies to manage a real-time traffic information database which supplies users with information via FM 107.7 radio, variable message panels and the Internet. In the future, information will be supplied directly to intelligent in-vehicle systems based on GPS technology. This data exchange will be a two-way process, with the vehicle sending information to the infrastructure, and even to other vehicles. Driver, vehicle and infrastructure will form an interactive triangle in which information exchange performs a key function. From this perspective ASFA members are participating to the CVIS, SAFESPOT and COOPERS projects to use the data collecting on the infrastructure, including in-vehicle data, to provide better information on board the vehicle. In this paper, we propose a comparative synthesis of these 3 major European projects dealing with cooperative systems; we introduce the contribution of ASFA members for CVIS and SAFESPOT; a section is dedicated to their contribution for COOPERS. We also introduce the European strategy regarding Intelligent Transport Systems.

Real-time crash avoidance system on crossroads based on 802.11 devices and GPS receivers

In this paper we study the contribution of inter-vehicular communication in ADAS applications. We thus propose a collaborative system on the crossroads using our 802.11g communications tools and a low cost GPS receiver. Once the vehicles positions exchanged, the crash avoidance is performed by predicting the future positions of both cars and calculating the time to impact and the region of high risk. The prediction is biased in time and space. The time error is due to the GPS and the communication latencies. We thus propose an estimation of both latencies and integrate them in the prediction loop. The space error is caused by the uncertainty of the positions delivered by the GPS receivers. We compensate this error not by considering a deterministic position of the vehicle but by a probability of being in a region of space and by proposing a Kalman filter which first reduces noise on the positioning and second estimates the variance of error on the measurement. This approach is validated through real scenarios of road crossing. We show thus the contribution of the cooperation via the communication devices in the reduction of accidents rate on the crossroads

Digitizing and 3D modeling of urban environments and roads using vehicle-borne laser scanner system

In this paper we present a system for three-dimensional environment modeling. It consists of an instrumented vehicle equipped with a 2D laser range scanner for data mapping, and GPS, INS and odometers for vehicle positioning and attitude information. The advantage of this system is its ability to perform data acquisition during the vehicle navigation; the sensor needed being a basic 2D scanner with opposition to traditional expensive 3D sensors. This system integrates the laser raw range data with the vehicles internal state estimator and is capable of reconstructing the 3D geometry of the environment by real-time geo-referencing. We propose a high level representation of the urban scene while identifying automatically and in real time some types of existing objects in this environment. Thus, our modeling is articulated around three principal axes: the segmentation, decimation, the 3D reconstruction and visualization. The road is the most important object for us; some road features like the curvature and the width are extracted.

PUVAME - New French Approach for Vulnerable Road Users Safety

In France, about 33% of roads victims are VRU. In its 3rd framework, the French PREDIT includes VRU Safety. The PUVAME project was created to generate solutions to avoid collisions between VRU and bus in urban traffic. An important part of these collisions take place at intersection or bus stop. In this paper, we detail the hardware and software architecture designed and developed in the project. This solution is based on offboard cameras observing particular places (intersection and bus stop in our case) to detect and track VRU present in the environment. The position of the bus is also computed and a risk of collision between each VRU and the bus is determined. In case of high risk of collision, the bus driver is warned. The HMI to warn the bus driver is also described. Finally, some experimental results are presented

Multi-sensor data fusion for land vehicle localization using /sup RT/MAPS

In this paper we present the use of /sup RT/MAPS in multi-sensor data fusion for land vehicle localization. The GPS, INS and odometers data time synchronisation and data logging and processing are performed in real time using our software /sup RT/MAPS. The fusion algorithm based on Kalman filtering integrates information coming from a single GPS receiver with inertial navigation system and wheel speed encoders. The filter is used to enhance positioning accuracy, especially during periods of losses of GPS signal and to achieve submeter accuracy for the positioning of a land vehicle. The results prove that /sup RT/MAPS is a good framework for prototyping multi-sensor automotive application. The integration results show that this system can be used as an autonomous localisation system for more than 10 minutes of GPS data outage.

Boosted algorithms for visual object detection on graphics processing units

Nowadays, the use of machine learning methods for visual object detection has become widespread. Those methods are robust. They require an important processing power and a high memory bandwidth which becomes a handicap for real-time applications. The recent evolution of commodity PC computer graphics boards (GPU) has the potential to accelerate those algorithms. In this paper, we present a novel use of graphics hardware for object detection in advanced computer vision applications. We implement a system for object-detection based on AdaBoost [1]. This system can be tuned to run partially or totally on the GPU. This system is evaluated with two face-detection applications. Those applications are based on the boosted cascade of classifiers: Multiple Layers Face Detection (MLFD), and Single Layer Face Detection (SLFD). We show that the SLFD implementation on GPU performs up to nine times faster than its CPU counterpart. The MLFD, in the other hand, can be accelerated using the (GPU) and performs up to three times faster than the CPU. To the best of our knowledge, this is the first attempt to implement a sliding window technique for visual object-detection on GPU, with promessing performance.

A new method and a clinical case for computer assisted dental implantology

This paper presents a new method for dental implant surgery. A pre-operative planning software is used to work with CT scanner data. Implant fixtures are placed with the help of a 3D reconstructed model of the patient’s jaw. An accurate robot is then used to drill a jaw splint, at the locations determined with the planning software, in order to make a surgical guide. The matching between image and robot referentials is performed with radio-opaque markers attached in a specific way to the jaw splint. A clinical case of this new technique is presented.

Computer Assisted Dental Implantology: A New Method and a Clinical Validation

This paper presents a new method for dental implant surgery. A preoperative planning software is used to work with CT scanner data. Implant fixtures are placed with the help of a 3D reconstructed model of the patients jaw. An accurate robot is then used to drill a jaw splint, at the locations determined with the planning software, in order to make a surgical guide. A validation case of this new technique is also presented.

Dense disparity estimation using Gabor filters and image derivatives

We tackle the recurrent problem of disparity estimation since the mapping from disparity to depth is well understood while the automatic disparity extraction is still subject to errors. We propose to use the image derivatives with the phase-based approach to overcome the tuning problem of the filter. Moreover we propose a quadratic model for the singularities neighborhood detection. The approach is characterized by the simplicity of its implementation. It also provides dense and accurate disparity maps. A numerical error analysis shows that the results are very satisfactory.