Phis-Lbp: Feature Descriptor for Vehicle Detection

International audience

PV-Alert: A fog-based architecture for safeguarding vulnerable road users

International audience; High volumes of pedestrians, cyclists and other vulnerable road users (VRUs) have much higher casualty rates per mile; not surprising given their lack of protection from an accident. In order to alleviate the problem, sensing capabilities of smartphones can be used to detect, warn and safeguard these road users. In this research we propose an infrastructure-less fog-based architecture named PV-Alert (Pedestrian-Vehicle Alert) where fog nodes process delay sensitive data obtained from smartphones for alerting pedestrians and drivers before sending the data to the cloud for further analysis. Fog computing is considered in developing the architecture since it is an emer…

Nouvelle stratégie de contrôle éco-énergétique pour les véhicules hybrides et électriques – Gestion de la batterie et optimisation prédictive

Background subtraction for aerial surveillance conditions

International audience; The first step in a surveillance system is to create a representation of the environment. Background subtraction is widely used algorithm to define a part of an image that most time remains stationary in a video. In surveillance tasks, this model helps to recognize those outlier objects in an area under monitoring. Set up a background model on moving platforms (intelligent cars, UAVs, etc.) is a challenging task due camera motion when images are acquired. In this paper, we propose a method to support instabilities caused by aerial images fusing spatial and temporal information about image motion. We used frame difference as first approximation, then age of pixels is …

An Online Time Warping based Map Matching for Vulnerable Road Users’ Safety

International audience; High penetration rate of Smartphones and their increased capabilities to sense, compute, store and communicate have made the devices vital components of intelligent transportation systems. However, their GPS positions accuracy remains insufficient for a lot of location-based applications especially traffic safety ones. In this paper, we developed a new algorithm which is able to improve smartphones GPS accuracy for vulnerable road users' traffic safety. It is a two-stage algorithm: in the first stage GPS readings obtained from smartphones are passed through Kalman filter to smooth deviated reading. Then an adaptive online time warping based map matching is applied to…

Rapid prototyping platform for stream-oriented reconfigurable computing applications

In this paper we present a methodology and tool for rapid prototyping of real time image processing applications. We describe our design flow of multiprocessor system on chip (MPSoC) architectures based on hardware/software components. This methodology provides automated methods to specify, generate the hardware, software, and the architectural interfaces between them. Our methodology starts from system level specification of the application with parallel processes described in C-code. The processes communicate through an abstract channel called streams. We describe also the solution that we proposed to synthesize a custom bus architecture for the reconfigurable computing applications, whic…

A2Ba: Adaptive Background Modelling for Visual Aerial Surveillance Conditions

International audience; Background modelling algorithms are widely used to define a part of an image that most time remains stationary in a video. In surveillance tasks, this model helps to recognize those outlier objects in an area under monitoring. Set up a background model on mobile platforms (UAVs, intelligent cars, etc.) is a challenging task due camera motion when images are acquired. In this paper, we propose A2Ba, a robust method to support instabilities caused by aerial images fusing different information about image motion. We used frame difference as first approximation, then age of pixels is estimated. This latter gives us an invariability level of a pixel over time. Gradient di…

Deep Learning-Based Real-Time Object Detection in Inland Navigation

International audience; Semi-autonomous and fully-autonomous systems must have knowledge about the objects in their environment to ensure a safe navigation. Modern approaches implement deep learning techniques to train a neural network for object detection. This project will study the effectiveness of using several promising algorithms such as Faster R-CNN, SSD, and different versions of YOLO, to detect, classify, and track objects in near real-time fluvial domain. Since no dataset is available for this purpose in literature, we first started by annotating a dataset of 2488 images with almost 35 400 annotations for training the convolutional neural network architectures. We made this data s…

Real Time Robust Embedded Face Detection Using High Level Description

Face detection is a fundamental prerequisite step in the process of face recognition. It consists of automatically finding all the faces in an image despite the considerable variations of lighting, background, appearance of people, position/orientation of faces, and their sizes. This type of object detection has the distinction of having a very large intra-class, making it a particularly difficult problem to solve, especially when one wishes to achieve real time processing. A human being has a great ability to analyze images. He can extract the information about it and focus only on areas of interest (the phenomenon of attention). Thereafter he can detect faces in an extremely reliable way.…