Louahdi Khoudour

Video Sequences Association for People Re-identification across Multiple Non-overlapping Cameras

This paper presents a solution of the appearance-based people re-identification problem in a surveillance system including multiple cameras with different fields of vision. We first utilize different color-based features, combined with several illuminant invariant normalizations in order to characterize the silhouettes in static frames. A graph-based approach which is capable of learning the global structure of the manifold and preserving the properties of the original data in a lower dimensional representation is then introduced to reduce the effective working space and to realize the comparison of the video sequences. The global system was tested on a real data set collected by two cameras installed on board a train. The experimental results show that the combination of color-based features, invariant normalization procedures and the graph-based approach leads to very satisfactory results.

A People Counting System Based on Dense and Close Stereovision

We present in this paper a system for passengers counting in buses based on stereovision. The objective of this work is to provide a precise counting system well adapted to buses environment. The processing chain corresponding to this counting system involves several blocks dedicated to the detection, segmentation, tracking and counting. From original stereoscopic images, the system operates primarily on the information contained in disparity maps previously calculated with a novel algorithm. We show that one can obtain a counting accuracy of 99% on a large data set including specific scenarios played in laboratory and on some video sequences shot in a bus during exploitation period.

Real-time passenger counting in buses using dense stereovision

We are interested particularly in the estimation of passenger flows entering or exiting from buses. To achieve this measurement, we propose a counting system based on stereo vision. To extract three-dimensional information in a reliable way, we use a dense stereo-matching procedure in which the winner-takes-all technique minimizes a correlation score. This score is an improved version of the sum of absolute differences, including several similarity criteria determined on pixels or regions to be matched. After calculating disparity maps for each image, morphological operations and a binarization with multiple thresholds are used to localize the heads of people passing under the sensor. The markers describing the heads of the passengers getting on or off the bus are then tracked during the image sequence to reconstitute their trajectories. Finally, people are counted from these reconstituted trajectories. The technique suggested was validated by several realistic experiments. We showed that it is possible to obtain counting accuracy of 99% and 97% on two large realistic data sets of image sequences showing realistic scenarios.

Background subtraction and 3D localization of moving and stationary obstacles at level crossings

This paper proposes an obstacle detection system for the purpose of preventing accidents at level crossings. In order to avoid the limits of already proposed technologies, this system uses stereo cameras to detect and localize multiple targets at the level crossing. In a first step, a background subtraction module is performed using the Color Independent Component Analysis (CICA) technique which allows to detect vehicles even if they are stopped (the main cause of accidents at Level Crossings). A novel robust stereo matching algorithm is then used to reliably localize in 3D each segmented object. Standard stereo datasets and real-world images are used to evaluate the performances of the proposed algorithm, showing the efficiency and robustness of the proposed video surveillance system.

A Video-Based Object Detection System for Improving Safety at Level Crossings

Improving transport users’ safety is one of the main priorities of research into transport system attractiveness. Level crossings are one of the most critical weak points involving road and rail users’ infrastructure. They have become increasingly dangerous and unsafe due to road and railway users’ behavior. Furthermore, rail and highway safety professionals from several countries must deal with the same subject: providing safer level crossing. Actions are planned in order to exchange and share knowledge on existing level crossings technologies between academic organizations and industrial operators, and provide experiments for improving the management of level crossing safety and performance. This has enabled the sharing of knowledge gained from research in order to improve safety at level crossings. This article provides research results about possible technological solutions to reduce the number of accidents at level crossings, in particular, discussion and proof of the effectiveness of the use of video sensing for object detection. The authors have tested and adapted a robust technique for moving object detection, which is followed by a new approach for 3D object localization.

People re-identification by classification of silhouettes based on sparse representation

The research presented in this paper consists in developing an automatic system for people re-identification across multiple cameras with non-overlapping fields of view. We first propose a robust algorithm for silhouette extraction which is based on an adaptive spatio-colorimetric background and foreground model coupled with a dynamic decision framework. Such a method is able to deal with the difficult conditions of outdoor environments where lighting is not stable and distracting motions are very numerous. A robust classification procedure, which exploits the discriminative nature of sparse representation, is then presented to perform people re-identification task. The global system is tested on two real data sets recorded in very difficult environments. The experimental results show that the proposed system leads to very satisfactory results compared to other approaches of the literature.

Object tracking using Harris corner points based optical flow propagation and Kalman filter

This paper proposes an objects tracking method using optical flow information and Kalman filtering. The basic idea of the proposed approach starts from the fact that interesting points based optical flow is more precise and robust when compared to the optical flow of the other pixels of objects. Firstly, objects to be tracked are detected basing on independent component analysis. For each detected object, Harris corner points are extracted and their local optical flow is calculated. The optical flow of the Harris points is then propagated using a Gaussian distribution based technique to estimate the optical flow of the remaining pixels. Finally, the estimated optical flow is corrected using an iterative Kalman Filter. Experimental results on real data set frames are presented to demonstrate the effectiveness and robustness of the method. This work is developed within the framework of the PANsafer project, supported by the ANR VTT program.

Robust visual tracking via MCMC-based particle filtering

We present in this paper a new visual tracking framework based on the MCMC-based particle algorithm. Firstly, in order to obtain a more informative likelihood, we propose to combine the color-based observation model with a detection confidence density obtained from the Histograms of Oriented Gradients (HOG) descriptor. The MCMC-based particle algorithm is then employed to estimate the posterior distribution of the target state to solve the tracking problem. The global system has been tested on different real datasets. Experimental results demonstrate the robustness of the proposed system in several difficult scenarios.

Gaussian propagation model based dense optical flow for objects tracking

In this paper we present a method for objects tracking within a surveillance zone. The tracking process starts by detecting moving objects using Independent Component Analysis. When a set of moving objects is detected, targets are extracted basing on an energy vector comparison strategy, which consists in clustering the pixels of the detected objects. Once the targets are extracted, the tracking is performed by calculating the optical flow of the pixels of the objects. This is achieved by a Harris points based optical flow propagation, followed by a Kalman filtering based correction. Experimental results are presented to demonstrate the effectiveness of the proposed method. This work is developed within the framework of PANsafer project (Towards a safer level crossing), supported by the Frensh ANR VTT program.

Using hidden markov model and dempster-shafer theory for evaluating and detecting dangerous situations in level crossing environments

In this paper we present a video surveillance system for evaluating and detecting dangerous situations in level crossing environments. The system is composed of the following main parts: a robust algorithm able to detect and separate moving objects in the perceived environment, a Gaussian propagation model based dense optical flow for objects tracking, a Hidden Markov Model to recognize trajectories of detected objects, and an uncertainty model using theory of evidence to calculate the level of danger allowing to detect dangerous situations in level crossings. This method is tested on real image sequences, and the results are discussed. This work is developed within the framework of PANsafer project, supported by the ANR VTT program.