ARGOS
(Automatic Remote Grand Canal Observation System)

A DISTRIBUTED VISION SYSTEM FOR BOAT TRAFFIC MONITORING
IN THE VENICE GRAND CANAL

D. Bloisi,  L. Iocchi,  G.R. Leone,  R. Pigliacampo
Dipartimento di Informatica e Sistemistica
University of Rome ”Sapienza”, via Ariosto 25, Rome, Italy
iocchi@dis.uniroma1.it, leone@dis.uniroma1.it
L.Tombolini
ECOTEMA-Ingegneria per l’Ambiente srl
Cannaregio 463, Venice, Italy
ltombolini@ecotema.it
 L.Novelli
Archimedes
via Tagliamento 9, Rome, Italy
lnovelli@archimedes.it



Introduction


ARGOS project (Automatic Remote Grand Canal Observation System) is a video-surveillance system for boat traffic monitoring, measurement and management along the Grand Canal of Venice. This new system will answer to the specific requirements for the boat navigation rules in Venice while providing a combined unified view of the whole Grand Canal waterway. Such features far exceed the performance of any commercially available product. Therefore, a specific software has been developed, based on the integration of advanced automated image analysis techniques.

The ARGOS system is going to control a waterway of about 4 km length, 80 to 150 meters width, through 14 observation points (Survey Cells). The system is based on the use of groups of IR/VIS cameras, installed just below the roof of several buildings leaning over the Grand Canal. Each survey cell is composed of 4 optical sensors: one center wide-angle (90 degree), orthogonal to the navigation axis, two side deep-field cameras (50-60 degree), and a pan-tilt-zoom camera for high resolution acquisition of boat details (e.g., license plates).

The main ARGOS functions are: 1) Optical detection and tracking of moving targets present in the FOV; 2) Computing position, speed and heading of any moving target within the FOV of each camera; 3) Elaboration at survey cell level of any event (target appears, exits, stops, starts within the cells FOV) and transmission of any event to the Control Center; 4) Connecting all the track segments related to the same target in the different cameras FOV into a unique trajectory and track ID; 5) Recording all the video frames together with the graphical information related to track IDs and trajectories; 6) Rectifying all the camera frames and stitching them into a composite plain image so as to show a plan view of the whole Grand Canal; 7) Allowing the operator to graphically select any target detected by the system and automatically activating the nearest PTZ camera to track the selected target.


Main techniques used for image analysis and tracking


SEGMENTATION BASED ON BACKGROUND SUBTRACTION

Our segmantation method is based on background modelling and subtraction that takes into account: gradual and sudden illumination changes (such as clouds), high frequency background objects (waves), changes in the background geometry (parked boats). Our approach models the background with a mixture of Gaussians. The system computes the bar chart for every pixel (i.e., the approximation of the distribution) in the RGB color space and it clusters the raw data in sets based on distance in the color space. 



Moreover, optical flow analysis is used to solve under-segmentation cases arising from boats passing close by in opposite directions, while an extended k-means algorithm has been developed to cluster blobs avoiding most of the over-segmentation cases.
A new clustering method, Rek-means, has been developed in order to overcome some of the limitations of k-means. Rek-means provides better results in clustering data coming from different Gaussian distributions; it does not require to specify k beforehand; it maintains real-time performance.



MULTI-HYPOTHESIS KALMAN FILTER TRACKING

The multi-target tracking problem considered in this project has been solved by using a set of Kalman filters and a Nearest Neighbors approach to data association. Moreover, in order to consider uncertainty in data association and filtering, a multi-hypothesis tracking has been implemented. The steps of the filter are: track formation, track update, track split, track merge, track deletion. These are determined by evaluating the parameters of the filter.



RECTIFICATION

Image rectification is used to produce a panoramic view for each cell (FOV larger than 180 degrees) and for a top-view image that is stitched on top of GIS and orthophoto images. Rectification also allows for converting image coordinates into metric coordinates (in particular, we use Gauss-Boaga coordinates) and thus to geo-referentiate the boats in the Grand Canal and estimate their velocity.


VIDEOS

Tracking demonstration (speed up) 9 MB

TV RAI-TG1 November 2007 (In Italian)


PUBLICATIONS

D. D. Bloisi, L. Iocchi, A. Pennisi, L. Tombolini.
ARGOS-Venice Boat Classification.
In Proc. of 12th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS), 2015, 2015.

D. D. Bloisi, L. Iocchi.
Independent Multimodal Background Subtraction.
In Proc. of the Third Int. Conf. on Computational Modeling of Objects Presented in Images: Fundamentals, Methods and Applications,, pp. 39-44, 2012.

L. Iocchi, L. Novelli, L. Tombolini, M. Vianello.
Automatic Real-Time River Traffic Monitoring Based on Artificial Vision Techniques.
In International Journal of Social Ecology and Sustainable Development (IJSESD), volume 1(2), pp. 40-51, 2010.

D. D. Bloisi, L. Iocchi.
ARGOS - A Video Surveillance System for Boat Trafic Monitoring in Venice.
International Journal of Pattern Recognition and Artificial Intelligence. Vol. 3(7), pp. 1477-1502, 2009.

D. D. Bloisi, L. Iocchi.
Rek-means: A k-means based clustering algorithm.
In Computer Vision Systems, volume 5008 of  LNCS, pages 109--118. Springer, 2008.

D. D. Bloisi, L. Iocchi, G. R. Leone, R. Pigliacampo, L. Tombolini, L. Novelli
A Distributed Vision System for Boat Traffic Monitoring in the Venice Grand Canal.
In Proc. of 2nd Int. Conf. on Computer Vision Theory and Applications (VISAPP), pp. 549--556, ISBN: 978-972-8865-74-0, 2007.

Roberta Pigliacampo
Multi-Tracking of Moving Objects with Unreliable Sensors for Mobile Robotic Platforms.
Extended abstract from Master Thesis. University of Rome “La Sapienza”, 2006.

ACKNOWLEDGEMENTS

The project has been realized thanks to the view of the future and to the active participation of the City Council of Venice. In particular, special thanks to Lord Vice-Major of Venice, On. Michele Vianello, for his foresight in applying innovative technologies in the delicate and complex historical city as Venice. We are also grateful to the Responsible Manager Arch. Manuele Medoro and his staff for their constant support and commitment.