The growing mobility of people and goods has a very high societal cost in terms of traffic congestion and of fatalities and injured people every year. The management of a road network needs efficient ways for assessment at minimal costs.
Road monitoring is a relevant part of road management, especially for safety, optimal traffic flow and for investigating new sustainable transport patterns. Current monitoring systems based on video lack of optimal usage of networks and are difficult to be extended efficiently.
The ORUSSI project focuses on road monitoring through a network of roadside sensors (mainly cameras) that can be dynamically deployed and added to the surveillance systems in an efficient way.
The main objective of the project is to develop an optimized platform offering innovative real-time media (video and data) applications for road monitoring in real scenarios. We exploit low-level efficient image features in order to enable our distributed system to extract semantic information from the imagery and to optimize the video compression adaptively.
In the BSc Thesis project of Leonardo Galteri we developed an application for real time low-level feature extraction that runs at 20 fps on a commercial surveillance camera. We extract image corners with a modified version of the FAST algorithm, and edges with the Sobel algorithm. The FAST feature detector is based on the concept of segment test and its efficiency can be increased through machine learning. To the end of removing the learning step we modified the FAST algorithm in order to cope with generic scenes and to improve its performance. Specifically we improved the early rejection procedure and employed a 9-pixel segment test instead of the 12-pixel one of FAST. Furthermore we perform a spatial clustering of the features using K-means in order to obtain a coarse estimation of feature distribution. Features can be visualized with a browser using our provided HTML5 and AJAX based web client.