Begise - Event detection system in escalators using intelligent video analysis
The advances on Video Content Analysis technologies, its lower cost and its growing maturity are the guiding principle for the development of an emerging and high potential business of technology-based products and services. Conceptually, a step forward has been taken from the use of traditional video systems (based on recording material which must be inspected and used by persons), to the use of machines which automatically analyse and extract the relevant information for each application.
The use of computer vision based systems and its use together with artificial intelligence methods is positioning as a trend in a wide variety of application sectors: the intelligent transport world (with Driver Assistance and Navigation Systems), the Biomedicine (for clinical decision support), the Traffic Systems and Infrastructure Management (for vehicle counting and classification), the security and biometry (facial recognition, event detection, entertainment and television (augmented reality), etc.
There are several reasons for this: the high performance and the practical feasibility of the Video Content Analysis technology, with falling costs of hardware components, with greater computing capacity, and the settlement of computer vision and artificial intelligence as mature technologies making a leap from the experimental scope to the industrial scope in the last 10 years. It is important to highlight the impact of this technology in the security and video surveillance sector, where the old CCTV systems are becoming powerful event analysis and interpretation systems, giving support to several applications of security, infrastructure management, forensic inspection etc.
The existence of cameras in CCTV systems has not guaranteed the persons' security in critical infrastructures such as train stations, airports, etc. since the high number of cameras and the 24/7 monitoring with the traditional need of a person to watch the scenes make it unfeasible to take preventive measures.
For example, the escalators (mechanical or electric) scene is typically monitored by CCTV cameras, since they are usually located in commercial centres, airports, public transport stations (metro, urban buses), etc. However, so far any analysis systems has not been applied to this field, even if risky situations may occur with an important impact in the security (fights, vandalism, etc.)
As any citizen service, it requires a set of basic security recommendations for its correct use with the aim of avoiding risky situations. Some of the recommendations mentioned below:
- Avoid to go up the escalator with sticks, walkers, prams or any kind of vehicles.
- Avoid to go up the escalator barefoot or with loose strings.
- Go up the escalator only when it is moving in the correct direction.
- Take steps carefully but resolutely going up and going down the escalator.
- Hold strongly the children with your hand.
- Hold strongly the small packages with your hand.
- Hold the handrail according to the movement of the escalator.
- Maintain the loose clothes far from the entrances and sides of the escalator.
- Do not use the stopped escalators for going up and down as normal escalators.
- Do not use an escalator for the transport of goods.
- Stay in the centre of the escalator and look forward.
- Maintain a constant control in the handrail.
- Do not leave the handbag or packages on the handrail.
- Do not stand on the sides.
- Do not run, do not sit, do not play, do not jump, etc.
- Go down resolutely and go far of the exit area of the escalator as soon as possible, since other passenger may be behind you.
To not follow these basic rules or recommendations could lead in minor or major accidents such as the loss of human lives (for example due to choking after a crowd of people in the exit area of the escalator).
The BEGISE project's general aim is to include the Video Content Analysis technologies to provide the analysis and interpretation capacity for typical scenes of escalator scenarios in a context of surveillance applications. To do this, a platform will be developed with remote processing units in stations. Moreover, computer vision and artificial intelligence technologies will be applied generating real time alarms at the time the escalator is being used incorrectly or in case of bad performance of the escalator.
An alarm management centre will be developed, where the information from remote units will be managed, stored and processed to ease the use of the information to security operators or alarm reception centres. This way the workers in charge of the security of the users of the area could act immediately after the anomaly and avoid this way any accident. The potential market of the BEGISE project is huge, strategic and international, due to the high number of escalators and ramps in the world. Moreover, the number of escalators and ramps is increasing very fast as the countries are considering the need of easing the access to elderly or disabled people.
DOMINION is taking part in the project as the project leader and coordinator. Dominion is a technology company and due to its market view it will be responsible of defining the requirements together with an end-user partner (Euskal Trenbide Sareak). VISUAL LINE will be responsible of the system design, being a technology company and coordinating the technical part of the project supported technologically by two leading technology centres: Vicomtech as an applied research technology centre in the multimedia technologies sector and Grupo de Tratamiento de Imágenes of Universidad Politécnica de Madrid specialised in analysis and compression. Finally, the managing company of the railways net of Gobierno Vasco, Euskal Trenbide Sareak (ETS), being an end-user, will provide its point of view about the system operation in real situations, taking part in the project without budget.
This project has received funding from the Ministry of Economy and Competitiviness. This program is funded by EU through the operational programme with the aim of "fostering quality technology development, innovation and research"