NEW TOOLS VIDEO ANALITICS BASED ON INTELLIGENT SENSOR

Minatulaev Sh. M.

ITASCKO,
e-mail: [email protected], [email protected] 

Abstract — Experimentally in an industrial system intelligent video shows the possibility of extending the functionality through the use of new tools for video analytics example retail chains. At the same time the possibility of creating intelligent sensors based on of the combined of sensors using a higher-level analytical systems. 

I. Introduction

In today's world-necessity surveillance tool for security monitoring and gathering intelligence. The appearance on the market of IP-cameras, with great functionality, capable of replacing a cascade of analog cameras by one device, a surveillance network demand increased markedly. Ability to obtain high-resolution images for subsequent processing, broadcast video in real-time optimization of video surveillance network costs by reducing the number of cameras - all of this is of great interest to enterprises of the transport sector, health care, retail chains, financial sector, as well as many others.Modern realities show that the use of a classical video surveillance is becoming less effective every year. The world has commissioned surveillance network for tens of thousands of cameras. Norma began to network hundreds of cameras. Automatically it was revealed problems of formation and scaling "giant" high load of information networks, distribution and storage of information flows. But if the problems described above can be solved by technical means, the search for relevant information becomes extremely difficult. In the classic CCTV (Closed-Circuit TeleVision, broadcast television system closed) there is only one search criterion - Date / Time. As soon as the user wants to use an event or an analysis of what is happening, it is required to connect the human resource that with this amount of information is ineffective [1].In dannoy? paper analyzes the problems of the classic video, the need, the basic problems of the development and implementation of video analytics technology and a new approach to their implementation to create a qualitatively different industrial IVN (Intelligent Video Surveillance).

II. Main part

Create as effective implementation of intelligent video surveillance, requires the coordination of the simultaneous solution of a number of problems, both in the effective management of information resources and in expanding the functionality of CCTV networks adaptively for a specific type of customer.A large number of cameras and the huge flow of information, which in real time can not handle full-time employees of the companies forced consumers to seek access to analytical tools. This paved the creation of IVN.Modern means of video analytics were born from the field of machine vision. To use them effectively requires advanced digital IP-camera. Advantages of IP-cameras are clear, IP-networks have higher data transfer speeds, resulting in possible transfer of high-quality full HD (High Definition) image. If necessary, you can record from a video camera to study in detail, scaled without loss of quality, which greatly helps in the analysis of the events (incidents).To transfer the IP-camera H.264 image format of the data have been used for intelligence, and MJPEG for face recognition. As a result of the trial operation has been concluded: need new formats, which would provide a number of new features: from motion compensation with high precision to a set of several adaptive coding methods.Creating a video surveillance system based on IP-based network requires not only knowledge of network technologies and smart segmentation of network traffic calculation and planning system file [3]. Of course, it is important to calculate the amount of disk space backup devices. In larger systems, the error in the calculation can result in significant additional costs associated with the acquisition of additional equipment or costly rework of the entire network infrastructure.It should be noted that video surveillance functional modern multi-service networks, which are built on the work of huge enterprises, trade, production, etc. not limited to, especially given the modern means of video analysis [1]. That is why the basis for our decision were laid characteristics such as scalability, reliability, high performance and fault tolerance. The basis for the whole system made IVN industrial ESB (Enterprise Service Bus).However, the main problem is the amount of CCTV entering and stored information [2]. New features that are provided for experiments created HVS system with an analysis of actions and events, a huge reservoir made it possible to weed out irrelevant information to the user and to focus on the interested events.IVN is the basis of video analytics tools. Basis analytics - a sensor is isolated from the frame stream by specialized mathematical algorithms to-date information with the formation of labels and accompanying meta information. That is a tool that generates event-driven information to help better and faster to manage incidents and their associated video data - both online and in operations with archives [2]. Specialized sensors form a specific set of signals that can be interpreted as a particular event, for example, display the current image to the operator. The more video analytics such sensors, the more data that can be interpreted as a high-level tools.Standard sensors analytics today is a motion detector, a broken cross, being in the zone, heat maps activity trajectory.

Fig. 1. Example of the sensors

All modern detectors basically emulate similar hardware solutions. Only the cost of modern CCTV orders of magnitude more expensive. Hence it is clear that the consumer wants to have something more than a simple security monitoring - this should provide our video analytics [2].several sets of standard sensors have been chosen to create a new analytical tool. The data from each set of analyzes own expert system (ES), which forms a data set. These are aggregated into a top-level expert system that interprets the information based on the user's request or the request of the preset template.The resulting system is incorporated into the three-dimensional effect, as a priori dimensional incoming image. We need to analyze the three-dimensional space, and place it in the anomaly. We also need to know the position in space of any visible object in three-dimensional coordinates [1]. Furthermore, we need to have different views of visual space without increasing the number of video cameras.

Fig. 2. Example actions of several expert systems average level

After you create a top-level expert system and connect it to the reporting tool, we were able to interpret the data on a more qualitative level - adaptive for the consumer (in this example - the network operator retail). When connecting to the system directories schemes of placing the goods on the shelves of an opportunity to get statistics of the dynamics of movement of goods, customer interests (for example, the product was taken off the shelf, but the buyer returned it to the place), and more. When connecting the three-dimensional multi-angle classifiers of human bodies, animals and objects recognition and tracking algorithms was obtained the opportunity to know the customer growth and to accompany him automatically from camera to camera, which is important for the consumer.

III. Conclusion

Meta information from analytics is essential. Application add-ins in the form of expert systems on the basis of mathematical data processing allows to obtain analytical tools qualitatively different level. It is this set of data allows you to search in the archive to a qualitatively different level, for example, in correlation with various events.This increases the quality of the execution of the business tasks of the consumer, who does not dare to earlier.

IV. References

 [1] Torsten Anst?dt, Ivo Keller, Harald Lutz. Intelligente Video-analyse: Handbuch für die Praxis. Wiley-VCH, 2010[2] Shan, C., Porikli, F., Xiang, T., Gong, S. (Eds.). Video Analytics for Business Intelligence. 2012, ISBN 978-3-642-28597-4[3] Jean-Yves Dufour. Intelligent Video Surveillance Systems. 2012, Wiley-ISTE. ISBN: 978-1-84821-433-0

"Microwave equipment and telecommunication technologies", Intern. Crimea. Conf (KryMiKo'2014). (24; 2014; Sevastopol). "Microwave equipment and telecommunication technologies", Intern. Crimea. Conf (CriMiCo '2014) (24; 2014; Sevastopol). Minatulaev, M. Sh New analytics tools based on smart sensors / M. Sh Minatulaev // 24-th International Crimean Conference "Microwave equipment and telecommunication technologies "(KryMiKo'2014), 7-13 September 2014, Sevastopol, Crimea, Russia: materials conf. :. 2 m - Sevastopol, 2014 - Volume 1, Section. 3a / 3: Processing and information protection. - S. 378-379: ph. - Bibliography. at the end of Art. (3). - ISBN 978-966-335-413-2