Multi-sensor effective application accelerates the pace of security intelligence
来源:Dongguan Yanghan Industrial Co., Ltd. 时间:2020-09-09
sensor is a kind of sensor that can sense information, and realize the transmission, processing, storage, display, recording and control of the sensed information And other required detection devices. Sensor technology, together with computer technology and communication technology, are called the three pillars of information technology. It is the first link to realize automatic detection and automatic control.
With the continuous development and progress of various software and hardware technologies related to sensor technology, multi-sensor information fusion So far, it has formed and developed into a specialized technology for comprehensive information processing, and it will soon be applied to various fields such as industrial robots, intelligent detection, automatic control, traffic management, and medical diagnosis. In the process of wide application, multi-sensor information fusion technology has also attracted more and more attention.
The basic principle of multi-sensor information fusion technology Just like the process of the human brain's comprehensive information processing, various sensors are used for multi-level and multi-space information complementary and optimized combination processing, and finally a consistent interpretation of the observation environment is produced. In this process, we must make full use of multi-source data for reasonable control and use, and the ultimate goal of information fusion is to derive more useful information based on the separated observation information obtained by each sensor, through the combination of multiple levels and multiple aspects of the information. This not only takes advantage of the cooperative operation of multiple sensors, but also comprehensively processes data from other information sources to improve the intelligence of the entire sensor system.
The multi-sensor fusion system has four notable features:
1. Information redundancy: For a certain feature of the environment, it can be obtained through multiple sensors (or multiple different moments of a single sensor) Multiple pieces of information are redundant and have different reliability. Through fusion processing, more accurate and reliable information can be extracted from them. In addition, the redundancy of information can improve the stability of the system, thereby avoiding the impact on the entire system caused by the failure of a single sensor.
2, complementary information Performance: Different types of sensors can provide the system with different types of information. The objects described by these information are different environmental characteristics, and they are complementary to each other. If a coordinate space composed of all features is defined, then the information provided by each sensor only belongs to a subspace of the entire space, and is independent of the space formed by other sensors.
3. Timeliness of information processing: each The processing processes of the sensors are independent of each other, and the entire processing process can adopt a parallel heat conduction processing mechanism, so that the system has a faster processing speed and provides more timely processing results.
4. Low cost of information processing: Multiple sensors can spend less cost to get the amount of information equivalent to a single sensor. On the other hand, if the information provided by a single sensor is not used to implement other functions, the sum of the cost of a single sensor and the cost of multiple sensors is equivalent.
information fusion as a synthesis of multi-sensor information The processing process is inherently complex. In the process of information fusion processing, according to the different original data processing methods, there are three main architectures of the information fusion system: centralized, distributed and hybrid.
1, centralized: centralized The raw data obtained by the sensor is directly sent to the central processor for fusion processing, which can realize real-time fusion. Its data processing accuracy is high and the algorithm is flexible. The disadvantage is that it requires high processor requirements, low reliability, and large data volume, so it is difficult to achieve .
2, distributed: each sensor pair The obtained raw data is first processed locally, including preprocessing, classification, and feature extraction of the raw data. Decisions are m& through respective decision criteria, and then the results are sent to the fusion center for fusion to obtain the final decision. Distributed has low demand for communication bandwidth, fast calculation speed, good reliability and continuity, but tracking accuracy is not as high as centralized.
3. Mixed type: in most cases The above two are combined in different ways to form a hybrid structure. It retains the advantages of the above two types of systems, but it has to pay a relatively expensive price in communication and calculation. However, this type of system also has the advantages that the above two types of systems are difficult to compare, and this type of structure is often used in practical situations.
Recently, research on multi-functional integrated sensors has received With extensive attention at home and abroad, integrated sensors in chemical analysis and biological protection have come out one after another. With the development of MEMS technology, people can make several sensitive components on the same material to make an integrated multifunctional sensor that can detect multiple parameters. Multifunctional sensors mainly have several different working principles and structural forms. Among them, due to the high integration and small size, the working atmosphere of each sensitive component is the same, and it is easy to realize the advantages of compensation and correction. Several different sensitive components Manufactured on the same silicon wafer and m& into integrated multifunctional sensors will be a direction for the development of multifunctional sensors.
In addition, the introduction of middleware technology is also more A common measure of sensor integration. Because the perception data structure provided by different types of sensors is not consistent, in order to achieve access to sensors of different types and different manufacturers, the system must solve the problem of access and processing of heterogeneous perception data, as well as the conversion and conversion of interaction protocols between various layers. Analyze the problem, and middleware is a kind of intermediate program that realizes the data transmission, filtering, data conversion and the conversion and analysis of the interactive protocol between the underlying hardware device and the application system.
middleware technology and system application are compliant The current development situation of the security market. The existing security system is mainly composed of the following characteristics:
①The security system software is mainly dedicated C/S architecture software provided by security equipment manufacturers. Under normal circumstances, this kind of software can only identify one or several products of the equipment manufacturer, and cannot be compatible with other manufacturers' products. Its internal communication protocol and information format are completely closed, resulting in poor scalability. And because of the disadvantages of C/S architecture software itself, such as installation before use and inability to unify upgr&s, once the equipment provider has a new product on the market, the original software cannot be compatible with the product. Upgr&, and then let all users reinstall it again, which greatly reduces the user experience. At the same time, for manufacturers, the cost of development, maintenance, and upgr&s are greatly increased, and the compatibility and scalability of the system are poor.
②Some security systems are only for a certain one A single system of type security equipment, or integrated multiple subsystems, but each subsystem works independently. If each subsystem needs to be monitored separately, it will consume a lot of manpower and material resources; and the same event causes independent alarms of each system, which causes the monitoring personnel to think that there are two unrelated alarms, which will also bring a lot of inconvenience to the monitoring personnel. Cause a waste of human resources.
And our actual project requirements are: Yes Access to different types of sensor equipment from different manufacturers, including infrared sensors, temperature sensors and cameras; good user experience, no need to install the software before using it, as long as you can view the real-time through a browser in a networked environment The alarm information; maintainability is good, if the system is updated and upgr&d, only the code on the server side needs to be updated, instead of upgrading all the client software; forming a whole security system, rather than each subsystem Single work; different types of sensors can coordinate alarms to provide more comprehensive alarm information; high intelligence, monitoring personnel no longer need to monitor the video wall 7 * 24 hours, once an abnormal situation occurs, the system will It automatically analyzes and informs the monitoring personnel in the form of images, sounds, and text messages. At the same time, the alarm period is recorded, so that there is no need to look for abnormal images from a large number of surveillance videos afterwards.
