(a) big data is the basis of video intelligent analysis

In the era of big data applications, video because of its highest information content, the largest amount of data, analysis and computing the most complex and has become the era of big data collection, analysis, transmission, storage and application of the most challenging international technical problems! Intelligent video analytics research never ends, analysis algorithms must be monitoring video as a resource, research real-time or historical surveillance video target feature extraction, enhancement and behavioral analysis and other key technologies, in order to promote the monitoring video application mode from the after the passive disposal to the active prevention of the change beforehand.

(II) Help realize smart city intelligence

One of the major challenges facing the construction of China's smart city is the inability to effectively integrate between urban systems due to standard issues, forming information islands. Therefore, in the field of big data fusion technology, on the one hand, it is necessary to strengthen the construction of big data standards, and on the other hand, it is necessary to strengthen the research and development of key technologies such as massive heterogeneous data modeling and fusion, massive heterogeneous data columns storage and indexing, in order to give the underlying data integration of information **** enjoyment to provide standards and technical protection. Large-scale data needs to be pre-processed during the flow of the smart city system for factors such as transmission efficiency, data quality and security. Big data processing technology often needs to be combined with parallel distributed technology based on cloud computing, which is also the technical solution commonly adopted by the international industry. Big data analysis and mining technology provides powerful decision support capabilities for smart city governance.

(C) Improve police efficiency

The rapid development of Internet technology has provided strong technical support for the construction of a large-scale nationwide professional alarm operation service platform. Through this alarm platform, alarm operation service providers will accumulate massive amounts of user data in their hands, such as user identity information, police data, consumption records, maintenance records and so on, which are all very valuable resources. Alarm operation service providers can, on this basis, apply big data technology to analyze and mine, giving full play to the commercial value of big data.

Public security, such as public security system in the map detection technology, application mode is diverse, active thinking, around the "discovery of clues" can be derived from the purpose of a variety of techniques and tactics, only from these specific techniques and tactics can be extracted from the needs of the real system to tell the designers, "What do we want? It is only from these specific techniques that the requirements can be extracted to truly tell the system designers "what we want".

So, what are the big data applications needed in TuDetect? Like commercial big data to find the application of the law seems to be far away, the most real is from the massive video data to the same clue characteristics of the image to find out, so that the police found a new case clues. As for "how to find?" This is the public security to mention the application mode. Therefore, the development of video big data is not simply by the technology vendors do lead, but the need for the public security system both criminal investigation experience, but also have the science and technology of the composite talent, *** with to participate in the development of video big data applications.

(D) let the smart home "smart" up

Smart home will produce big data, but also an important application of big data, otherwise it is very likely to be stagnant. Big data generated by the family can make the smart home more "smart", but need to be effectively dealt with according to the actual situation, rather than any data "pot", through the combination of big data and cloud computing technology, the smart home system can be the first time on the user's smart devices in the home data, information and information. The data and information of intelligent devices in the family can be effectively analyzed and memorized, and the corresponding laws obtained can be applied to intelligent devices in turn, so as to enhance the intelligent effect of the smart home.

Two, security big data application difficulties

(a) data integration issues

Different sources of big data, respectively, are stored in each other independent systems, the data will be centralized in a unified platform, is the implementation of security big data fundamental work, but the industry, departmental barriers is the biggest obstacle. Even if only the video data within the public security, the provinces, municipalities are not connected to each other, want to collect centralized is not an easy thing. Even after centralization, how to find the relationship between these different types of data, so as to dig out valuable data, is also difficult.

(ii) the maturity of data mining, analysis algorithms

For the most important video data in the security data, its intelligent video analysis and mining is a very difficult thing. At present, in addition to license plate recognition, headcount and other algorithms are more mature, the video for event analysis, face recognition, summary and other technologies have not yet reached a large-scale commercial level, which also greatly restricts the implementation of security big data.

(C) the timeliness problem

One of the purposes of security big data is to solve the existing security system to view after the fact, analysis of data (especially video data) application form, but also to increase to the pre-warning, real-time processing, which requires a high degree of real-time big data processing technology. This timeliness determines the video security big data high computing power, high transmission bandwidth requirements.

(D) information security and user privacy issues

The security industry, especially the public security industry, the security requirements of the data is very high, which is an important reason for the regional isolation of data. At the same time, in the use of security big data, how to protect the privacy of users, is also a very important topic, the current main use of data desensitization approach. It is imperative that the security data security level needs to have a clear hierarchical definition, can not be emphasized security and their respective closure, otherwise it will certainly lead to security big data analysis to become a source of water.

(E) video image data mining difficulties

1. identify what features? An image or a video can have countless angles of the label attributes to describe, what is the attribute we need? This is closely related to the purpose we need to get, which requires public security figure investigation of the talent to summarize the end.

2. Recognition algorithm development is difficult, because it is a flat image, so the main principle of feature recognition is to look at the image area in the outline, color, texture and feature library for comparison. But in the same object in different monitoring angle of the camera shows the outline are not the same, so can not do to recognize.

3. Large-scale data processing is difficult, even if the recognition algorithm is achieved, but if large-scale video is structured in the form of data processing servers, the construction cost is huge, and its energy consumption is also impractical in China, where electricity is limited in summer.

(F) police service platform big data difficulties:

1. How to integrate the data between different alarm operation service providers together?

2. Most of China's alarm operation network has not yet completed the construction of scale, the user scale, cross-provincial and municipal operation of the network is very small, each alarm operation service provider's police concurrency is not large, and the alarm operation service providers generally exist between the information silos, it is difficult to achieve value-added data through big data analysis.

3. The mining of big data is a long-term process, which requires companies to constantly try to dig out meaningful information or laws, and take the results to the market to test.

4. Big data itself is also facing challenges, the use of data is still facing a variety of technical difficulties in the boundaries, the lack of talent in big data, big data products are not yet mature and other issues are restricting the development of big data in the field of alarm operation services.

Summary

Aiming at these problems and difficulties, individuals put forward their own insights on one aspect, the information collection and monitoring of big data. For the moment, big data and the Internet are an interrelated whole. Then, in terms of data mining, the forum, bar, microblogging, WeChat information collection becomes very necessary. After data mining, the data should be screened and processed. At this point, the monitoring of information comes into play. At present, the information collection and information monitoring can be combined and applied to the actual enterprise is not much, you can pay attention to this, two words, happy "happy", thinking "thinking", in this regard has a certain amount of precipitation and strength. Big data is a new industry. Therefore, to find a certain technology, in order to be applied to the field of security, and produce the desired effect.