The same as the Internet of Things, video surveillance also needs key technology breakthrough to finish transformation and elevate to another phase.
To achieve key technology breakthrough, the right technical direction and appropriate start point should be chosen.

Automatic interpretation of image is the basic symbol of video surveillance system intellectualization. However it took us a long time to explore and gradually make clear of its real meaning, fundamental function and realization approach. It is the most urgent requirement of security system to identify the individual through images. Recently many achievements about image identification has been realized and it has been brought into preliminary application in various fields ,especially in security field. There are two main modes:

    ·Verification. Identifying the surveilling objective and legality of the happening activity is the fundamental function and typical application of  security system. 

    ·Identification. Comparing input features with a large number of references in database, identifying the objective and discovering relevant clues of various events are the typical applications of reconnaissance system. However such kind of system needs to set up a huge number of sample databases.

The above two modes and applications all require system to define characteristic firstly and ensure correlation between primary inputs and spot inputs, that is to say, to establish a little bit limited environment. However, common video surveillance system can not meet such requirements. A new technical direction and application start point must be found. Image content analysis is the most appropriate choice.

Image content analysis has automatically defining (simple) characteristic and no comparison with primary inputs, but achieved by image correlations through objective identification, behavior analysis and prediction, etc. The common video surveillance system is its foundation platform, which is the direction and realization approach of surveillance system intellectualization.
Image content analysis research can be divided into two levels namely video detection and video semantic interpretation.

An initial probe can separate (moving) objectives from video images and classify objectives under a simple environment (simple objective and pure background ).Behavior analysis can judge its moving direction, mode, objective recombination and dissociation. On the ground of analysis, it is easy to judge whether the status is normal, discover abnormal behavior and give an alarm. Then the objective trajectory comes into being. It can track the objective automatically and conduct objective statistics, relevance, filtration and trend prediction, etc. Main differences between the video detection and general intrusion detection is its focusing point. At present some products are sold in the market. But application under complex environment,(that is, under common video surveillance environment), needs key technology breakthrough.

Video semantic interpretation is to understand the real information contained in image sequences by analyzing them. Video semantic interpretation is the highest technical level. It shows that machines have the ability to understand the images just as humans and it has the efficiency that humans can not reach. Although we haven’t reached such a degree so far, image information index from content analysis can be achieved, so do the fast and accurate retrieval of image information. Then achieving mutual relevance of image information, images and sounds, graphics and text messages. This is the premise of image information deepening application. It may be called video half-semantic interpretation.
The Application of Surveillance Platform:

Image content analysis is different from image identification. It is the real meaning and fundamental function of video surveillance system intellectualization.
Image content analysis not only disposes single frame image(objective separation),but also deals with image sequences(behavior analysis ,etc),not only relates every single image from image source camera, but also relates multisource images, It can be processed front-end, but also needs system (background) processing. Therefore, distributed enhancement processing structure must be used. The concept of cloud computing should be applied and become core technology of the internet . 

Video surveillance system intellectualization is a process of gradual development. It cannot be accomplished in an action and has no ultimate result. Constant technology accumulation is needed, especially core technology breakthrough. Image content analysis is an important direction.

To achieve innovation of system model and the breakthrough of key system technology, firstly, the system design ideas should be transformed from designing system under specified resources and programming image quality to adapt to the system environment into the needed resources in programming and configurating the system according to the requirement of image quality. To finish this transformation, the innovations in system control equipments and image transmission technologies are needed, thus creating a new system to control theory and central equipment.
To conform to the trend of the current development of science and technology, integration is the trend, information technology is the arrow and triple play service is the model. Security technology should keep pace with and integrate with information technology, and form a unified whole with information security, open service system and intelligent building, etc, then become the original part of these technical systems. To achieve the own integration of security technology, video surveillance technology is the key of integration.

To achieve the specialized service of general products, the transformation from specialized products into specialized service of general product is the application technology development trend. Intelligent building system and characteristic identification technology are typical examples. The core of application system would transform from hardware into software(products softening). Middle ware plays an important role in the process.

The basic concepts of middleware can be intuitively interpreted as the common services between platforms(hardware and operating system) and applications. It is an essential application supporting platform which can solve many problems like inter operating between systems, data exchanging, relevant distributivity, reliability and security, etc. It can also achieve relevance, integration, collaboration, interaction and service on demand. At present surveillance management software, system surveillance platform, soft answering alarm platform and application server all have middleware products.

Establishing the open service system(platform) is a development tendency of public affairs management system. Video surveillance system should become a new service system supported by the internet and cloud technologies.

The key to achieve the above technical breakthrough is software. It is the weakest part in China's security industry, also the highest part of technology added value. We should conform to this tendency, promote adjustment of industrial structure, change the phenomenon and condition from hardware first software second to hardware better to software weaker ,improve industry system integration ability and the capability of providing integrated solutions. 

Only the above breakthrough has been accomplished, can the advantages and characteristics of digit, intelligence and high definition, etc, be adequately showed and widely used.