Intelligent Video Analytics Technology Active Market

The current highway safety construction is emerging a variety of intelligent analysis technologies, such as new license plate recognition, vehicle trajectory tracking, flow detection and analysis and other new technologies. Through these new technologies, it is possible to expand the ability to capture speeding vehicles along the route, measure vehicle speeds and traffic statistics in specific areas, and realize the function of illegally occupying car lanes in large vehicles; it can realize the illegal use of emergency lanes, random parking, and retrograde capture of vehicles. Function; It can realize the detection of key entrances and exits of highways, detection of illegal passengers and passengers, and capture video; it can capture video for accident detection. With the development, the application of video surveillance in highways, high-speed railways, etc. has become more and more mature.

For high-speed railways, the main monitoring content of their safety monitoring can be broadly divided into 5 categories according to the target: natural disaster monitoring (such as earthquakes, rainfall and flood, wind speed and wind direction, etc.), line monitoring (such as shaft temperature, roadbed disasters Etc.) Monitoring of large-scale structures (such as monitoring of natural conditions in and around the station buildings, tunnels, traction substations, and communication signals in machinery rooms, etc.), object intrusion monitoring, train operation status, and car monitoring. Traditional high-speed railway integrated safety and disaster prevention systems are based on various sensors. However, with the continuous development and progress of video surveillance and analysis technologies, video-based integrated monitoring systems have received more attention and applications.

The front-end video surveillance point of the active infrared night vision technology integrated safety and disaster prevention system is usually in the outdoor environment where the natural environment is relatively harsh, and the target needs to be monitored all-weather at the same time. There is no street light, building lighting, and other ambient light sources available at night. For rainy days, foggy days and other special weather conditions required to maintain a good visual distance, the best visual distance can be about 1-2 km. Currently there are three types of night vision technologies available on the market: passive infrared imaging technology, active infrared night vision technology based on infrared light, active infrared night vision technology based on lasers.

Passive infrared imaging technology is suitable for monitoring nighttime people or vehicles and other moving objects. It is incapable of monitoring the environment. The active infrared night vision technology based on infrared lamps has a visible distance (within 300 meters). The light source has a large volume and a short lifetime. Adapt to the harsh natural environment outdoors. Laser-based active infrared night vision technology is currently relatively mature, including infrared laser luminaires, ultra-low light color black camera, infrared night vision surveillance camera three major parts.

The characteristics of the laser determine the laser can take a longer distance, and the light intensity is much stronger than the conventional light source. The laser has the advantages of high brightness, good monochromaticity, good directivity, etc. The selection of the wavelength can increase the CCD's sensitivity to the wavelength. Laser lamps have a long life span, which is generally more than 10,000 hours. The choice of camera is mainly to take into account two aspects: one is the imaging effect during the day, and the other is the low illumination of the infrared sensor. The daytime requirement is good, that is, it requires a filter switch structure to filter out invisible light during the day to make the image color more realistic. At night, it switches to night vision mode and puts invisible light in. This can not only guarantee daytime color, but also Guarantee night vision effect.

The night-vision lens F-number is generally smaller than the conventional lens F-number, which means that the lens into the light than the conventional lens is larger, there is a strong light collection ability, good night vision lens lens to near-infrared light The antireflection was performed to increase the transmittance of near-infrared light. In order to clearly see the distant objects, electric telephoto lenses are generally used.

Intelligent video analysis technology Intelligent video technology originates from the research of computer vision and artificial intelligence. Its development goal is to establish a mapping relationship between images and event descriptions, so that the computer distinguishes and identifies key target objects from numerous video images. This research is applied to video surveillance systems and will be able to use computer's powerful data processing capabilities to filter out unwanted or interfering information in images, automatically analyze and extract key useful information from video sources, so that cameras in traditional surveillance systems are not only Being human eyes also makes "intelligent video analytics" computers become human brains, and has a more "smart" way of thinking about learning.

This fundamental change can greatly expand and expand the role and capabilities of the video surveillance system, make the surveillance system more intelligent, significantly reduce resource and staffing, and at the same time, it will surely increase the efficiency of security work. . Intelligent video analysis can play a number of positive roles in fire alarms, foreign object intrusion alarms, station and bridge monitoring.

This type of technology, as an important means of intelligent road monitoring, is also gaining more and more attention from the industry and is also being applied to practical projects. For now, the road video intelligent analysis system architecture mainly has three ways:

Intelligent Analysis Camera: This is currently the main implementation method, that is, the intelligent video analysis algorithm is embedded in the front-end professional design of the camera, when there is a driving vehicle trigger camera capture, the camera immediately through the embedded algorithm for video information retrieval and analysis. Then send the picture and analysis results to the back-end linkage processing center. This type of front-end identification method has two major advantages: First, it has high processing efficiency, and second, it only outputs the analysis result, and its transmission of information traffic and bandwidth occupancy rate is low. Of course, you can also output dual streams according to your needs.

Stand-alone video analytics device: This is a stand-alone device dedicated to road intelligent video analytics. It incorporates intelligent analysis algorithms that identify vehicle and traffic violation information from the video stream provided by the front-end camera and send the analysis results to the background . This kind of analysis equipment is generally installed at a position close to the front-end camera. Its powerful intelligent video analysis function can even support 4 channels of 1080P HD video analysis, and can support the output of pictures and video streams. This stand-alone device method has two major advantages. First, the analysis performance is powerful. Second, it can be combined with a variety of high-definition cameras.

Background processing method: Nowadays, many traffic system platforms generally have developed road information intelligence detection modules, which will send the video stream of front-end cameras to the back-end and analyze them by the back-end platform. The advantage of this mode is its strong concentration, and it also makes use of the powerful intelligent video analytics performance of the back-end platform hardware, and it also has more diverse applications for identification. Of course, there are certain risks, such as very high demands on network bandwidth. When a large number of front-end cameras need synchronous analysis, even if the platform has load balancing functions, the front-end cameras still have a large amount of information, which will still affect the processing performance and network bandwidth of the system.

In short, the use of intelligent analysis functions is filtered on the premise of clear image recognition. Whether it is for intelligent front-end analysis mode or back-end platform analysis and processing, high-definition picture quality is an important factor in improving the accuracy of intelligent analysis. It can be seen that high-definition image quality is not only satisfied with a single visual effect, but also provides a more useful analysis environment for intelligent analysis.

Integrated monitoring host The traditional matrix monitoring system has a separate alarm host, but usually can only access the switch alarm, while the existing network video encoder has fewer alarm access, can not meet a variety of monitoring systems: strong wind, Rainfall, earthquakes, and various data protocols: 485 bus, CANBUS, and other access requirements, do nothing for multiple analog and digital acquisitions. The use of ARM9-based embedded systems provides extremely high system stability. Multi-channel card design, through a variety of cards can support multiple access control, RS-485 bus or CANBUS bus, digital input and output, RS-232 interface on a host.

Conclusion Video surveillance is an important part of the security system, including front-end cameras, transmission cables, and video surveillance platforms. Video surveillance is widely used in many occasions for its intuitive, accurate, timely and rich information content. In recent years, with the rapid development of computers, networks, and image processing and transmission technologies, video surveillance technology has also made considerable progress. It is believed that under the steady development of video surveillance technology, our traffic safety will receive more and more guarantees.