As an important pillar of modern transportation, the safe and stable operation of EMUs and high-speed trains depends on an efficient monitoring system. The visual monitoring system is the visual perception center of high-speed rail safety management. Dark light full-color technology in the rail transit visual monitoring system, the core role is to break through the night, low light environment imaging restrictions, to solve the traditional infrared, black and white monitoring "unclear, unclear, low data value" pain points, covering the train on-board monitoring, rail monitoring, station monitoring and other three core areas.

It includes the following subsystems and key equipment:

1. vehicle visual monitoring: to ensure night driving safety, improve fault identification accuracy

1. Forward driving assistance and obstacle recognition. When high-speed trains and high-speed trains are running at night, traditional low-illumination cameras are prone to the problems of much noise and low contrast, making it difficult to identify hidden dangers such as falling rocks, pedestrians and foreign bodies in front of the track. The dark full-color camera can restore the true color and outline of obstacles in extremely low-light environments such as moonlight and starlight, assist the driver or the automatic driving system to warn in time, and greatly improve the safety of driving at night.

2. There is no dead corner for night security in the carriage. The cabin is dimly lit at night, and the traditional monitoring screen is blurred, so it is impossible to distinguish the specific characteristics of passenger abnormal behavior (such as theft, fighting) or left behind items. The dark full-color technology can clearly capture the details of the personnel's clothing and items in the carriage, support intelligent behavior analysis, and realize the rapid identification and alarm of night emergencies.

2. track along the monitoring: to achieve night all-weather inspection, reduce labor costs

1. Foreign body invasion and perimeter security accurate warning. Key areas such as bridges, tunnel openings, and fences along the track are high-incidence periods for foreign matter intrusion at night. Traditional infrared monitoring can only find "heat source targets" and cannot distinguish pedestrians, animals or sundries. Dark-light full-color technology can directly restore the shape and color characteristics of the intrusion target, accurately judge whether it is dangerous foreign matter (such as large vehicles and construction equipment), avoid false alarm, and at the same time, there is no need to install supplementary light to ensure the concealment of monitoring.

2. Automatic inspection of track and catenary at night. Manual night inspection track, contact network low efficiency, high risk, the lack of details of the data collected by traditional monitoring equipment. The inspection unmanned aerial vehicle or rail side camera equipped with dark full-color load can carry out high-definition imaging of rail cracks, loose fasteners, deviation of catenary hanging strings and other problems at night. The color picture can more intuitively present the defect location and degree. The automatic defect classification can be realized by combining AI algorithm, replacing manual night inspection, and reducing labor cost and operation risk.

3. Bridge tunnel structure disease night monitoring. Corrosion of bridge steel structure and leakage of tunnel lining are difficult to be identified by traditional monitoring in low light environment at night. The dark full-color camera can clearly capture the color difference and crack details of the structure surface. Through long-term night data comparison, it can accurately evaluate the development trend of the disease and provide a reliable basis for maintenance and repair.

Product advantages:

1, improve the overall data value of the system, enabling intelligent applications.

Compared with traditional visible light cameras, dark-light full-color cameras can be tuned according to different scenes of rail transit, and their color reproduction, high dynamic range, high noise reduction and other characteristics in complex environments are easier to be analyzed and trained by AI algorithms. Whether it is on-board fault identification, foreign body detection along the line, or station behavior analysis, color data can greatly improve the identification accuracy and efficiency of the algorithm, realize the closed-loop management of "data collection-intelligent research and judgment-automatic early warning", and promote the upgrading of rail transit visual monitoring system from "passive monitoring" to "active early warning.

2. High reliability and stability of products

The product has military-grade quality, can withstand extreme high and low temperature (-30 ℃ ~ +85 ℃) harsh environment, can quickly restart the recovery work after power failure, with strong stability and reliability to ensure all-weather uninterrupted monitoring.

3, with accurate, real-time, low-cost perception ability, improve the boundary of assisted driving.

Through binocular depth camera, the three-dimensional depth information of the front track and the surrounding environment can be captured in real time under high-speed driving scenes, and the double-camera stereo matching algorithm can be matched to accurately identify hidden dangers such as foreign objects, obstacles, line wear, turnout deviation, etc. on the track. At the same time, the real-time distance and relative speed between obstacles and trains can be measured, and millisecond-level early warning can be realized in combination with AI target detection algorithm. Effectively improve the safety and intelligent level.