System introduction
By using depth learning algorithm, the system can track pedestrian movement and extract pedestrian features, and identify the external attributes such as hairstyle, age, sex, race, whether wearing a hat, knapsack, jacket type and color, trousers type and color in real time. After detecting pedestrian features, pedestrian features can be retrieved, that is, pedestrian features can be retrieved. Video retrieval of pedestrians in adjacent cameras is performed based on pedestrian pictures.
Depth neural network algorithm based on bone key points detection can also automatically identify 18 human postures, such as joints, five senses, etc. The human skeleton information is described by key points to judge the type of action.
Statistical identification and analysis of people, mainly to achieve statistical analysis of people, including the number of statistics, population density, human activity, etc.
Human behavior anomaly analysis, mainly to achieve intrusion detection, abnormal running, illegal trailing, on-the-job and other abnormal behavior analysis.
Normalization analysis of human posture,The standardized analysis of human body posture mainly detects some specified normative actions or dress requirements, and extracts and alarms those who do not conform to the norms.
The system can be applied to many scenarios, including intelligent transportation, rail transit, intelligent commerce, safe production, film and television entertainment, judicial analysis, tourist attractions, video coding, intelligent live broadcast and other fields.
Intelligent Transportation
Based on human body recognition, it can be widely used in intelligent transportation, such as pedestrian warning at intersections, dynamic warning when pedestrians are detected at zebra crossings to reduce traffic accidents, pedestrian running red light snapping, dynamic control of pedestrian crossings, pedestrian on-highway and urban expressway main road pre-planning. two passengers and one dangerous driver's fatigue driving and identification of illegal actions. Pedestrian detection is also a very important function in the coming era of autopilot.
Rail transit
The main function of rail transit is to transmit passengers, so the demand for pedestrian detection is the largest, including the number of stations and corridors statistics, crowd density analysis, abnormal running events, intrusion tripping detection, crowd gathering detection and so on. In addition, it can also automatically detect the elderly and children in the carriage, and remind them to give up seats.
Security monitoring
Security surveillance is mainly used for the structuralization of surveillance video, which is easy to retrieve and control personnel. At the same time, the system can also be used to detect intrusion trips, abnormal running, crowd density analysis and crowd gathering warning.
Wisdom for the aged
In the intelligent environment, the most important thing is to pay attention to the safety of the elderly, so we can use this system to realize the fall detection of the elderly, the elderly activity analysis (and then analyze the mental state of the elderly and the trend of physical changes).
Wisdom class
It can be used to analyze the state of everyone in kindergartens, primary and secondary schools, etc.
The human posture recognition based on depth learning is used to recognize the action of the people in the video, and the activity of each individual is evaluated. The warning is given to those who are below or above the activity threshold.
Tourist attractions
Tourist attractions pay special attention to the daily number of tourists, as well as the density of tourists in key areas, to prevent stampede incidents. Through this system, the number of entrances and exits can be counted, the population density analysis of key areas and the abnormal warning of population density can be achieved, and reliable data and risks can be provided for tourist attractions. Remind.
Intelligent business
In smart business, customers are always the center, so we can realize customer flow statistics, customer trajectory analysis, customer focus thermodynamic analysis based on this system, through these optimized commodity visits, auxiliary location, in addition to the realization of in-store service and cashier work status detection, including is No long time playing cell phone, off duty and so on.
Safety production
Safety in production wearing safety helmet, wearing work clothes is the requirement, so through this system to achieve staff wearing safety helmet detection, wearing work clothes detection. In addition, it can also achieve some staff's unusual ground detection, abnormal climbing detection and so on.
Physical training
By modeling standardized sports movements and comparing and analyzing athletes'real-time movements through video, standardized sports movements can be achieved. At the same time, it can also conduct statistical analysis of athletes' Sports volume and scientifically guide physical training teaching.