We have entered the era of Industry 4.0, and the mobile edge computing laboratory has emerged as the times require. Combining edge computing and cloud intelligent services is the trend of every manufacturing company's update and iteration, and it is also the development direction strongly supported by the government. No matter how advanced the technical means are, the standardized management of employees and items is the most basic and fundamental. What solutions does the Mobile Edge Computing Lab have for intelligent video analysis in manufacturing enterprises? Let's take a look below.
Pain points in the management of the security department: In addition to the entrance and exit gates, there are also entry and exit of various workshops and office buildings in the factory area. There are still practical difficulties in identifying internal personnel, visitors and strangers in real time. After entering the factory area, it is difficult to grasp the activity area of various personnel in real time. For various incidents, the method of recalling video surveillance can only be used to deal with them afterwards. Moreover, the operating cost is limited, and it is impossible to invest a large amount of manpower to manage the personnel behavior norms of each key node and respond to security incidents in various areas of the factory in real time.
Cost pain points of the production department: the status of employees in production positions, dress codes such as whether to wear work clothes, behavior standards such as whether they are smoking and looking at mobile phones, monitoring equipment is introduced into production line defect detection, process optimization and other deeper artificial intelligence. The on-the-job movement trajectory of the staff, tool identification and placement are also important optimization factors to improve production efficiency. If these data can be collected and analyzed in near real time at a controllable cost, it will be of great help to improve product quality and work efficiency as a whole.
Service pain points of the logistics department: Enterprises hope that employees will take the factory area as their home, so they naturally need to provide various logistics services such as employee canteens, convenience stores, public health and other facilities. How to keep up with the overall pace of the enterprise, provide intelligent logistics services while providing intelligent production management, and provide employees with real and effective convenience is also a problem that needs to be solved.
Design
The solution aims to realize real-time monitoring of key areas of the factory, identification and early warning of personnel at entrances and exits, situational analysis of important activity places, and comprehensive and multi-point data statistics, helping to realize intelligent and convenient digital smart factories.
Scene selection:
Access at all levels: Connect the existing or newly added camera video stream to the mobile edge AI device, realize face capture and feature value extraction, realize stranger warning, blacklist control warning, and assist attendance;
Production workshop: Connect the existing or newly added camera video stream to the mobile edge AI device to realize employee on-the-job status, identification of clothing and behavioral norms, identification of objects and tools, and collection of operation trajectories;
Canteen back kitchen: Connect existing or newly added camera and video streams to mobile edge AI devices to realize dining population statistics, personnel identification, dining diversion, and back kitchen food hygiene specifications;
Computer room/warehouse: Connect existing or newly added camera video streams to mobile edge AI devices for non-staff identification and early warning to prevent unrelated personnel from intruding into key areas;
Core Scenario 1: Fast passage at factory entrances and exits to improve factory experience
Deploy face access control personnel channel system, face capture and recognition system and other systems at the entrance and exit of the factory to realize the identification of factory personnel and prevent strange and suspicious personnel from intruding into the factory, posing personal and property threats to factory personnel. Personnel can also do quick detection and language reminders.
Core Scenario 2: Recognition of dress and behavioral norms of employees in the production workshop, collection of work trajectories
Using mobile edge AI devices, situation awareness, real-time monitoring of the working status of employees in the production workshop, standardizing the operation of workers, and realizing personnel control without going to the production site in person, improving work efficiency
value of customer:
Real-time detection: Real-time detection and alarm can be realized for whether the employees are in the production position, the movement trajectory required in the operation, the dress code and behavior code in the workshop, etc. At the same time, customized identification is also performed on the movement and placement of production tools. In addition, you can monitor the production situation of the factory anytime, anywhere, and assist in commanding and dispatching resources in the production workshop through the online inspection fusion video on the WeChat mobile terminal, which is easy to use.
Situation monitoring: The formed event log is kept on file, which can be used for further time series and semantic correlation analysis in the future. After an abnormal situation occurs, accurately restore the on-site situation in high-definition, and assist in troubleshooting the cause of the incident.
Device-side decoding: Through the use of edge computing boxes, only alarm information is uploaded and saved, reducing bandwidth and saving storage.
Core scenario 3: counting the number of diners at the peak of the cafeteria, helping the fine dining operation of the cafeteria
Real-time identification of the number of employees dining in the cafeteria, crowd density and other data, feedback to the logistics service department and disclosure to employees. It helps the canteen to accurately formulate food procurement data, optimize and adjust the staff scheduling of the canteen, and can also realize the diversion of meals to avoid crowding during peak dining periods.
value of customer:
Stranger alarm: Stranger alarm can be set in the kitchen and the kitchen door to prevent malicious intrusion into the kitchen.
Blacklist alarm: control blacklist personnel to realize illegal personnel alarm.
Number of people counting: counting the number of people eating can help the canteen to accurately formulate food procurement data and optimize the adjustment of canteen staff scheduling.
Dining diversion: By counting the number of diners and exceeding the set threshold alarm, smooth diversion, avoiding crowding during peak dining periods.
Core Scenario 4: 24/7 guarding of key areas in the computer room/warehouse, "rest assured" factory
The computer room and monitoring room are the key places for the safe operation of the enterprise park, and the protection around the computer room and monitoring room is the guarantee for the safe operation of the park. The intelligent face recognition tube machine can be used to alert suspicious persons approaching the monitoring room and remind security personnel to take relevant precautions.
value of customer:
Stranger recognition: Through the intelligent face recognition machine, timely identify and alarm non-staff personnel to prevent illegal intrusion into key areas.
Abnormal behavior analysis: By setting the time period, the fast moving, wandering and crossing the line in the sensitive time period are detected, and the alarm is given in time.
Blacklist alarm: use intelligent face recognition machine to realize deployment and identification of blacklist personnel.
Uncivilized phenomena are reminded in time to create a high-quality factory environment:
The solution achieves high-definition image collection through the front-end camera, and realizes accurate identification of people through deep learning algorithms. The detection type can be extended to vehicles, non-motor vehicles, etc. Connect the sound column through the alarm output signal, trigger the built-in voice, and transmit it to the back-end video recorder through the mobile edge AI device, realize the functions of end-side alarm pop-up window, quick review and supervision, and also realize the mobile edge AI device through the microphone Shout out to the front camera.