Apr. 14, 2025
In the lifting operation scene of lifting machinery, there are many high-risk areas, such as under the boom, under the hoisted object, in front of the hoisted object, in the triangle area of the guide pulley steel rope, around the fast rope, and in the force direction of the inclined hook or guide pulley. Once a danger occurs, it is difficult for personnel to avoid it. The main types of danger are as follows:
Danger 1: If employees directly support the shaking objects with their hands during hoisting operations, it is very easy to cause dangerous accidents.
Danger 2: Hidden dangers of binding objects: If the binding of objects does not meet the specifications, it may cause the objects to fall off, which may cause danger.
Danger 3: If the rope breaks during the hoisting process, the people below may be injured or even lose their lives.
Danger 4: When the crane hook is tilted and tilted, the people below are at risk of being crushed to death.
These safety hazards not only threaten the lives of workers, but may also cause business shutdowns, high compensation and other losses. Therefore, it is urgent to effectively control the dangerous areas of lifting equipment.
Solution
The AI anti-intrusion system for dangerous areas of Weite lifting equipment is based on the independently developed visual AI algorithm. It accurately identifies personnel behavior, environmental risks and security risks and triggers early warnings through real-time monitoring and analysis of front-end video data. The system can meet the safety management needs of different construction site application scenarios, intelligently detect violations such as working without a helmet and intrusion into dangerous areas, automatically archive early warning screenshots and videos to generate safety reports, realize 24-hour intelligent duty, and effectively save labor costs.
1. System diagram
2. Video image acquisition and processing
The AI anti-intrusion system for dangerous areas of lifting equipment combines traditional target detection technology, target detection and recognition technology based on deep learning, false alarm removal technology and various other advanced technologies to carry out research. The technical solution is shown in the figure
For the video images in the monitoring area, the preprocessing task of the video images is completed through background modeling, grayscale transformation, smoothing filtering, image enhancement and edge detection. For the target and the operating background, based on the time series image, the hough transform and other methods are used to complete the identification and regional division of the operating area. The foreground image is extracted by the background subtraction method, and the abnormal target detection is preliminarily realized based on the multi-dimensional characteristics such as color features, spatial geometric features and texture features. Based on the deep learning SSD and YOLOV3 target detection algorithm, the learning and training of intrusion foreign objects are completed to realize the real-time monitoring of intrusion foreign objects and the identification of foreign objects.
According to the principle of binocular stereo imaging, the target stereo matching is completed through the target detection results of the video images of adjacent cameras to obtain the target disparity map; the three-dimensional modeling of the target is completed according to the target disparity map, and the distance between the intruding target and the suspended object is calculated according to the target three-dimensional coordinates; finally, according to the gantry crane operation safety regulations, it is judged whether the intruding foreign object has crossed the boundary to decide whether the incident should be warned.
3. System platform
The complete system platform is divided into three layers from the technical service level: support layer, platform layer, and application layer. The application layer consists of three parts: early warning disposal, business support, and business technology. It mainly completes the functions of regional management, unit management, equipment management, personnel management, early warning event management, safety monitoring management, foreign body intrusion warning, standard shutdown warning, tilted crane warning, and introduction and description of related functions; the platform layer consists of public services, data services, and integrated buses, and mainly completes the functions of map services, statistical services, data services, computing services, security affairs, authority management, database and software buses, and integrated buses based on standard protocols that are commonly associated with the environmental safety monitoring process of gantry cranes;The support layer mainly provides hardware support for security monitoring and early warning technology, including data acquisition technology, video processing and storage technology, server systems and various mobile terminals.
Industrial computer display interface in the driver's cab
When someone passes through the monitored dangerous area, the AI intelligent recognition system will accurately identify the person, send out an audible and visual warning signal, and stop the vehicle operation in time.
System advantages
1. Real-time monitoring and active warning
Through self-developed AI visual algorithms and high-definition cameras, the system can identify intrusions in real time and trigger sound and light alarms, voice expulsion or equipment restriction, achieving a response in seconds.
2. Intelligent analysis and low false alarm rate
The deep learning algorithm is more scientific and more accurate, can distinguish between normal operations and abnormal intrusions, adapt to complex environments such as shadows and low light, and reduce false alarms and missed alarms.
3. Data tracing and process optimization
The system automatically saves intrusion videos and pictures, supports post-event retrospective analysis, and provides a basis for security management improvements.
4. Remote management and multi-scenario adaptation
Through the cloud platform, managers can remotely monitor multiple work points and flexibly adjust the identification area and alarm threshold, which is suitable for scenes such as construction sites, factories, and transportation hubs.
5. Significant cost-effectiveness
Compared with traditional manual inspections, the AI system has low operation and maintenance costs after a one-time investment, and can reduce indirect losses such as downtime and compensation caused by accidents.
Project cases
The application of single-beam crane in the factory of a compressor manufacturing enterprise. As a supporting factory of Gree Group, the enterprise has a large production scale and a large number of workers in the factory area. During the interlaced operations on the ground, it is inevitable that they neglect to guard against the overhead lifting equipment. Therefore, the safety issue of crane operation should not be underestimated.
Once the system was installed and debugged, it was immediately recognized by customers, who praised Weite as a protection expert in the field of lifting safety. Only customer recognition can motivate us to keep moving forward. Weite has been committed to providing customers with valuable products and services. We hope to reduce the probability of crane safety accidents through our products and provide customers with more reliable safety management solutions.
The AI anti-intrusion technology in the dangerous area of lifting equipment effectively makes up for the shortcomings of traditional protection measures through intelligent and precise safety control, and is a necessary choice to reduce accident risks and improve work efficiency. Its core value lies in transforming passive response into active prevention, providing reliable protection for industrial safety.
Latest
HOT PRODUCTS
