Give the city a pair of 'intelligent eyes' and use AI to protect safety

It is the nature of the people to settle down and relocate. For a long time, cities have meant more than just a place for people to live, but they also carry the eager expectation for family and friends to reunite and live a stable life. This is why from the agricultural era to the industrial era and now to the digital era, humans have never stopped pursuing a more comfortable and safer urban environment.

Especially as the construction of smart cities in various countries accelerates, the new generation of infrastructure around the Internet of Things, AI, machine vision, and big data has quietly been integrated into every corner of people's lives. It has not only become a "smart city" that improves efficiency and comfort. "Butler" becomes a "safety guard", issuing "early warning" before danger strikes to protect the safety of people's lives and property.

Chuangtong Lianda: Edge intelligence ensures the safety of gas station operations

People have been learning how to protect themselves as they grow up. For example, they have been reminded to stay away from fire, water, sharp objects and other sources of danger since they were babies. However, despite this, there are still hidden dangers in people's daily lives. In many crisis scenarios, a little carelessness may cause irreparable harm. In particular, many special scenarios already have clearly stated safety regulations in the manual, but safety awareness and operational rationality vary from person to person, but little does one know that it may cause catastrophe and endanger the safety of others.

• Hidden dangers in gas station operations

Gas stations belong to the hazardous chemicals industry. An incompletely extinguished cigarette butt may cause a fire or explosion. The main reason is that fuel is volatile and will turn into small fuel molecules and mix into the air. Any open flame may become a ignition The cable ignites the fuel-gas mixture, causing an explosion.

Although warnings such as "No open flames" and "No making phone calls" have been deeply rooted in people's hearts, fire/explosion accidents at gas stations have never stopped. Especially for gas station staff, unlike the short stay of consumers, they are in a high-risk environment for a long time. In addition to adhering to the safety bottom line in behavior, how to eliminate hidden dangers as soon as possible and ensure the refueling and unloading operations? The safety of people and property is of paramount importance.

For a long time, although the gas station industry has clear safety operation requirements, there are still problems with chaotic management and personnel not performing according to standards. For example, the oil unloading operation process is not standardized, the oil unloading process is not started after waiting for the oil to be stable, the fire extinguisher is not placed correctly, the static electricity removal operation is not performed, there is a lack of on-site supervision during the oil unloading process, etc.

In recent years, the country has launched a series of policies to accelerate the monitoring of safety production risks of hazardous chemicals. Among them, Document No. 207 of the Shandong Provincial Department of Emergency Management in 2021 requires the promotion of the application of full-process information management and video surveillance systems for special operations. Intelligent video surveillance systems can be used to intelligently analyze, monitor and alarm abnormal conditions during the operation process. It is true that compared with human supervision, AI-based intelligent video analysis can not only save labor costs, but more importantly, it can improve real-time performance and accuracy, and turn post-event accountability into pre-warning risks.

• Chuangtong Lianda Intelligent Video Analysis Solution Strengthens Risk Management

As the world's leading provider of intelligent IoT products and solutions, Chongqing Chuangtong Lianda Intelligent Technology Co., Ltd. (hereinafter referred to as "Chuangtong Lianda")In view of the safety hazards in the gas station scene, around For safe operating behaviors in refueling and unloading operations, we provide an intelligent video analysis solution integrating algorithm and computing power, using edge computing and AI video analysis, based on the EB5 edge box with built-in Qualcomm QCS8250 IoT solution, equipped with refueling Station scene algorithm realizes real-time monitoring of gas station personnel's dangerous behaviors such as smoking/talking on the phone, gas station staff's dress code, and oil unloading work process. It realizes the upgrade of the intelligent safety monitoring and oil unloading process of gas stations from "manual supervision" to "Intelligent monitoring" strengthens gas station safety risk management and control to comprehensively ensure the safety of gas stations.

Specifically, Chuangtong Lianda provides users with an integrated solution of algorithm computing power platform, which collects on-site data through smart cameras and sensor equipment on the end side, and performs processing on the edge side based on edge computing boxes. Real-time analysis, centralized management and visualization of equipment, AI algorithms, early warnings, and on-site images in the cloud.

Among them, in terms of algorithms, this solution can realize smoke and flame detection, work clothes identification, personnel intrusion identification, off-duty and sleeping duty identification, smoking identification, phone call identification, motor vehicle/non-motor vehicle identification, and tank truck identification. , fire extinguisher placement identification, electrostatic clamp identification, oil unloading identification, personnel on duty monitoring.

• Qualcomm QCS8250 IoT solution helps intelligent video analysis

According to Zhang Shuan, general manager of Chuangtong Lianda Industry Product Department, Chuangtong Lianda gas station safety operation solution improves safety management efficiency by intelligentizing the process and reducing manual intervention, while also realizing dangerous Incidents are quickly sensed and responded to, and incident reporting forms a closed-loop processing process. On the other hand, it can also achieve multiple protocol video stream access, fully utilize the original intelligent system of the old gas station, reduce access costs, and provide a standard API interface that can be quickly integrated with third-party platforms.

It is worth mentioning that

the EB5 edge box with built-in Qualcomm QCS8250 IoT solution, as a high-performance AI edge smart product that supports 5G connection, has 15TOPS super AI computing power, supports up to 24 channels of full HD video parallel processing. The Qualcomm QCS8250 IoT solution not only supports Qualcomm Wi-Fi 6 solutions and 5G connections, but also achieves top performance through Qualcomm Kryo 585 CPU architecture, fifth-generation Qualcomm AI engine and powerful image signal processor (ISP) to support Up to 7 concurrent cameras, encoding capabilities up to 4K, 120 frames/second. This provides sufficient computing power support for Chuangtong Lianda’s intelligent video analysis solution.

Ajiaxi Huaqin: Helping effective supervision of construction work sites

Gas stations mainly need to prevent fires and explosions caused by illegal operations. However, construction sites with more complex environments and denser workers involve multiple systems of water, electricity, and heating at the same time, so it is even more necessary to strengthen safety risk management. However, the project site covers a wide area, the personnel and departments involved in the construction are complex, and the equipment used in on-site operations is numerous and is in charge of different systems and companies. It is difficult to clearly assign responsibilities whether it is equipment loss or major safety accidents. There is a lack of effective supervision means for construction personnel, management personnel, construction equipment, and the construction environment. It can be said that "people are difficult to manage, things are difficult to handle, and responsibilities are difficult to bear."

• Construction site management is difficult, safety production is the first priority

In order to solve the problem of difficult accountability for various emergencies and major accidents at the construction site, the management has installed cameras at key locations. However, the traditional video access and manual supervision methods are still commonly used, which cannot monitor the on-site situation in real time and cannot achieve "prevention beforehand and control during the incident" of security incidents

With the rapid development of artificial intelligence vision, in recent years, more and more video monitoring scenarios have begun to try to replace "manual prevention" with "technical prevention". With the support of powerful computing power and high-performance algorithms, it not only achieves all-weather real-time monitoring, but also improves the accuracy and timeliness of non-compliant operations, and can detect potential risks at the first time, reduce accident risks, and fulfill Responsibility for ensuring safe production

More importantly, construction work is only the first step in building or park operations. A complete video monitoring system is not only a "scout" for hidden dangers on the project site, but can also be upgraded to a "safety" after the park is completed. Guards" and continue to patrol the park for potential dangers. For example, many office parks today do not have traditional fences. Instead, they use green belts as boundaries. Although beautiful, it can easily cause security loopholes. Intelligent video monitoring systems can monitor in real time those who enter in violation of regulations. Capture and report to management.

• Edge computing AI vision works together to ensure security

In view of the safety protection needs during the construction phase and later operation and maintenance phase of the park, Huaqin Technology Co., Ltd. (hereinafter referred to as "Huaqin"), which is deeply involved in safety production and digital transformation of the park, has joined forces with Chengdu Ajiaxi Intelligent Technology Co., Ltd. (hereinafter referred to as "Ajiaxi") has created the Huaqin AI vision software platform based on the AidLux edge computing device with built-in Qualcomm QCS8250 IoT solution, and has assisted China Electronics System Engineering Second Construction Co., Ltd. in Wuxi Huaqin The second phase project of Qin R&D Center achieved more than 400 days of safe production.

The second phase of the Wuxi Huaqin R&D Center project has a planned construction area of ​​58,000 square meters and is expected to be completed in June 2024. After the project is fully completed, it will gather 5,000 mid-to-high-end R&D talents. In this project, Huaqin AI vision software platform carried out intelligent AI monitoring of construction specifications, safety risks, personnel behavior, and equipment. It can realize construction specification monitoring such as hard hat detection, reflective clothing detection, elevator people counting, open flame detection, smoke detection, etc. Safety risk monitoring such as inspection and perimeter protection, personnel behavior monitoring such as personnel entry and exit detection, personnel fall detection, personnel smoking detection, monitoring room departure detection, etc., vehicle detection, license plate recognition, traffic flow statistics and other equipment monitoring.

According to Nanyang , chief architect of Huaqin AI vision business, due to project cycle constraints, it takes two weeks to complete the project deployment and implementation. In order to ensure the timely implementation of the project, Huaqin cooperated with Ajiaxi and built the Huaqin AI vision software platform into Ajiaxi's Aidlux system to quickly and efficiently complete the adaptation of the Qualcomm QCS8250 IoT solution match. At the same time, the dozens of algorithm functions integrated in the AI ​​platform can also complete the adaptation and fusion of algorithms and chips on this basis, ensuring the accuracy of the algorithm while maximizing chip performance.

Based on this, Huaqin quickly developed algorithm applications based on actual application scenarios, and connected cameras to test the effectiveness of deployed safety production projects. For scenarios with poor test results, Huaqin can quickly collect on-site training data and conduct incremental training and iterative models to improve recognition results. According to Nanyang, the entire project from chip adaptation to effect tuning was completed within two weeks, greatly shortening the delivery cycle

After the application of the Huaqin AI vision software platform, it can detect, identify and analyze various on-site video images 24 hours a day, and provide real-time, visual and intelligent AI services to managers. Conduct real-time situation detection, promptly alert potential risks, eliminate safety hazards, and reduce the occurrence of safety accidents. Based on this platform, the safety of the construction site has been comprehensively improved by 80%. Take the scene of wearing safety helmets in the construction site as an example. In the previous week, an average of about 10 people were found not wearing safety helmets in accordance with safety regulations at the construction site. After adding an AI vision software platform for supervision and alarm notification, was triggered every week at the construction site The number of people calling the police has dropped to only 1 or 2 people, and there will no longer even be alarms for such violations.

• Qualcomm QCS8250 IoT solution helps Huaqin AI Vision Platform bloom in many places

It is worth mentioning that, thanks to the superposition of the Ajiaxi AidLux platform and Huaqin’s business strength, the Huaqin AI vision software platform can be

quickly reused in other scenarios, such as parks and communities Management etc. At present, it has been deployed and used in Huaqin's own campus, providing integrated intelligent services of video monitoring, analysis and identification, feedback and early warning. Real-time monitoring of fire escape occupancy/blockage, climbing behavior, density/distribution of people flow, smoke, open flames, special crowd control and other behavioral status, and real-time detection, identification and early warning of collected video streams.

After deploying the Huaqin AI vision software platform,

the security of Huaqin’s own campus has been comprehensively improved by 90%. Taking cross-border recognition as an example, when AI vision was not added before, it would happen every month There were several incidents of people crossing illegally. After adding the 7*24-hour electronic fence monitoring function, no one inside the park has ever climbed over the fence again.

According to

Nanyang, the chief architect of Huaqin AI vision business, based on the Ajiaxi AidLux platform, it is compatible with Qualcomm chip platform, which improves the underlying structural capabilities of Huaqin AI software platform and does not require special optimization. , you can maximize the adaptability of Huaqin’s self-developed algorithms and Qualcomm chip platform. However, Huaqin AI software platform has encountered some challenges in the process of adapting to Qualcomm QCS8250 IoT solution, but with the help of Ajiaxi’s cross-platform AidLux system based on ARM architecture, it has cleverly solved the underlying compatibility problem. It also gives full play to the super computing power of Qualcomm chip platform, which greatly improves Huaqin’s algorithm performance in this solution.

Chuangtong Lianda: Putting food safety on the table

"Food is food to safety first". Food safety is also crucial to human health. In recent years, with the improvement of people's living standards, rich and diversified diets have also increased food hygiene risks. Especially after more and more illegal operations in the catering industry have been exposed, consumers are extremely concerned about food safety. Because of this, the "Bright Kitchen and Bright Stove" project has been more widely promoted. More and more restaurants and canteens have begun to replace the partition walls of the kitchen with transparent glass, or place display screens in the dining area for real-time display. Live image of the cooking area.

• Visible food safety

China’s cuisine enjoys an international reputation, and whether it is a small restaurant or a high-end restaurant, there may be unexpected delicacies hidden in it. However, precisely because of the rapid development of China's catering industry, various restaurants have sprung up like mushrooms after a rain, resulting in uneven management levels. In particular, there are often illegal operations in the back kitchen because they are far away from the sight of consumers. Therefore, the emergence of "Bright Kitchen and Bright Stove" can be said to display every action of the kitchen to regulators and consumers in real time, strengthening the management and restraint of food safety

In fact, as early as 2014, the State Food and Drug Administration deployed various localities to carry out the work of brightening the kitchen and brightening the stove in the catering industry, and successively issued notices and requirements such as guidance on the brightening of the kitchen and the campus food safety protection action plan, and implemented " Smart management models such as "Internet Bright Kitchen and Bright Stove". The national policy clearly requires that "market supervision and education departments in various regions must continue to promote the quality and expansion of the 'Internet Bright Kitchen and Bright Stove' in off-campus catering units and school canteens, accelerate the empowerment of smart supervision, and improve the level of school food safety management."

The traditional "bright kitchen and bright stove" supervision method mainly relies on manual visual inspection and video surveillance, which is prone to missed inspections. Even if an AI review platform is adopted, there are problems with real-time performance, and it is difficult to prevent food safety problems in advance. However, as AI capabilities decline, the edge side begins to take on more complex tasks. Intelligent video analysis can now be performed in real-time within restaurants and provide timely warnings to prevent non-compliant operations, eliminating food safety hazards at their earliest stages

• Chuangtong Lianda’s “edge cloud” collaborates to protect video security

In view of the food safety hazards in the catering industry, chuangtonglianda uses edge computing and AI video analysis technology to build an "edge cloud" collaborative kitchen intelligent analysis system to empower kitchen smart supervision. By deploying the EB2 edge computing box with built-in Qualcomm QCS610 IoT solution and loading the kitchen scene algorithm, it upgraded the existing camera to conduct real-time dynamic monitoring and analysis of the kitchen and detect whether personnel are wearing Issues such as irregularities, whether behaviors violate regulations, and whether there are hidden health hazards will be automatically warned, reported and recorded to achieve a closed-loop business intelligence.

This solution effectively solves the pain points of traditional monitoring methods such as difficulty in supervision, low efficiency, lack of prior prevention mechanism, and difficulty in traceability, and builds an upgradable, easy-to-deploy, and easy-to-manage Mingchu Liangzao smart upgrade solution. Currently, It has provided services to many schools, corporate canteens, and chain catering companies.

Specifically, Chuangtong Lianda provides the algorithm computing platform integrated solution for the Mingchu Liangzao intelligent monitoring platform, accesses the kitchen video monitoring data, and conducts real-time analysis based on the edge-side edge computing box , the cloud provides centralized management and visualization of equipment, AI algorithms, early warnings, and monitoring, and supports docking with enterprise business platforms or government food safety supervision platforms.

• Performance increased by 50%, Qualcomm IoT solution escorts "bright kitchen and bright stove"

According to Zhang Shuan, general manager of Chuangtong Lianda Industry Product Department, Chuangtong Lianda’s bright kitchen and bright stove solution is compatible with the original video monitoring environment and supports multiple protocol access for cameras from major manufacturers. Old buildings can be reused and renovated, and new buildings can also be built. At the same time, the algorithm has been optimized for the complex environment of the kitchen to improve the accuracy of alarms and reduce false alarms. More importantly, this solution can realize 7*24 hours of real-time intelligent monitoring. If any abnormality occurs, it will automatically capture and report it, and notify the company management personnel. It can be said that this solution makes up for problems such as untimely personnel inspections and omissions, and truly realizes a closed loop of business intelligence. It is worth mentioning that this solution also supports docking with the "Internet Bright Kitchen and Bright Stove" business supervision platform to achieve unified data aggregation and integration, break information islands, and visualize supervision data.

As the edge-side core product of this solution, Chuangtong Lianda EB2 edge smart station has

1.7TOPS computing power, supports 6-channel FHD video decoding, and Gigabit Ethernet, 4G, Wi-Fi network, which can flexibly adapt to different network access requirements during device deployment. The QCS610 IoT solution is based on Qualcomm Kryo CPU, Qualcomm Adreno GPU and Qualcomm Hexagon DSP, as well as Qualcomm AI engine whose AI performance is 50% higher than the previous generation products. At the same time, its dual image signal processor (ISP) supports video shooting, integrated audio, and global satellite navigation system GLONASS, which can provide stable connection guarantee for more smart camera products.

Write at the end

Initially, for most people, smart cities give the impression of being a mirage far away from the real world, and it takes a lot of time, manpower, and financial resources to get a first glimpse of its secrets. Nowadays, under the joint guidance of policy and industry, China's smart city construction has been quite effective. Especially as the development of digital China accelerates, citizens are more able to feel the real "security" brought by smart cities.

The above is the detailed content of Give the city a pair of 'intelligent eyes' and use AI to protect safety. For more information, please follow other related articles on the PHP Chinese website!