Construction site safety early warning system and method based on AI model identification

Abstract

The invention provides a construction site safety pre-warning system based on AI model identification. The construction site safety pre-warning system comprises a multi-source sensing layer, an edge calculation layer, a pre-warning decision module, a data middle platform and a report center, the invention discloses a construction site safety early warning method based on AI model identification. The method comprises the following steps that S1, a multi-source sensing layer collects field state data in real time; s2, the edge calculation layer receives and processes the field state data, and sends the field state data to an early warning system core and a command large screen for real-time feedback; s3, the face recognition unit dynamically monitors personnel access, updates on-site personnel information, and counts individual working hours according to the on-site personnel information for management personnel to check; s4, an early warning decision module performs intelligent analysis by using a configured early warning model according to the field state data; and S5, if the early warning rule is triggered, the system immediately displays early warning details on the commanding and dispatching large screen and gives out an alarm sound. According to the invention, full-process digital management of second-level intelligent sensing and risk early warning and disposal of potential safety hazards on a construction site is realized.

Description

Technical Field

[0001] This invention relates to the field of construction engineering safety supervision technology, and in particular to a construction site safety early warning system and method based on AI model recognition. Background Technology

[0002] In recent years, national regulations on work safety have become increasingly stringent (such as the revised Work Safety Law), placing higher demands on safety management at construction sites. While the traditional model relying on manual inspections by safety officers constitutes a core supervisory method, it suffers from the following significant drawbacks: First, insufficient timeliness: The discovery of hidden dangers depends on the frequency of inspections, with an average response cycle of more than 2 hours, which is difficult to meet the second-level early warning requirements of high-risk operation scenarios (such as deep foundation pits and tower crane groups); Second, the coverage is limited: manual inspections are concentrated in the main areas, and the missed inspection rate in key locations such as construction corners and high-altitude blind spots exceeds 30%. Third, costs continue to rise: a single large project requires 6 to 8 full-time safety officers, with labor costs increasing by an average of 12% to 15% annually.

[0003] Fourth, traditional video surveillance systems mainly have recording functions but lack the ability to automatically identify risks and handle them in a closed loop.

[0004] Breakthroughs in artificial intelligence and computer vision technologies, particularly in behavioral recognition (such as fall detection) and cross-modal data fusion (video + sensor), have provided a technological foundation for building an intelligent safety early warning system. Therefore, there is an urgent need for a systematic solution that integrates AI intelligent analysis, real-time early warning, and response tracking to address the shortcomings of existing management models. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing a construction site safety early warning system and method based on AI model recognition.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: A construction site safety early warning system based on AI model recognition includes: The multi-source sensing layer includes a video acquisition unit, an intelligent environmental monitoring unit, an intelligent tower crane monitoring unit, and a portable intelligent work card unit; The edge computing layer receives and initially processes real-time status data from the multi-source sensing layer. The early warning and decision-making module includes a risk classification engine, an intelligent push engine, and a hidden danger closed-loop handling subsystem. The data platform and reporting center integrate and analyze data and early warning information from various perception layers, providing: risk heat maps based on geographical location, risk profile analysis of different job types, trend analysis of risk levels monitored by various intelligent devices, and comprehensive assessment reports of overall project safety risk indicators.

[0007] Furthermore, the video acquisition unit is equipped with 4K / 8K high-definition cameras and mobile surveillance cameras to achieve comprehensive monitoring of the construction site; The intelligent environmental monitoring unit monitors the dynamic environmental parameters of the construction site around the clock, including temperature, humidity, wind speed, and concentration of toxic gases, to ensure that construction personnel work in a safe environment. The intelligent tower crane monitoring unit integrates the tower crane's own operating status data and surrounding real-time video to provide real-time early warning of abnormal tower crane conditions, and the data is synchronized to the system decision center. The portable smart work card unit integrates positioning and emergency rescue functions. It uploads personnel location and movement trajectory at regular intervals through multiple signal transmission methods such as 4G, Bluetooth, and UWB. When personnel cross the boundary into a dangerous area or send out a distress signal, the system automatically triggers an alarm and links with the command and dispatch screen to issue a warning or initiate rescue.

[0008] Furthermore, the tower crane's own operating status data includes tilt angle, load, and wind speed; the tower crane's abnormal status includes overload and collision risk.

[0009] Furthermore, the edge computing layer utilizes the Internet of Things platform through data preprocessing and transmission to process the perceived data according to a preset rule chain, and transmits the key data to the core of the early warning system and the command screen through the message middleware (MQTT). The edge computing layer includes a face recognition and attendance statistics unit, which dynamically monitors personnel entry and exit, displays real-time personnel changes at the construction site, and calculates individual working hours based on entry and exit records by week, month, year, and cumulative dimensions, and reports to management personnel. Early warning model configuration platform: Based on industry norms and national standards, it classifies the attributes of personnel and equipment at construction sites and establishes a standardized single indicator library; it supports projects to add personalized single or mixed indicators according to their own characteristics, so as to realize flexible configuration and compatibility of early warning rules for multiple scenarios, multiple environments and multiple devices.

[0010] Furthermore, the early warning decision-making module includes a risk classification engine, an intelligent push engine, and a hidden danger closed-loop handling subsystem.

[0011] Furthermore, the risk grading engine is based on standards such as the "Standard for Safety Inspection of Building Construction" JGJ59-2011, and constructs a risk quantification model. It automatically outputs four levels of warning signals: red, orange, yellow, and blue, for the identified hidden dangers. Among them, the red warning signal represents the highest level, the orange warning signal represents the high level, the yellow warning signal represents the medium level, and the blue warning signal represents the low level.

[0012] Furthermore, the intelligent push engine automatically selects and executes the optimal push strategy based on the warning level and specific scenario, with target terminals including WeChat mini-programs for administrators, on-site intelligent broadcasting systems, and command center dispatch screens.

[0013] The construction site safety early warning method based on AI intelligent analysis includes the following steps: S1: The multi-source sensing layer collects on-site status data in real time; S2: The edge computing layer receives and processes the field status data, and sends the field status data to the early warning system core and command screen for real-time feedback; S3: The facial recognition unit dynamically monitors personnel entry and exit, updates the information of personnel present, and calculates individual working hours (weekly / monthly / yearly / cumulative) for management personnel to view; S4: The early warning decision module performs intelligent analysis based on on-site status data and the configured early warning model (based on industry standards and project-specific indicators); S5: If the warning rule is triggered, the system will immediately display the warning details on the command and dispatch screen and issue an alarm sound.

[0014] Compared with the prior art, the beneficial effects of the present invention are: the speed of hazard identification of the present invention is improved from 2-4 hours in the traditional manual mode to the second level, and the hazard response time is shortened from more than two hours in the traditional mode to less than 30 minutes.

[0015] The system provided by this invention optimizes human resource costs, reducing the number of safety officers per project by 30%, saving on average annual salary and training expenses, while improving production efficiency and increasing the early warning rate for major accidents such as tower crane collisions to nearly 100%.

[0016] The system provided by this invention improves management efficiency by visualizing the risk values ​​of construction areas based on a hazard heat map, guiding the precise allocation of safety resources; and by using job risk profiles, it enables targeted safety training, effectively reducing training costs. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the construction site safety early warning system based on AI model recognition according to Embodiment 1 of the present invention; Figure 2 This is a flowchart of the construction site safety early warning method based on AI model recognition in Embodiment 1 of the present invention. Detailed Implementation

[0018] To provide a further understanding of the purpose, structure, features, and functions of the present invention, detailed descriptions are provided below with reference to specific embodiments. Example

[0019] A construction site safety early warning system based on AI model recognition includes: The multi-source sensing layer includes a video acquisition unit, an intelligent environmental monitoring unit, an intelligent tower crane monitoring unit, and a portable intelligent work card unit; The edge computing layer receives and initially processes real-time status data from the multi-source sensing layer. The early warning and decision-making module includes a risk classification engine, an intelligent push engine, and a hidden danger closed-loop handling subsystem. The data platform and reporting center integrate and analyze data and early warning information from various perception layers, providing: risk heat maps based on geographical location, risk profile analysis of different job types, trend analysis of risk levels monitored by various intelligent devices, and comprehensive assessment reports of overall project safety risk indicators.

[0020] Furthermore, the video acquisition unit is equipped with 4K / 8K high-definition cameras and mobile surveillance cameras to achieve comprehensive monitoring of the construction site; The intelligent environmental monitoring unit monitors the dynamic environmental parameters of the construction site around the clock, including temperature, humidity, wind speed, and concentration of toxic gases, to ensure that construction personnel work in a safe environment. The intelligent tower crane monitoring unit integrates the tower crane's own operating status data and surrounding real-time video to provide real-time early warning of abnormal tower crane conditions, and the data is synchronized to the system decision center. The portable smart work card unit integrates positioning and emergency rescue functions. It uploads personnel location and movement trajectory at regular intervals through multiple signal transmission methods such as 4G, Bluetooth, and UWB. When personnel cross the boundary into a dangerous area or send out a distress signal, the system automatically triggers an alarm and links with the command and dispatch screen to issue a warning or initiate rescue.

[0021] Furthermore, the tower crane's own operating status data includes tilt angle, load, and wind speed; the tower crane's abnormal status includes overload and collision risk.

[0022] Furthermore, the edge computing layer utilizes the Internet of Things platform through data preprocessing and transmission to process the perceived data according to a preset rule chain, and transmits the key data to the core of the early warning system and the command screen through the message middleware (MQTT). The edge computing layer includes a face recognition and attendance statistics unit, which dynamically monitors personnel entry and exit, displays real-time personnel changes at the construction site, and calculates individual working hours based on entry and exit records by week, month, year, and cumulative dimensions, and reports to management personnel. Early warning model configuration platform: Based on industry norms and national standards, it classifies the attributes of personnel and equipment at construction sites and establishes a standardized single indicator library; it supports projects to add personalized single or mixed indicators according to their own characteristics, so as to realize flexible configuration and compatibility of early warning rules for multiple scenarios, multiple environments and multiple devices.

[0023] Furthermore, the early warning decision-making module includes a risk classification engine, an intelligent push engine, and a hidden danger closed-loop handling subsystem.

[0024] Furthermore, the risk grading engine is based on standards such as the "Standard for Safety Inspection of Building Construction" JGJ59-2011, and constructs a risk quantification model. It automatically outputs four levels of warning signals: red, orange, yellow, and blue, for the identified hidden dangers. Among them, the red warning signal represents the highest level, the orange warning signal represents the high level, the yellow warning signal represents the medium level, and the blue warning signal represents the low level.

[0025] Furthermore, the intelligent push engine automatically selects and executes the optimal push strategy based on the warning level and specific scenario, with target terminals including WeChat mini-programs for administrators, on-site intelligent broadcasting systems, and command center dispatch screens. Example

[0026] The construction site safety early warning method based on AI intelligent analysis includes the following steps: S1: The multi-source sensing layer collects on-site status data in real time; The on-site status data includes monitoring video, environment, tower crane status, and work cards; S2: The edge computing layer receives and processes the field status data, and sends the field status data to the early warning system core and command screen for real-time feedback; The edge computing layer uses a pre-defined rule chain from the IoT platform to process the on-site status data and sends key data to the early warning system core and command screen for real-time feedback through message middleware such as MQTT.

[0027] S3: The facial recognition unit dynamically monitors personnel entry and exit, updates the information of personnel present, and calculates individual working hours (weekly / monthly / yearly / cumulative) for management personnel to view; S4: The early warning decision module performs intelligent analysis based on on-site status data and the configured early warning model (based on industry standards and project-specific indicators); S5: If the warning rule is triggered, the system will immediately display the warning details on the command and dispatch screen and issue an alarm sound.

[0028] Upon receiving an alarm, managers can immediately contact the relevant area manager or person in charge via telephone or other means, based on system prompts, to direct on-site personnel to carry out rescue or rectification and resolve the danger in a timely manner.

[0029] The system automatically generates rectification orders, tracks the handling process, and completes the closed loop through AI or manual review.

[0030] This invention establishes a construction site safety early warning system and method based on AI model recognition. The system deeply integrates intelligent video analysis, IoT sensor data fusion, and building information modeling / geographic information system spatial positioning technology, achieving: second-level intelligent perception of safety hazards at construction sites; full-process digital closed-loop management of risk warning and handling; and structural optimization of safety management costs, significantly improving the intelligence level and efficiency of construction safety supervision.

[0031] The present invention has been described in the above-described embodiments; however, these embodiments are merely examples for implementing the present invention. It must be noted that the disclosed embodiments do not limit the scope of the present invention. Conversely, any modifications and refinements made without departing from the spirit and scope of the present invention are within the scope of patent protection of the present invention.

Claims

1. A construction site safety early warning system based on AI model recognition, characterized in that... ,include: The multi-source sensing layer includes a video acquisition unit, an intelligent environmental monitoring unit, an intelligent tower crane monitoring unit, and a portable intelligent work card unit; The edge computing layer receives and initially processes real-time status data from the multi-source sensing layer. The early warning and decision-making module includes a risk classification engine, an intelligent push engine, and a hidden danger closed-loop handling subsystem. The data platform and reporting center integrate and analyze data and early warning information from various perception layers, providing: risk heat maps based on geographical location, risk profile analysis of different job types, trend analysis of risk levels monitored by various intelligent devices, and comprehensive assessment reports of overall project safety risk indicators.

2. The construction site safety early warning system based on AI model recognition as described in claim 1, characterized in that: The video acquisition unit is equipped with 4K / 8K high-definition cameras and mobile surveillance cameras to achieve comprehensive monitoring of the construction site. The intelligent environmental monitoring unit monitors the dynamic environmental parameters of the construction site around the clock, including temperature, humidity, wind speed, and concentration of toxic gases, to ensure that construction personnel work in a safe environment. The intelligent tower crane monitoring unit integrates the tower crane's own operating status data and surrounding real-time video to provide real-time early warning of abnormal tower crane conditions, and the data is synchronized to the system decision center. The portable smart work card unit integrates positioning and emergency rescue functions. It uploads personnel location and movement trajectory at regular intervals through multiple signal transmission methods such as 4G, Bluetooth, and UWB. When personnel cross the boundary into a dangerous area or send out a distress signal, the system automatically triggers an alarm and links with the command and dispatch screen to issue a warning or initiate rescue.

3. The construction site safety early warning system based on AI model recognition as described in claim 2, characterized in that: The tower crane's own operating status data includes tilt angle, load, and wind speed; the tower crane's abnormal status includes overload and collision risk.

4. The construction site safety early warning system based on AI model recognition as described in claim 1, characterized in that: The edge computing layer utilizes the Internet of Things platform for data preprocessing and transmission, processes the sensed data according to a preset rule chain, and transmits key data to the core of the early warning system and the command screen through the message middleware (MQTT). The edge computing layer includes a face recognition and attendance statistics unit, which dynamically monitors personnel entry and exit, displays real-time personnel changes at the construction site, and calculates individual working hours based on entry and exit records by week, month, year, and cumulative dimensions, and reports to management personnel. Early warning model configuration platform: Based on industry norms and national standards, it classifies the attributes of personnel and equipment at construction sites and establishes a standardized single indicator library; it supports projects to add personalized single or mixed indicators according to their own characteristics, so as to realize flexible configuration and compatibility of early warning rules for multiple scenarios, multiple environments and multiple devices.

5. The construction site safety early warning system based on AI model recognition as described in claim 1, characterized in that: The early warning decision-making module includes a risk classification engine, an intelligent push engine, and a hidden danger closed-loop handling subsystem.

6. The construction site safety early warning system based on AI model recognition as described in claim 5, characterized in that: The risk grading engine is based on standards such as the "Standard for Safety Inspection of Building Construction" JGJ59-2011. It constructs a risk quantification model and automatically outputs four levels of warning signals: red, orange, yellow, and blue, for identified hidden dangers. The red warning signal represents the highest level, the orange warning signal represents the high level, the yellow warning signal represents the medium level, and the blue warning signal represents the low level.

7. The construction site safety early warning system based on AI model recognition as described in claim 5, characterized in that: The intelligent push engine automatically selects and executes the optimal push strategy based on the warning level and specific scenario. Target terminals include WeChat mini-programs for administrators, on-site intelligent broadcasting systems, and command center dispatch screens.

8. The construction site safety early warning method based on AI model recognition as described in claim 1, characterized in that: Includes the following steps: S1: The multi-source sensing layer collects on-site status data in real time; S2: The edge computing layer receives and processes the field status data, and sends the field status data to the early warning system core and command screen for real-time feedback; S3: The facial recognition unit dynamically monitors personnel entry and exit, updates the information of personnel present, and calculates individual working hours for management personnel to view. S4: The early warning decision module performs intelligent analysis based on on-site status data and the configured early warning model (based on industry standards and project-specific indicators); S5: If the warning rule is triggered, the system will immediately display the warning details on the command and dispatch screen and issue an alarm sound.

Images