Construction hoist safe driving assistance method and device based on artificial intelligence

Abstract

The application provides a construction elevator safe driving assistance method and device based on artificial intelligence, comprising: receiving a parking request sent by a first associated terminal from a construction personnel, the parking request being obtained by scanning a parking layer two-dimensional code pasted at a current parking layer where the construction personnel is located through the first associated terminal, and the parking request carrying an identifier of a target parking layer; generating at least one target parking task corresponding to the current parking layer based on the parking request and a first monitoring image frame of the current parking layer; determining a target competitive request of the target parking task in all competitive requests for any target parking task, so as to lock the target competitive request and a corresponding construction elevator with the target parking task; and generating a parking operation guide based on all target parking tasks locked by the construction elevator. The application can significantly improve the safety of the construction elevator.

Description

Technical Field

[0001] This invention relates to the field of hoist technology, and in particular to a method and device for safe driving assistance of construction hoists based on artificial intelligence. Background Technology

[0002] As an indispensable vertical transportation device in modern construction, the safe operation of construction hoists is directly related to the safety of personnel and equipment on construction sites. With the construction industry moving towards higher and more complex structures, the frequency of use and load requirements of construction hoists are constantly increasing, placing higher demands on hoist safety control technology. Currently, key operations such as starting, running, and stopping construction hoists still rely entirely on manual control by the operator, requiring the operator to possess sustained high levels of concentration and rapid reaction capabilities. During prolonged operation or in adverse weather conditions, human factors can easily lead to safety hazards. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide a construction hoist safety driving assistance method and device based on artificial intelligence, which can significantly improve the safety of construction hoists.

[0004] In a first aspect, the present invention provides a method for safe driving assistance of a construction hoist based on artificial intelligence, comprising: The system receives a docking request from the first associated terminal of the construction worker. The docking request is obtained by scanning the docking QR code affixed at the current docking floor where the construction worker is located by the first associated terminal. The docking request carries the identifier of the target docking floor. Based on the docking request and the first monitoring image frame of the current docking layer, at least one target docking task corresponding to the current docking layer is generated through a pre-deployed target model; Among all competing requests for any target docking task, identify the target competing request for the target docking task and lock the target competing request and its corresponding construction hoist to the target docking task. Based on all target docking tasks locked by the construction hoist, a docking operation guide is generated.

[0005] In one implementation, the docking request also carries images of the first number of passengers and the items to be transported; based on the docking request and the first monitoring image frame of the current docking level, at least one target docking task corresponding to the current docking level is generated through a pre-deployed target model, including: By using a pre-deployed target model, target detection is performed based on the first monitoring image frame to determine the total number of construction personnel and the type of items to be transferred corresponding to the current stop floor, in order to generate an initial docking task; Based on the images of the first number of passengers and the items to be transported carried by each docking request, the initial docking task is divided into multiple docking sub-tasks, which are the same as the number of docking requests. Based on the identifier of the target docking layer carried in the docking request, identify the running direction of the docking subtask corresponding to each docking request; Based on the safety requirements for riding construction hoists, the docking sub-tasks are clustered or split according to the direction of operation to obtain at least one target docking task. The target docking task carries a second number of passengers and the type of items to be transferred.

[0006] In one implementation, determining the target contention request for a target docking task from all contention requests for any given target docking task includes: For any target docking task, perform the following operations: Based on the second number of passengers and the type of items to be transferred carried by the target docking task, and the current operating status and other locked target docking tasks carried by each of the competing requests for the target docking task, a preliminary screening is performed on all competing requests for the target docking task to obtain candidate competing requests for the target docking task. Based on the second number of passengers and the type of items to be transferred carried by the target docking task, and the historical operation and maintenance data, current operating status and other locked target docking tasks carried by each candidate competition request, the multi-dimensional score value corresponding to the candidate competition request is determined. The multi-dimensional score value is used to evaluate the candidate competition request from the dimensions of item type adaptation, remaining capacity redundancy, operation efficiency, operation and maintenance and task conflict. Based on multi-dimensional scoring values, the target competition requests for the target docking task are determined.

[0007] In one implementation, docking operation instructions are generated based on all target docking tasks locked by the construction hoist, including: Based on the target floors to which all target docking tasks are locked by the construction hoist, a docking operation guide corresponding to the construction hoist is generated. The docking operation guide includes a dynamic hoisting map, which describes all target floors that need to be docked and the current floor. The dynamic hoisting map is sent to the second associated terminal so that when the hoist arrives at the target floor in sequence, the hoist operator is prompted to perform the docking operation through the second associated terminal.

[0008] In one implementation, the method further includes: The identifier of the construction hoist whose target docking task is locked is sent to the first associated terminal corresponding to the target competition request for the target docking task, so that the first associated terminal prompts the construction personnel to go to the hoist door of the construction hoist whose target docking task is locked.

[0009] In one embodiment, a monitoring device is installed inside the construction hoist, and the method further includes: During the operation of the construction hoist, when the hoist door of the construction hoist performs an opening and closing operation, the monitoring equipment is triggered to collect a second monitoring image frame. Target detection is performed based on the second monitoring image frame to determine the current number of passengers on the construction hoist, and the distribution of construction personnel on each floor is estimated by combining the opening and closing operations with the corresponding landing floor. And / or, dynamically track the elevator operator based on the second monitoring image frame to identify the elevator operator's action data, and determine whether the elevator operator has violated regulations based on the action data.

[0010] In one implementation, the method further includes: Receive login requests sent by the target associated terminal, and verify the permission scope corresponding to the target associated terminal based on the login request; Receive the query request sent by the target associated terminal for the construction hoist, and if the query request meets the permission scope, send the query result corresponding to the construction hoist back to the target associated terminal.

[0011] Secondly, the present invention also provides an artificial intelligence-based safety driving assistance device for construction hoists, comprising: The docking request receiving module is used to receive docking requests sent from the first associated terminal of the construction personnel. The docking request is obtained by the first associated terminal scanning the docking QR code affixed at the current docking floor where the construction personnel are located. The docking request carries the identifier of the target docking floor. The task generation module is used to generate at least one target docking task corresponding to the current docking layer based on the docking request and the first monitoring image frame of the current docking layer, through a pre-deployed target model. The task allocation module is used to determine the target competition request for any target docking task from all competition requests for any target docking task, so as to lock the target competition request and its corresponding construction hoist to the target docking task; The task guidance module is used to generate docking operation guidance based on all target docking tasks locked by the construction hoist.

[0012] Thirdly, the present invention also provides an electronic device including a processor and a memory, the memory storing computer-executable instructions executable by the processor, the processor executing the computer-executable instructions to implement any of the methods provided in the first aspect.

[0013] Fourthly, the present invention also provides a computer-readable storage medium storing computer-executable instructions, which, when invoked and executed by a processor, cause the processor to implement any of the methods provided in the first aspect.

[0014] This invention provides an artificial intelligence-based method and device for safe driving assistance of construction hoists. Applied to a construction hoist safe driving assistance system, the system stores equipment attribute information for all construction hoists on the construction site and personnel information for their associated hoist drivers. Each hoist door at a different floor has a QR code for that floor affixed. The system receives a docking request from a first associated terminal of a construction worker. This request is obtained by scanning the QR code at the current floor location using the first associated terminal. The docking request carries the identifier of the target floor. The system also triggers the monitoring equipment at the current floor to capture a first monitoring image frame. The first monitoring image frame displays at least the construction worker located outside the hoist door at the current floor and... / or the item to be transported; based on the docking request and the first monitoring image frame, generate at least one target docking task corresponding to the current docking floor; receive a competition request from the construction hoist, the competition request carrying the construction hoist's historical operation and maintenance data, current operating status, and other locked target docking tasks, the current operating status including at least the current docking floor, current number of passengers, and current type of transported item; among all competition requests for any target docking task, determine the target competition request for the target docking task, and lock the target competition request and its corresponding construction hoist with the target docking task; generate docking operation instructions based on all locked target docking tasks of the construction hoist, so that the second associated terminal of the hoist operator bound to the construction hoist prompts the hoist operator to perform the docking operation according to the docking operation instructions. The above method combines monitoring image frames and docking requests uploaded by construction personnel by scanning QR codes to generate target docking tasks to be executed. Each construction hoist competes for the target docking request, and based on the competition results, it locks the target competition request and its corresponding construction hoist and target docking task, which facilitates the subsequent generation of docking operation instructions to guide the hoist operator to perform the docking operation, significantly improving the safety and automation level of construction hoists.

[0015] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention are realized and obtained in accordance with the structures particularly pointed out in the description, claims and drawings.

[0016] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0017] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0018] Figure 1 A flowchart illustrating a construction hoist safety driving assistance method based on artificial intelligence, provided for an embodiment of the present invention; Figure 2 A schematic diagram of a user interface provided in an embodiment of the present invention; Figure 3 A schematic diagram of a construction hoist safety driving assistance device based on artificial intelligence provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Currently, existing construction hoists suffer from high operational risks and low levels of automation. To address this, this invention provides an artificial intelligence-based method and device for safe driving assistance of construction hoists, which can significantly improve the safety and automation of construction hoists.

[0021] To facilitate understanding of this embodiment, a detailed description of an artificial intelligence-based construction hoist safety driving assistance method disclosed in this embodiment of the invention will be provided first. This method is applied to a construction hoist safety driving assistance system. The construction hoist safety driving assistance system stores equipment attribute information corresponding to all construction hoists in the construction site and personnel information corresponding to the hoist drivers bound to them. A landing QR code is affixed to the hoist door at each landing level in the construction site. See [link to relevant documentation]. Figure 1 The diagram shows a flowchart of a construction hoist safety driving assistance method based on artificial intelligence. The method mainly includes the following steps S102 to S108: Step S102: Receive a docking request sent from the first associated terminal of the construction personnel.

[0022] The docking request is obtained by scanning the QR code affixed to the current docking floor where the construction worker is located using the first associated terminal. The docking request carries the identifier of the target docking floor, as well as the first number of passengers and an image of the item to be transported. In one example, after the construction worker logs into the first associated terminal and scans the docking floor QR code, they are redirected to an information entry interface. They then fill in the target docking floor they wish to reach. If no passengers or images of the item to be transported are entered, the first number of passengers is assumed to be 1, and the item to be transported is assumed to be "none". If the number of passengers and items to be transported are entered, the docking request will then carry the first number of passengers and the image of the item to be transported.

[0023] Step S104: Based on the docking request and the first monitoring image frame of the current docking layer, generate at least one target docking task corresponding to the current docking layer through a pre-deployed target model.

[0024] The first monitoring image frame displays at least the construction workers and / or items to be transported located outside the elevator doors on the current stopping floor. In one example, monitoring equipment is deployed at the current stopping floor to capture data on the area outside the elevator doors, thus obtaining the first monitoring image frame containing the construction workers and / or items to be transported.

[0025] The target docking task carries a second passenger number and the type of item to be transferred. In one implementation, an initial docking task is first roughly generated based on a first monitoring image frame, and then the initial docking task is divided into multiple docking sub-tasks with the same number of docking requests, taking into account the safety riding requirements corresponding to the construction hoist as constraints and the direction of operation, to obtain at least one target docking task.

[0026] Step S108: Among all competing requests for any target docking task, determine the target competing request for the target docking task, and lock the target competing request and its corresponding construction hoist to the target docking task.

[0027] The competition request carries the construction hoist's historical operation and maintenance data, current operating status, and other locked target docking tasks. The current operating status includes at least the current docking floor, the current number of passengers, and the type of goods being transported. The historical operation and maintenance data can include the time of the last operation and maintenance, as well as the construction hoist's working data from the last operation and maintenance to the current time, such as working time, total lifting distance, and number of stops at each floor. In one example, one or more channels for construction hoist operation can be deployed on the construction site. For example, if two channels are deployed, that is, there are two construction hoists available for construction workers to choose from on the same docking floor, and both construction hoists can send competition requests for target docking tasks.

[0028] In one example, preliminary screening based on numerical comparison can be performed on the data carried by the competition request and the target docking task to select candidate competition requests that match the target docking task. Then, based on the data carried by the candidate competition requests and the target docking task, all candidate competition requests can be evaluated in multiple dimensions to select the target competition request locked by the target docking task and its corresponding construction hoist.

[0029] Step S108: Generate docking operation instructions based on all target docking tasks locked by the construction hoist.

[0030] The docking operation guidance can be a dynamic elevation map, which describes all target floors that need to be docked and the current docking floor. It can also dynamically display the next target floor that will be docked. In one example, a dynamic elevation map can be generated according to the target floor corresponding to the target docking task, and the dynamic elevation map can be displayed on the second associated terminal. At the same time, voice, vibration or other forms of prompts can be generated based on the dynamic elevation map to prompt the elevator operator to perform the docking operation.

[0031] The artificial intelligence-based safe driving assistance method for construction hoists provided in this invention combines monitoring image frames and docking requests uploaded by construction personnel via scanning QR codes to generate target docking tasks to be executed. Each construction hoist competes for the target docking request, and based on the competition result, the target competition request and its corresponding construction hoist and target docking task are locked, which facilitates the subsequent generation of docking operation guidance to guide the hoist operator to perform the docking operation, significantly improving the safety and automation level of construction hoists.

[0032] For ease of understanding, this embodiment of the invention provides a specific implementation of an artificial intelligence-based safe driving assistance method for construction hoists.

[0033] Before performing the aforementioned step S102, the construction hoist needs to be pre-configured and calibrated. See the table below for details: Table 1 Host Number

[0034] Table 2 Host Information

[0035] Table 3 Project Name

[0036] Table 4 Equipment Numbering

[0037] Table 5 Manufacturer Selection

[0038] Table 6 Model Selection

[0039] Table 7 Left Cage / Right Cage / Single Cage

[0040] Table 8 Passenger Capacity Limit

[0041] Table 9 Front Camera Detection

[0042] Table 10 AI Function On / Off

[0043] Table 11 AI Function Selection

[0044] Table 12 Identification Channel Selection

[0045] Table 13 Confidence Level

[0046] Table 14 Audio Upload Control

[0047] Table 15 System Restart

[0048] Table 16 Accessing Wi-Fi Settings

[0049] Table 17 Channel Activation

[0050] Table 18 Camera Web Configuration

[0051] Table 19 Front / Rear Door Status Calibration

[0052] Table 20 Height Calibration

[0053] Table 21 Stop Floor Calibration

[0054] In practical applications, height calibration must be completed before floor stop calibration can proceed; otherwise, a prompt will appear and the user will be forced to return to the previous page. Enter the unit project design height and number of design floors. The number of standard floors is optional. If the standard floor height is entered, the product of the standard floor height and the number of design floors must be calculated to ensure it exceeds the design height. If it does, the next step cannot be completed. If no standard floor height is entered, the average floor height will be automatically calculated and displayed in a list for individual or batch revisions of floor stop calibrations for special floor heights.

[0055] Furthermore, you can also revise the floor height, including: selecting the floor height to be revised, clicking the revision button, and adjusting the floor height within the allowed range.

[0056] Table 22 Elevator Lifting Section

[0057] In practical applications, the initial height and number of floors are displayed in a list format (the height indicates the maximum number of floors), and the current maximum number of floors can be adjusted in the elevator height column. The adjustment range must be greater than the initial height and number of floors, but less than or equal to the designed maximum number of floors.

[0058] Table 23 Weight Calibration

[0059] Table 24 Relay Connections

[0060] Based on the aforementioned setup and calibration of the construction hoist, subsequent steps S102 to S112 can be performed.

[0061] In step S102 above, a QR code for each floor stop is affixed, also known as an elevator call QR code. On-site construction personnel can scan the code to call the elevator for the corresponding floor. This function is available after the floor stop calibration is completed. It automatically synchronizes the floor stop information in the calibration information to the program and automatically generates QR codes in batches according to the floor stop information when generating QR codes.

[0062] Regarding the aforementioned step S104, a monitoring device is deployed at the current stopping floor, which can collect data on the area outside the elevator door to obtain a first monitoring image frame containing construction personnel and / or items to be transported, and generate a target stopping task according to steps 1A to 1D as follows: Step 1A: Using a pre-deployed target model, target detection is performed based on the first monitoring image frame to determine the total number of construction personnel and the type of items to be transferred corresponding to the current stop floor, in order to generate the initial docking task.

[0063] In one example, a pre-trained target detection model can be used to detect targets in the first monitoring image frame. Specifically, the total number of construction workers in the first monitoring image frame can be determined by recognizing head contours and shoulder features, and the type of the items to be transmitted can be matched and identified by combining the historical item database.

[0064] In another example, environmental sensor data at the current stopping floor is acquired in real time. This data includes at least light intensity, dust concentration, and / or humidity. The environmental sensor data is then fused with a first monitoring image frame using multimodal data fusion to generate an environmental perception feature vector. Based on this feature vector, a target detection model with the highest adaptability to the current environmental conditions is dynamically selected from a pre-deployed set of target models, or the image preprocessing parameters and network weights of a base model are dynamically adjusted to perform target detection using the adjusted model. The details are as follows: The system acquires environmental sensor data in real time from environmental sensors deployed outside the elevator door of the current floor. The environmental sensors include at least one or more of light intensity sensors, dust concentration sensors, and humidity sensors. The environmental sensor data reflects the real physical environment of the current floor at the time the docking request is triggered.

[0065] Multimodal data fusion is performed between environmental sensor data and the first monitoring image frame: First, the environmental sensor data is normalized and mapped into a structured environmental feature sequence with time synchronization markers; then, the environmental feature sequence is concatenated with the pixel-level feature map of the first monitoring image frame in the channel dimension, or a unified environmental perception feature vector is generated through a lightweight cross-modal attention module (such as Cross-ModalTransformerEncoderLayer), which simultaneously encodes visual content semantics and environmental physical constraint information.

[0066] Based on environmentally aware feature vectors, dynamic model selection or parameter adjustment operations are performed on a set of target models pre-deployed in local edge computing units or cloud servers: If a model ensemble approach is adopted, the target detection model with the highest fit is selected based on the light intensity value, dust concentration level and humidity range represented in the environmental perception feature vector, and matched with a preset environment-model mapping rule table. The target model ensemble contains multiple specialized models trained on datasets of different environmental scenarios, such as: strong light low dust model, weak light high dust model, high humidity fog model, etc. Each model has been robustly enhanced to address typical degradation phenomena such as blurred human outlines, degraded object textures and decreased image contrast in its corresponding environment. If a single base model is used, the image preprocessing parameters (such as automatic white balance gain coefficient, histogram equalization intensity, number of dehazing algorithm iterations) and network weights (such as applying environment-aware gating modulation to shallow convolution kernels to adjust their response sensitivity to low-contrast edges) of the base model are dynamically adjusted based on the environment-aware feature vector, so that the adjusted model is more suitable for the current actual imaging quality.

[0067] Finally, using the selected target detection model or the dynamically adjusted base model, target detection is performed on the first monitoring image frame to identify the number of construction workers, their spatial distribution, and the types of goods to be transported, thereby supporting the generation of subsequent initial docking tasks.

[0068] This invention addresses the problem of a sharp decline in the detection accuracy of fixed AI models under harsh construction environments such as sudden changes in lighting and pervasive dust. By introducing an environmental perception mechanism and dynamically adjusting the model, the system maintains high reliability and robustness in identifying personnel and objects in complex and ever-changing construction site environments, avoiding misjudgments or omissions caused by environmental interference, and fundamentally guaranteeing the accuracy of the task generation process.

[0069] Step 1B: Based on the first passenger and the image of the item to be transported carried in each stop request, the initial stop task is divided into multiple stop sub-tasks with the same number of stop requests. The passenger number corresponding to the stop sub-task is the first passenger carried in the corresponding stop request, and the type of item to be transported corresponding to the stop sub-task is obtained by recognizing the image of the item to be transported carried in the corresponding stop request.

[0070] Step 1C: Based on the target stop layer identifier carried in the stop request, identify the running direction of the stop subtask corresponding to each stop request. In one example, if the target stop layer is higher than the current stop layer, the running direction is upward; if the target stop layer is lower than the current stop layer, the running direction is downward.

[0071] Step 1D: Based on the safety requirements for the construction hoist, cluster or split the docking sub-tasks according to the direction of travel to obtain at least one target docking task. Each target docking task carries a second number of passengers and the type of goods to be transported. In one example, if the number of passengers or the weight of the goods corresponding to a docking sub-task does not meet the safety requirements, task splitting is necessary. If the safety requirements are met, the docking sub-tasks can be clustered under the premise of meeting the safety requirements. For example, docking sub-tasks with the same direction of travel and the same type of goods can be grouped into a unified target docking task.

[0072] For example, there are sub-tasks: Sub-task 1: Upward, target floor 5, item "steel bar (heavy)," 3 people, weight 1.2 tons; Sub-task 2: Upward, target floor 7, item "tool kit (light)," 2 people; Sub-task 3: Downward, target floor 2, item "paint bucket (flammable)," 1 person.

[0073] After clustering: Target docking task 1 (upward clustering): contains subtask 1 + subtask 2, with parameters "upward direction, target docking level 5 / 7, total number of people 5, item type {heavy, light}"; Target docking task 2 (downward separate): contains subtask 3, with parameters "downward direction, target docking level 2, number of people 1, item type {flammable}".

[0074] Regarding step S106 above, each construction hoist can send a competition request for all target docking tasks. The competition request must carry the construction hoist's historical operation and maintenance data, current operating status, and other locked target docking tasks. Based on this, the following operation is performed for any target docking task: Step 2A: Based on the second number of passengers and the type of items to be transferred carried by the target docking task, and the current operating status and other locked target docking tasks carried by each of the competing requests for the target docking task, perform a preliminary screening of all competing requests for the target docking task to obtain candidate competing requests for the target docking task.

[0075] In one implementation, based on the core parameters of the target docking task (second passenger capacity, volume / type of goods to be transported), all competing requests are initially filtered, retaining elevators that meet the basic compatibility criteria. For example, the filtering rules may include: passenger capacity compatibility: the sum of the elevator's current passenger capacity, the second passenger capacity of other locked target docking tasks, and the second passenger capacity of the target docking task is less than or equal to the construction elevator's rated passenger capacity; volume compatibility: the total volume of the goods to be transported is less than or equal to the remaining car volume of the construction elevator; and item type matching: the construction elevator's functional attributes must include all item types of the target docking task.

[0076] Step 2B: Based on the second number of passengers and the type of items to be transferred carried by the target docking task, and the historical operation and maintenance data, current operating status and other locked target docking tasks carried by each candidate competing request, determine the multi-dimensional score value corresponding to the candidate competing request.

[0077] The multi-dimensional scoring system evaluates candidate competition requests based on four dimensions: item type compatibility, remaining capacity redundancy, operational efficiency, maintenance, and task conflict. Item type compatibility describes the match between the type of item to be transported and the elevator's functional attributes. Remaining capacity redundancy describes the relationship between the sum of the elevator's current load capacity and the weight of the item to be transported, and the elevator's rated load capacity. Operational efficiency describes the distance between the elevator's current stop floor and the current stop floor corresponding to the docking request. Maintenance uses historical maintenance data to describe whether the elevator can safely execute the target docking task. Task conflict further confirms whether the direction of other target docking tasks already locked by the elevator is consistent with the direction of this target docking task.

[0078] Step 2C: Based on multi-dimensional scoring values, determine the target competition request for the target docking task. In one example, the candidate competition request with the highest multi-dimensional scoring value can be selected as the target competition request. If the multi-dimensional scores of multiple subsequent competition requests are the same or less than a preset difference threshold, the target competition request for the target docking task can be further determined according to arbitration rules, such as: priority matching for hazardous materials: if the target task includes flammable, explosive, or other hazardous materials, the elevator with a higher explosion-proof rating is selected first; precise matching of remaining volume: compare the difference between the remaining volume of the elevator and the volume of the task items, the smaller the difference, the higher the priority; driver qualification weighting: query the qualification level of the driver bound to the elevator, the higher the level, the higher the priority; random allocation: if the above conditions are still the same, the system random algorithm will be used for selection, etc.

[0079] Regarding the aforementioned step S108, a docking operation guide corresponding to the construction hoist can be generated based on the target landing floors to which all target docking tasks locked by the construction hoist belong. The docking operation guide includes a dynamic hoisting map, describing all target landing floors requiring docking and the current landing floor. The dynamic hoisting map is sent to the second associated terminal so that the hoist operator is prompted to perform the docking operation upon sequential arrival at the target landing floors. For example, the second associated terminal has a program or app installed, such as... Figure 2The diagram shows a user interface that can display the current weight and its corresponding danger level, such as a current weight of 9.3 being safe and a current weight of 14.5 being relatively dangerous. It can also display the current floor and direction of travel, as well as other operating statuses of the construction hoist, such as the opening and closing status of the front and rear doors and their tilt angles. Furthermore, the construction hoist can be operated through a second terminal.

[0080] Table 25 Homepage Overview

[0081] Table 26 Overview of Work Pages

[0082] Furthermore, the second associated terminal can also display the current operating status of the construction hoist and issue an operating alarm when the current operating status is abnormal. Optionally, the operating alarm is divided into five levels (initial recommended setting), which can be customized. Logical settings have been configured so that a higher-level alarm is greater than or equal to a lower-level alarm.

[0083] Table 27 Alarm Logic

[0084] For example, the logic is: triggering at each level; if multiple levels are triggered simultaneously, the highest-level alarm is recorded. The linkage is: relay alarms, SMS alarms, and alarms exceeding level three automatically start recording for 5 minutes.

[0085] Furthermore, after locking the task and request, the identifier of the construction hoist locked by the target docking task can be sent to the first associated terminal corresponding to the target competition request of the target docking task, so that the first associated terminal can prompt the construction personnel to go to the hoist door of the construction hoist locked by the target docking task.

[0086] Furthermore, during the operation of the construction hoist, when the hoist door is detected to perform an opening and closing operation, a second monitoring image frame is triggered and collected by the monitoring equipment; target detection is performed based on the second monitoring image frame to determine the current number of passengers on the construction hoist, and the distribution of construction personnel on each floor of the construction site is estimated by combining the floor corresponding to the opening and closing operation; and / or, the hoist driver is dynamically tracked based on the second monitoring image frame to identify the hoist driver's action data, and the hoist driver's violation is determined based on the action data.

[0087] The monitoring equipment needs to cover the entire interior of the elevator from a fixed perspective to avoid blind spots and thus more accurately detect the current number of passengers on the construction elevator.

[0088] Furthermore, the current number of passengers determined by the second monitoring image frame can be used to update the current operating status of the construction hoist, so as to more reliably compete for the new target docking task when generating it.

[0089] In one implementation, the violation detection process is as follows: Step 3A: The system calls the pre-stored elevator driver face template library (containing facial feature point coordinates, such as 21 key points like the corners of the eyes, the tip of the nose, and the corners of the mouth) to perform face detection and matching on the second monitoring image frame. Specifically: The MTCNN (Multi-Task Convolutional Neural Network) algorithm is used to quickly locate the face region in the image and generate a minimum bounding rectangle to ensure that the selected area covers the entire head and the upper edge of the shoulders; the detected facial feature points are compared with the template library, and when the matching degree is greater than a threshold, the target corresponding to the bounding box is confirmed as the current elevator driver; the detection results of the elevator driver are verified and tracked by combining historical monitoring image frames to avoid matching failures due to factors such as occlusion or turning.

[0090] Step 3B: Bounding Box Image Correction and Feature Enhancement. To eliminate the impact of monitoring angle deviations (such as camera tilt and facial distortion caused by the driver looking down) on action recognition, geometric correction and image optimization are performed on the bounding box region. Specifically, this includes: calculating the image tilt angle based on facial feature points, and correcting the bounding box region to a frontal viewpoint through affine transformation to ensure that the driver's actions in different frames are in the same coordinate system; denoising and contrast enhancement are performed on the corrected bounding box image so that subsequent action recognition algorithms can clearly capture key features such as eye and hand features, including blinking frequency and operational actions.

[0091] Step 3C: Bounding box tracking and behavioral feature extraction based on time-series frames. Multi-frame correlation tracking technology is used to capture the driver's continuous actions over a period of time and extract behavioral feature parameters. Specifically, this includes: using the DeepSORT (Deep Simple Online and Real-Time Tracking) algorithm to match the driver's bounding box in the current frame with the previous frame, achieving continuous tracking across frames. Even if the driver is briefly obscured, trajectory prediction can re-lock the tracking after the obstruction ends. The following dynamic parameters are extracted from the tracking trajectory: Head movements: pitch angle (i.e., vertical head turning angle), yaw angle (i.e., horizontal head turning angle), and dwell time; Hand movements: relative position of hands and control sticks, hand lifting height, and frequency of button operation; Body posture: upper body tilt angle, whether a safety harness is worn, etc.

[0092] Step 3D involves the construction and matching of a pre-defined behavioral feature library. The system has a built-in library of violation behavior feature templates. Extracted driver behavioral features are compared with these templates to determine if a violation exists. For example, the cosine similarity between driver behavioral features and each pre-defined template is calculated in real time. When the similarity of a template exceeds a matching threshold, a violation warning is triggered, and three key images are captured as evidence. If multiple templates are matched simultaneously, they are sorted by violation level, with higher-level warnings prioritized.

[0093] Furthermore, it receives login requests from the target associated terminal and verifies the permission scope corresponding to the target associated terminal based on the login request; it also receives query requests from the target associated terminal regarding the construction hoist, and if the query request meets the permission scope, it sends the query results corresponding to the construction hoist back to the target associated terminal. For example: Table 28 Status Viewing Permissions

[0094] Table 29 Alarm Viewing Permissions

[0095] Table 30 Driver Information Viewing Permissions

[0096] Based on the foregoing embodiments, this invention provides an artificial intelligence-based construction hoist safety driving assistance device. This device is applied to a construction hoist safety driving assistance system. The system stores equipment attribute information for all construction hoists within the construction site and personnel information for their associated hoist drivers. A QR code for each floor's hoist door is affixed to the hoist door at each floor. (See also...) Figure 3 The diagram shows a structural schematic of an artificial intelligence-based safety driving assistance device for construction hoists. The device mainly includes the following parts: The docking request receiving module 302 is used to receive docking requests sent from the first associated terminal of the construction personnel. The docking request is obtained by the first associated terminal scanning the docking QR code affixed at the current docking floor where the construction personnel are located. The docking request carries the identifier of the target docking floor. The task generation module 304 is used to generate at least one target docking task corresponding to the current docking layer based on the docking request and the first monitoring image frame of the current docking layer, through a pre-deployed target model. The task allocation module 306 is used to determine the target competition request for the target docking task among all competition requests for any target docking task, so as to lock the target competition request and its corresponding construction hoist to the target docking task. Task guidance module 308 is used to generate docking operation guidance based on all target docking tasks locked by the construction hoist.

[0097] The artificial intelligence-based construction hoist safety driving assistance device provided in this embodiment of the invention combines monitoring image frames and docking requests uploaded by construction personnel by scanning QR codes to generate target docking tasks to be executed. Each construction hoist competes for the target docking request, and based on the competition result, locks the target competition request and its corresponding construction hoist and target docking task, which facilitates the subsequent generation of docking operation guidance to guide the hoist operator to perform the docking operation, significantly improving the safety and automation level of the construction hoist.

[0098] The device provided in this embodiment of the invention has the same implementation principle and technical effect as the aforementioned method embodiment. For the sake of brevity, any parts not mentioned in the device embodiment can be referred to the corresponding content in the aforementioned method embodiment.

[0099] This invention provides an electronic device, specifically, the electronic device includes a processor and a storage device; the storage device stores a computer program, and the computer program, when run by the processor, executes the method described in any of the above embodiments.

[0100] Figure 4 The present invention provides a schematic diagram of the structure of an electronic device 100, which includes a processor 40, a memory 41, a bus 42 and a communication interface 43. The processor 40, the communication interface 43 and the memory 41 are connected through the bus 42. The processor 40 is used to execute executable modules, such as computer programs, stored in the memory 41.

[0101] The memory 41 may include high-speed random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Communication between this system network element and at least one other network element is achieved through at least one communication interface 43 (which can be wired or wireless), such as the Internet, wide area network, local area network, metropolitan area network, etc.

[0102] Bus 42 can be an ISA bus, PCI bus, or EISA bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 4 The symbol is represented by a single double-headed arrow, but this does not mean that there is only one bus or one type of bus.

[0103] The memory 41 is used to store programs. After receiving an execution instruction, the processor 40 executes the program. The method executed by the device for defining the flow process disclosed in any of the foregoing embodiments of the present invention can be applied to the processor 40 or implemented by the processor 40.

[0104] Processor 40 may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by the integrated logic circuitry in the hardware of processor 40 or by instructions in software form. Processor 40 can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this invention can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in memory 41. The processor 40 reads the information in memory 41 and, in conjunction with its hardware, completes the steps of the above method.

[0105] The computer program product of the readable storage medium provided in the embodiments of the present invention includes a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the methods described in the foregoing method embodiments. For specific implementation, please refer to the foregoing method embodiments, which will not be repeated here.

[0106] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0107] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit it. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the technical scope disclosed in the present invention, or make equivalent substitutions for some of the technical features; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for safe driving assistance of construction hoists based on artificial intelligence, characterized in that, include: The system receives a docking request from the first associated terminal of the construction worker. The docking request is obtained by the first associated terminal scanning the docking QR code affixed at the current docking floor where the construction worker is located. The docking request carries the identifier of the target docking floor. Based on the docking request and the first monitoring image frame of the current docking layer, at least one target docking task corresponding to the current docking layer is generated through a pre-deployed target model; Among all the contention requests for any of the target docking tasks, a target contention request for the target docking task is determined, and the target contention request and its corresponding construction hoist are locked to the target docking task; Based on all the target docking tasks locked by the construction hoist, a docking operation guide is generated.

2. The artificial intelligence-based safe driving assistance method for construction hoists according to claim 1, characterized in that, The docking request also carries images of the first number of passengers and the items to be transported; based on the docking request and the first monitoring image frame of the current docking level, at least one target docking task corresponding to the current docking level is generated through a pre-deployed target model, including: By using a pre-deployed target model, target detection is performed based on the first monitoring image frame to determine the total number of construction personnel and the type of items to be transferred corresponding to the current stop floor, in order to generate an initial docking task; Based on the first number of passengers and the image of the item to be transported carried in each of the stop requests, the initial stop task is divided into multiple stop sub-tasks with the same number of stop requests. Based on the identifier of the target stopping layer carried in the docking request, the running direction of the docking subtask corresponding to each docking request is identified; Based on the safety riding requirements corresponding to the construction hoist, the stopping sub-tasks are clustered or split according to the running direction to obtain at least one target stopping task, which carries a second number of passengers and the type of items to be transferred.

3. The artificial intelligence-based safe driving assistance method for construction hoists according to claim 1, characterized in that, Determining the target contention request for a target docking task from among all contention requests for any of the target docking tasks includes: For any of the aforementioned target docking tasks, perform the following operations: Based on the second number of passengers and the type of items to be transferred carried by the target docking task, and the current operating status and other locked target docking tasks carried by each of the competing requests for the target docking task, a preliminary screening is performed on all the competing requests for the target docking task to obtain candidate competing requests for the target docking task. Based on the second number of passengers and the type of items to be transferred carried by the target docking task, and the historical operation and maintenance data, the current operating status and other locked target docking tasks carried by each of the candidate competition requests, a multi-dimensional score value corresponding to the candidate competition request is determined. The multi-dimensional score value is used to evaluate the candidate competition request from the dimensions of item type adaptation, remaining capacity redundancy, operation efficiency, operation and maintenance and task conflict. Based on the multi-dimensional scoring values, the target competition request for the target docking task is determined.

4. The artificial intelligence-based safe driving assistance method for construction hoists according to claim 1, characterized in that, Based on all the target docking tasks locked by the construction hoist, a docking operation guide is generated, including: Based on the target stopping floors to which all the target stopping tasks locked by the construction hoist belong, a stopping operation guide corresponding to the construction hoist is generated. The stopping operation guide includes a dynamic lifting map, which describes all the target stopping floors that need to be stopped and the current stopping floor. The dynamic lifting map is sent to the second associated terminal so that when the hoist arrives at the target stopping floor in sequence, the hoist operator is prompted by the second associated terminal to perform the stopping operation.

5. The artificial intelligence-based safe driving assistance method for construction hoists according to claim 1, characterized in that, The method further includes: The identifier of the construction hoist locked by the target docking task is sent to the first associated terminal corresponding to the target competition request locked by the target docking task, so that the first associated terminal prompts the construction personnel to go to the hoist door of the construction hoist locked by the target docking task.

6. The artificial intelligence-based safe driving assistance method for construction hoists according to claim 1, characterized in that, The construction hoist is equipped with monitoring equipment, and the method further includes: During the operation of the construction hoist, when the hoist door of the construction hoist is detected to perform an opening and closing operation, the monitoring equipment is triggered to collect a second monitoring image frame. Target detection is performed based on the second monitoring image frame to determine the current number of passengers on the construction hoist, and the distribution of construction personnel on each floor of the construction site is estimated by combining the floor corresponding to the opening and closing operation. And / or, dynamically track the elevator operator based on the second monitoring image frame to identify the elevator operator's action data, and determine whether the elevator operator has violated regulations based on the action data.

7. The artificial intelligence-based safe driving assistance method for construction hoists according to claim 1, characterized in that, The method further includes: Receive a login request sent by the target associated terminal, and verify the permission scope corresponding to the target associated terminal based on the login request; The system receives a query request sent by the target associated terminal for the construction hoist. If the query request meets the permission scope, the system feeds back the query result corresponding to the construction hoist to the target associated terminal.

8. A construction hoist safety driving assistance device based on artificial intelligence, characterized in that, include: The docking request receiving module is used to receive docking requests sent from the first associated terminal of the construction personnel. The docking request is obtained by the first associated terminal scanning the docking QR code affixed at the current docking floor where the construction personnel are located. The docking request carries the identifier of the target docking floor. The task generation module is used to generate at least one target docking task corresponding to the current docking layer based on the docking request and the first monitoring image frame of the current docking layer, through a pre-deployed target model. The task allocation module is used to determine the target competition request for the target docking task among all the competition requests for any target docking task, so as to lock the target competition request and its corresponding construction hoist to the target docking task; The task guidance module is used to generate docking operation guidance based on all the target docking tasks locked by the construction hoist.

9. An electronic device, characterized in that, The method includes a processor and a memory, the memory storing computer-executable instructions executable by the processor, the processor executing the computer-executable instructions to implement the method of any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions that, when invoked and executed by a processor, cause the processor to perform the method according to any one of claims 1 to 7.

Images