Intelligent construction hoist control method and system

Abstract

The invention relates to an intelligent construction elevator control method and system, and relates to the technical field of elevators, and the intelligent construction elevator control method comprises the steps that a historical bearing distribution sequence is collected, elevator inclination accumulative influence analysis is carried out, and accumulative influence inclination parameters are obtained; monitoring images in the elevator are collected, real-time bearing recognition decision is made according to the accumulated influence inclination parameters, bearing recognition is conducted on the monitoring images, and bearing distribution is obtained; according to the bearing distribution, real-time inclination influencing parameters are obtained through prediction, and inclination influencing parameters are obtained through calculation; and according to the inclination influencing parameters, compensation level control analysis is conducted on the elevator, compensation control parameters are obtained, error correction is conducted through accumulative inclination influencing parameters, a compensation control parameter interval is obtained, and compensation level control is conducted on the elevator. The problem of inclination caused by uneven load bearing of the building hoist is solved, accurate adjustment of the horizontal state of the lifting cage is achieved, the stability and accuracy of equipment operation are improved, and the safety and efficiency of construction are guaranteed.

Description

Technical Field

[0001] This application relates to the technical field of elevators, and particularly to an intelligent construction elevator control method and system. Background Art

[0002] With the intelligent development of the construction industry, the safe and stable operation of construction elevators has become an important link in ensuring project progress and operation safety. At present, when a construction elevator transports heavy objects such as construction materials, the problem of the cage tilting due to uneven load bearing is relatively prominent. Traditional control methods are difficult to accurately adjust the horizontal state of the cage in real time, which easily causes obvious shaking during operation and misalignment with the building plane when docking. This not only affects the service life of the equipment but also may pose safety hazards.

[0003] Most existing elevator control methods rely on adjusting fixed parameters, do not fully combine historical operation data and real-time load-bearing distribution, and have insufficient ability to predict and compensate for dynamic tilting risks. As a result, control delays and poor adjustment accuracy occur, making it difficult to meet the higher requirements for equipment stability and safety in complex construction scenarios. Summary of the Invention

[0004] To solve the above technical problems, this application provides an intelligent construction elevator control method and system, which overcomes the problems of existing elevator control not combining dynamic data and lagging tilt adjustment, and improves the accuracy of cage horizontal control and the stability of operation.

[0005] The embodiments of this application disclose the following technical solutions: In a first aspect, the embodiments of this application provide an intelligent construction elevator control method, and the method includes: According to the historical operation records of the elevator, collect the historical load-bearing distribution sequence, conduct an analysis of the cumulative impact of elevator tilt, and obtain the cumulative impact tilt parameter; During the operation of the construction elevator, collect the monitoring images inside the elevator, make a real-time load-bearing recognition decision based on the cumulative impact tilt parameter, and perform load-bearing recognition on the monitoring images to obtain the load-bearing distribution; Based on the load-bearing distribution, predict the real-time impact tilt parameter, and combine the cumulative impact tilt parameter to calculate the impact tilt parameter; Based on the impact tilt parameter, conduct a compensation horizontal control analysis on the elevator to obtain the compensation control parameter, use the cumulative impact tilt parameter for error correction, obtain the compensation control parameter interval, and perform compensation horizontal control on the elevator.

[0006] In a second aspect, the embodiments of this application provide an intelligent construction elevator control system, and the system includes: Historical data tilt analysis module, which is used to collect historical load-bearing distribution sequences according to the historical operation records of the lift, conduct cumulative impact analysis of lift tilt, and obtain cumulative impact tilt parameters; Real-time load-bearing identification and decision-making module, which is used to collect monitoring images inside the lift during the operation of the construction lift, make real-time load-bearing identification and decision according to the cumulative impact tilt parameters, conduct load-bearing identification on the monitoring images, and obtain load-bearing distribution; Tilt parameter calculation module, which is used to predict real-time impact tilt parameters according to the load-bearing distribution, and combine with the cumulative impact tilt parameters to calculate impact tilt parameters; Compensation level control module, which is used to conduct compensation level control analysis on the lift according to the impact tilt parameters, obtain compensation control parameters, use the cumulative impact tilt parameters for error correction, obtain a compensation control parameter range, and conduct compensation level control on the lift.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages: This application proposes an intelligent construction lift control method and system. By collecting the historical load-bearing distribution sequence, real-time monitoring images and operation data of the lift, comprehensively analyzing the cumulative impact tilt parameters, real-time load-bearing distribution and dynamic tilt trend, it combines historical data with real-time monitoring information and dynamic error correction mechanism. At the same time, based on multiple types of data such as historical load-bearing changes, real-time load distribution, and equipment structure characteristics, it conducts cumulative impact analysis of tilt, load-bearing identification and compensation control, and dynamically adjusts the control parameters according to the error correction model, effectively improving the accuracy of the analysis of the tilt state of the construction lift and the dynamic adaptability of compensation control, achieving efficient horizontal adjustment in scenarios such as uneven load-bearing and complex working conditions, and avoiding potential operation safety hazards caused by control lag and inaccurate adjustment. Through the steps of tilt trend analysis, real-time load-bearing identification and dynamic compensation control, it integrates multi-source data and evaluates multiple influencing factors, effectively avoiding control deviation problems caused by relying on fixed parameters and single data judgment. At the same time, the closed-loop feedback mechanism effectively improves the accuracy and reliability of compensation control.

[0008] The technical solution of this application realizes precise regulation of the tilt state of the construction lift by combining historical trend data, real-time monitoring data and dynamic error correction mechanism, solves the problems of untimely and inaccurate adjustment in traditional control due to relying on fixed adjustment parameters, improves the operation stability and safety of the equipment, and avoids the occurrence of equipment loss and safety accidents caused by tilt. Description of the Drawings

[0009] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the accompanying drawings required for the description of the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other accompanying drawings can be obtained based on these drawings without creative efforts.

[0010] Figure 1 It is a schematic flowchart of a control method for an intelligent construction elevator provided by an embodiment of the present application; Figure 2 It is a schematic structural diagram of an intelligent construction elevator control system provided by an embodiment of the present application; In the accompanying drawings, the components represented by each reference numeral are described as follows: Historical data inclination analysis module 01, real-time load-bearing identification and decision-making module 02, inclination parameter calculation module 03, compensation level control module 04. Detailed implementation manners

[0011] The present application provides a control method and system for an intelligent construction elevator, which are used to solve the technical problems existing in the prior art, such as the cage tilting due to uneven load-bearing when the construction elevator transports heavy objects, the traditional control method has a lag in adjustment and insufficient accuracy, obvious running shaking, and easy docking misalignment.

[0012] The following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, rather than all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present application.

[0013] In the description of the present application, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of the described features. In the description of the present application, "a plurality" means two or more, unless otherwise specifically defined.

[0014] In the description of the present application, the term "for example" is used to mean "serving as an example, illustration, or explanation". Any embodiment described as "for example" in the present application is not necessarily to be construed as more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to implement and use the present invention. In the following description, details are set forth for purposes of explanation. It should be understood that those of ordinary skill in the art can recognize that the present invention can be implemented without the use of these specific details. In other instances, well-known structures and processes are not elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is to be accorded the widest scope consistent with the principles and features disclosed in the present application.

[0015] Embodiment 1, as shown in the appendix Figure 1 The present application provides an intelligent construction elevator control method, and the method includes the following steps: S100: According to the historical operation records of the elevator, collect the historical load-bearing distribution sequence, conduct an analysis of the cumulative influence of elevator inclination, and obtain the cumulative influence inclination parameter; In the embodiment of the present application, during the operation of the construction elevator, in order to solve the problem that the inclination state evaluation does not match the actual load distribution, it is necessary to collect the historical operation records of the elevator, obtain the historical load-bearing distribution sequence, and then conduct an analysis of the cumulative influence of elevator inclination to obtain the cumulative influence inclination parameter.

[0016] Specifically, the monitoring images inside the elevator are collected in real time through an image acquisition device (such as a high-definition industrial camera) deployed inside the elevator to obtain visual data including the distribution of personnel and goods.

[0017] At the same time, according to the historical operation record database of the elevator, the load-bearing distribution data at each moment detected by devices such as weighing sensors during its operation in the historical time period is extracted, and the historical load-bearing distribution sequence is formed by arranging them in time series.

[0018] Furthermore, an analysis of the cumulative influence of elevator inclination is carried out for the historical load-bearing distribution sequence. Through a machine learning model or a statistical analysis algorithm, the historical load-bearing data is deeply analyzed to quantify the cumulative tilting effect of the uneven historical load on the elevator structure. Finally, the cumulative influence inclination parameter including the cumulative inclination direction (such as tilting left / right / forward / backward) and the cumulative inclination angle (characterizing the degree of inclination by an angle value) is obtained, providing a historical trend reference basis for subsequent real-time control.

[0019] Step S100 in the method provided by the embodiment of the present application includes: During the operation of the construction elevator, collect the monitoring images inside the elevator; According to the historical operation records of the lift, collect the historical load-bearing distribution detected during the operation of the lift within the historical time, and obtain the historical load-bearing distribution sequence; Conduct an analysis of the cumulative impact of lift inclination based on the historical load-bearing distribution sequence to obtain the cumulative impact inclination parameters, where the cumulative impact inclination parameters include the cumulative inclination direction and the cumulative inclination angle.

[0020] In the embodiments of the present application, during the operation of the construction lift, multiple high-definition industrial cameras installed in the lift car are used to collect real-time monitoring images of the personnel distribution and cargo stacking situation inside the lift.

[0021] Among them, the personnel distribution and cargo stacking situation inside the lift are collected by high-definition industrial cameras installed on the top or side wall of the car, and the real-time images are further processed using computer vision technology.

[0022] Specifically, first perform object detection to identify the positions, quantities, and dynamic distributions of personnel, perform instance segmentation to extract the contours and categories of the cargo, and then combine the preset coordinate system to convert the personnel coordinates and the center of gravity of the cargo into digital data, generating a comprehensive monitoring interface to provide data support for the inclination risk assessment.

[0023] Exemplarily, when a person enters the car, the top camera captures an image containing the person's contour in real time. The computer vision algorithm identifies the human body through an object detection model (such as YOLO) and marks the position with a bounding box. Then, through the coordinate system mapping algorithm, the two-dimensional pixel coordinates are converted into three-dimensional space coordinates (Xn, Yn, Zn) with the lower left corner of the car as the origin. At the same time, the number of people in the bounding box is counted, and the distribution situation is judged by tracking the trajectory (such as whether they gather on the left or right). Finally, the generated digital coordinate data is displayed in real time on the heat map of the control system interface.

[0024] Similarly, when the cargo enters the car, the top camera captures an image containing the cargo in real time. The computer vision algorithm identifies the cargo through an instance segmentation model (such as Mask-R-CNN) and marks the position with a contour line. Similarly, through the coordinate system mapping algorithm, the two-dimensional pixel coordinates are converted into three-dimensional space coordinates. At the same time, the floor area of the cargo enclosed by the contour line is calculated, and the cargo category is judged by analyzing the contour shape (such as boxed, bagged, irregular objects). Finally, the generated digital coordinate data and category information are displayed in real time on the distribution map of the control system interface to help the staff judge the balanced placement of the cargo.

[0025] Furthermore, according to the historical operation records of the lift, extract the load-bearing distribution data at each monitoring time point during the operation process, and arrange them in chronological order to form a historical load-bearing distribution sequence. This sequence completely records the load-bearing situation of the lift under different usage scenarios, laying a data foundation for subsequent analysis.

[0026] Optionally, data processing can also be performed on the historical load-bearing distribution sequence by combining manual verification with preset rules.

[0027] Specifically, mark the data during the sensor failure period according to the equipment maintenance record, screen out abnormal overload data according to the specified load standard, eliminate the data that does not conform to the normal operation law, and ensure the integrity and accuracy of the data through cross-checking to ensure that the historical load-bearing data meets the analysis requirements.

[0028] The step of "performing an analysis of the cumulative influence of the inclination of the lift based on the historical load-bearing distribution sequence to obtain a cumulative influence inclination parameter" in the method provided by the embodiment of the present application includes: Collect a set of sample load-bearing distribution sequences based on the operation and maintenance historical data of the construction lift, and collect the cumulative inclination direction and cumulative inclination angle of the lift when it is inclined under different sample load-bearing distribution sequences, and mark to obtain a set of sample cumulative influence inclination parameters; Construct an inclination cumulative influence analyzer for analyzing the cumulative influence of the lift inclination, wherein the inclination cumulative influence analyzer is constructed by machine learning, the input feature is the load-bearing distribution sequence, and the output feature is the cumulative influence inclination parameter; Use the set of sample load-bearing distribution sequences and the set of sample cumulative influence inclination parameters to perform supervised training on the inclination cumulative influence analyzer until convergence; Input the historical load-bearing distribution sequence into the converged inclination cumulative influence analyzer, and output to obtain the cumulative influence inclination parameter.

[0029] In the embodiment of the present application, based on the operation and maintenance historical data of the construction lift, a set of sample load-bearing distribution sequences is collected and formed, and further, the cumulative inclination direction and angle of the lift corresponding to the sequence are statistically analyzed, and marked to form a set of sample cumulative influence inclination parameters.

[0030] Among them, the collection of the sample load-bearing distribution sequence is to retrieve the historical monitoring data in the operation and maintenance database of the construction lift, extract the real-time load-bearing values of each partition of the car (such as the front, rear, left, right, and central areas) at a frequency of 1 minute interval, and arrange them in chronological order to form a sequence, so as to obtain a set of sample load-bearing distribution sequences.

[0031] For missing data, it is supplemented with the mean value of adjacent time periods, and abnormal data is corrected after cross-checking with the equipment operation log to ensure the integrity and accuracy of the sample sequence.

[0032] Further, according to the collected sample load-bearing distribution sequence, synchronously retrieve the records of the elevator inclination monitoring system, screen out the corresponding inclination events according to the sequence time period, and algebraically accumulate the direction (such as 30° north of east, 25° south of west) and inclination angle (such as 2°) of each inclination event to calculate the cumulative inclination direction and angle values corresponding to each sample load-bearing distribution sequence.

[0033] Further, label the cumulative inclination direction and cumulative inclination angle values corresponding to each sample load-bearing distribution sequence with the unique identification number of the sequence, and integrate the data into a set of sample cumulative influence inclination parameters containing three kinds of information: "sample load-bearing distribution sequence - cumulative inclination direction - cumulative inclination angle", providing intuitive data for subsequent analysis of the relationship between load-bearing distribution and elevator inclination.

[0034] Further, based on the collected set of sample load-bearing distribution sequences and the set of sample cumulative influence inclination parameters, construct an inclination cumulative influence analyzer for the cumulative influence analysis of the construction elevator inclination, with the load-bearing distribution sequence as the input feature and the cumulative influence inclination parameter as the output feature.

[0035] Among them, the analyzer is built based on the machine learning algorithm framework, with multiple linear regression as the core algorithm. First, perform data sorting and preprocessing. Normalize the load-bearing data of each region in the load-bearing distribution sequence according to a unified standard. Specifically, map the load-bearing values of each region to the interval from 0 to 1 through linear transformation to eliminate the influence caused by differences in measurement units or numerical magnitudes.

[0036] Exemplarily, the load-bearings of the left, middle, and right regions of the car are 200 kg, 500 kg, and 300 kg respectively. Among them, the maximum load-bearing is 500 kg and the minimum load-bearing is 200 kg. Calculated according to the formula (load-bearing value - minimum load-bearing) / (maximum load-bearing - minimum load-bearing), the normalized values are 0 ((200 kg - 200 kg) / (500 kg - 200 kg) = 0), 1 ((500 kg - 200 kg) / (500 kg - 200 kg) = 1), 0.33 ((300 kg - 200 kg) / (500 kg - 200 kg) = 0.33) respectively, to ensure the unified standard of the data format and provide a standardized data basis for subsequent analysis.

[0037] Then, perform model construction and training. Use the historical sample load-bearing distribution sequence as the independent variable, and the cumulative inclination direction and angle actually occurred in the elevator during the corresponding time period as the dependent variable to construct a multiple linear regression model.

[0038] During the training process, the least squares method is used to repeatedly calculate and optimize the model, and various coefficients in the model are continuously adjusted. For each iteration, the predicted results of the model are compared with the actual inclination data. According to the error situation, the coefficients are adjusted to gradually narrow the gap between the two until the model reaches the best fitting state, and finally a regression equation that can accurately reflect the relationship between load-bearing and inclination is determined, that is, the construction of the inclination cumulative impact analyzer for the analysis of the cumulative impact of the elevator inclination is completed.

[0039] Furthermore, after the construction of the inclination cumulative impact analyzer is completed, the analyzer is supervised and trained using the sample load-bearing distribution sequence set and the sample cumulative impact inclination parameter set.

[0040] Specifically, the sample load-bearing distribution sequence set is used as the training data set, and the sample cumulative impact inclination parameter set is used as the training target. By repeatedly adjusting the internal parameters of the analyzer, the prediction effect of the inclination direction and angle of the elevator is continuously optimized, and the error between the final predicted value and the actual inclination data is minimized.

[0041] During the training process, part of the data is divided as the validation data set to further test the accuracy of the analyzer's prediction. When the error value of the analyzer on the validation set is reduced to the pre-set range, and the prediction accuracy gradually stabilizes after multiple iterations without obvious fluctuations, it is determined that the analyzer training converges. At this point, the inclination cumulative impact analyzer is constructed and can be used to accurately predict the inclination of the construction elevator, providing a reliable guarantee for the safe operation of the equipment.

[0042] Furthermore, the collected historical load-bearing distribution sequence is input into the converged inclination cumulative impact analyzer, and finally accurate and reliable cumulative impact inclination parameters are obtained, providing a quantitative basis for the assessment of the safe operation status of the construction elevator.

[0043] Among them, the cumulative impact inclination parameter is a quantitative manifestation of the impact of the load-bearing distribution during the operation of the construction elevator.

[0044] Specifically, during long-term operation, the continuously changing load-bearing distribution will cause the cage to tilt, and these tilt angles accumulate continuously, forming an irreversible tilt trend. For example, the cumulative tilt reaches 3°, and the specific tilt direction is 20° south by east.

[0045] Optionally, the tilt direction and tilt angle before the elevator cage currently carries goods can also be collected as the cumulative impact inclination parameter, for example, detected by an angle sensor.

[0046] By sorting out and analyzing the historical load-bearing distribution sequence, the cumulative influence inclination parameters including the cumulative inclination angle and inclination direction are calculated comprehensively, so as to intuitively reflect the long-term influence of the load-bearing distribution on the operating state of the lift, and provide a scientific basis for equipment maintenance and risk warning.

[0047] S200: During the operation of the construction lift, collect the monitoring images inside the lift, make a real-time load-bearing recognition decision according to the cumulative influence inclination parameters, perform load-bearing recognition on the monitoring images, and obtain the load-bearing distribution. In the embodiment of the present application, during the operation of the construction lift, in order to solve the problem of load-bearing recognition deviation caused by cumulative inclination, it is necessary to comprehensively analyze by combining the cumulative influence inclination parameters and the real-time monitoring images.

[0048] Specifically, during the operation of the construction lift, high-definition industrial cameras installed on the top, bottom and four walls of the car are used to continuously collect the monitoring images inside the lift at a fixed frequency. The collected image data is transmitted to the system background in real time for preprocessing such as image enhancement and noise reduction to improve the image clarity and lay a foundation for subsequent analysis.

[0049] Since cumulative inclination will cause errors in the monitoring images and affect the accuracy of load-bearing recognition, it is necessary to analyze the cumulative influence inclination parameters. That is, by calculating the ratio of the cumulative inclination angle to the historical maximum cumulative inclination angle, the load-bearing recognition coefficient is obtained to measure the influence degree of the current inclination trend on load-bearing recognition.

[0050] Furthermore, according to this coefficient, the image analysis process is dynamically adjusted. When the inclination degree of the lift is relatively high, more parallel analysis paths are enabled to check the images from multiple angles and reduce the interference of image deviation caused by inclination on load-bearing judgment.

[0051] Finally, the load-bearing recognition coefficient is applied to the optimized image analysis process. By adjusting the analysis focus, the system pays more attention to the car area that is greatly affected by inclination, so as to accurately obtain the real-time load-bearing distribution inside the lift.

[0052] The steps in the method provided by the embodiment of the present application for S200 include: During the operation of the construction lift, collect the monitoring images inside the lift; According to the cumulative inclination angle in the cumulative influence inclination parameters, calculate the ratio to the historical maximum cumulative inclination angle, and make a decision as the load-bearing recognition coefficient; Obtain a load-bearing recognizer preconfigured with multiple load-bearing recognition branches, and randomly select the load-bearing recognition branch with the proportion of the load-bearing recognition coefficient; Input the monitored image into the selected load-bearing recognition branch, identify and output the load-bearing distributions of multiple branches, and obtain the load-bearing distribution through arithmetic averaging, where the load-bearing distribution includes the load weights at multiple positions inside the lift.

[0053] In the embodiment of the present application, during the operation of the construction lift, high-definition industrial cameras deployed on the top, bottom, and four walls of the car continuously collect monitored images of the inside of the lift at a fixed frequency (such as 25 frames per second).

[0054] Among them, the camera integrates an automatic dimming system and wide dynamic technology, which can adapt to complex environments such as high dust, strong light, and low illuminance at night on the construction site, ensuring stable operation within the working temperature range of -10°C to 50°C. At the same time, the camera uses a 120° wide-angle lens, combined with a 2K resolution and a 25fps frame rate, which can comprehensively cover the interior space of the car and accurately capture the distribution status of personnel and goods.

[0055] Furthermore, the collected image data will be transmitted to the system background in real time. The image data will be compressed using the H.264 encoding technology, which can reduce the data volume while ensuring the image clarity, making the transmission faster and smoother.

[0056] During the transmission process, image data verification is also carried out to check whether the data is complete, whether there is any loss or error. Once a problem is found, it will be promptly feedback and corrected to ensure that each image transmitted to the background is accurate, providing true and reliable data support for subsequent analysis of the car load-bearing by manual or traditional algorithms and ensuring the safe operation of the lift.

[0057] Furthermore, according to the cumulative tilt angle within the cumulative impact tilt parameter, calculate its ratio to the historical maximum cumulative tilt angle, and this ratio is used as the load-bearing recognition coefficient.

[0058] Specifically, divide the current cumulative tilt angle by the maximum cumulative tilt angle in the system historical record, and the resulting value is the load-bearing recognition coefficient. For example, if the current cumulative tilt angle is 2° and the historical maximum cumulative tilt angle is 5°, then the load-bearing recognition coefficient is 0.4 (2 / 5 = 0.4).

[0059] This coefficient is used to measure the influence degree of the current tilt trend on load-bearing recognition. The larger the coefficient, the closer the current tilt state is to the historical limit, and the more significant the impact on the load-bearing distribution. The subsequent analysis strategy can be dynamically adjusted according to the size of this coefficient, such as enhancing the monitoring intensity of the tilt direction area or increasing the sensitivity of the load limit warning threshold, providing more accurate decision-making support for the safe operation of the lift.

[0060] Furthermore, obtain a load-bearing recognizer preset with multiple load-bearing recognition branches.

[0061] The steps of "configuration of the load-bearing identifier" in the method provided by the embodiments of this application include: According to the operation data record of the construction elevator, collect a set of sample monitoring images, perform load-bearing distribution annotation based on different sample monitoring images, and obtain a set of sample load-bearing distributions, where the annotation is performed according to the number, weight, and position of people and goods in the sample monitoring images; Perform multiple divisions on the set of sample monitoring images and the set of sample load-bearing distributions to obtain multiple load-bearing recognition branch data, where there is overlapping data between every two load-bearing recognition branch data; Use a convolutional neural network to construct multiple load-bearing recognition branches, where the input data of each load-bearing recognition branch is a monitoring image and the output data is a load-bearing distribution; Respectively use multiple load-bearing recognition branch data to perform supervised training and verification on multiple load-bearing recognition branches. After convergence, complete the configuration to obtain a load-bearing identifier.

[0062] In the embodiments of this application, according to the operation data record of the construction elevator, collect multiple groups of sample monitoring images from different operation periods, load conditions, and environmental conditions to form a set of sample monitoring images.

[0063] Among them, for each sample monitoring image, organize professional personnel to perform load-bearing distribution annotation based on the actual number, weight, and spatial position information of people and goods in the image.

[0064] Specifically, estimate the total weight of people according to the average weight, calculate the weight of goods based on their types, volumes, and combined with empirical density data, and accurately mark the position coordinates of each object in the car through the mapping relationship between image coordinates and physical space, and finally form a set of sample load-bearing distributions containing weight values and spatial distribution information.

[0065] This annotation process strictly follows a unified annotation specification and quality control process to ensure the accuracy and consistency of the annotation results and provide reliable supervision data for subsequent model training.

[0066] Furthermore, adopt a combination of stratified sampling and cross-validation to perform multiple random divisions on the set of sample monitoring images and the set of sample load-bearing distributions.

[0067] Specifically, each time of division, allocate the data to different load-bearing recognition branches according to a preset ratio (such as 70% training set, 30% validation set), and ensure that some samples (about 20% - 30%) are reused in adjacent two divisions to form multiple load-bearing recognition branch data that contain both unique data and overlapping samples. In this way, multiple load-bearing recognition branch data that are both different and partially overlapping are obtained.

[0068] Through the above steps, each branch of data can cover different usage scenarios and load conditions. At the same time, overlapping data is used to verify and calibrate the analysis results, improving the accuracy and reliability of judging the load distribution under various working conditions.

[0069] Furthermore, multiple parallel load-bearing recognition and analysis branches are built based on a convolutional neural network. Each branch is equipped with an independent image processing module, a regional key analysis module, and a result calculation module.

[0070] Among them, in the image processing module, the monitored image first enters a processing channel composed of multiple convolutional layers. The first convolutional layer uses a convolutional kernel with a size of 3×3 and scans the image in a step size of 1 to initially extract edge and texture information.

[0071] Furthermore, subsequent convolutional layers gradually extract more complex features by adjusting the size and number of convolutional kernels. For example, in the third convolutional layer, a 5×5 convolutional kernel is used with a step size of 2, which can capture the general shape of the human contour and the basic geometric features of the goods. As the number of layers increases, the convolutional layer continuously extracts and abstracts features from the image, and finally outputs a feature map containing key features such as the human contour and the shape of the goods.

[0072] The regional key analysis module analyzes based on the pre-set car layout information. The coordinate ranges of key positions such as the built-in car corners and common cargo stacking areas are obtained. When the feature map is input into this module, first, according to these coordinate ranges, the image cropping technology is used to accurately cut out the key areas from the overall image to form independent local images, realizing the separation of the key areas from other parts.

[0073] Furthermore, the separated local images are processed using an enhanced contrast algorithm and edge sharpening technology. For example, for the car corner area, by increasing the local brightness and enhancing the contour contrast, the features of small goods stacked in the corner or curled-up people are made more prominent, effectively enhancing the recognition of the key areas.

[0074] After receiving the processing results output by the regional key analysis module, the fully connected layer starts the data integration and calculation process. Through a pre-established weight estimation model, combined with parameters such as the average weight of people and the corresponding relationship between the material density and volume of the goods, the feature data is converted into specific weight values.

[0075] Exemplarily, by counting the number of people in the image and multiplying by the standard average weight, the total weight of the people is obtained. According to the size measurement results of the goods in the image and combined with the density parameters of the known materials, the weight of the goods is obtained through the volume calculation formula.

[0076] Meanwhile, by using the mapping relationship between the image pixel coordinates and the actual physical space of the car, the calculated weight is accurately corresponded to each area of the car, and finally the load-bearing distribution within the entire car is determined.

[0077] Furthermore, to cope with the complex construction site environment, different parameter settings are adopted for each branch when processing image data. That is, some branches adjust the rotation angle of the image to simulate the tilted state, some branches change the image scaling ratio to test the influence of different perspectives, and there are also branches that adjust the brightness parameter to adapt to different lighting conditions.

[0078] Meanwhile, different calculation parameters are adopted for each branch to ensure the analysis of the same monitoring image from multiple dimensions. Finally, the results of each branch are summarized, and the load-bearing distribution within the elevator is comprehensively determined to provide reliable data support for safety assessment.

[0079] Furthermore, for each load-bearing recognition branch, corresponding branch data is used for independent supervised training.

[0080] Specifically, during the training process, the sample monitoring image is input into the branch analysis process and processed layer by layer according to the preset algorithm rules to calculate the preliminary load-bearing distribution prediction result.

[0081] Furthermore, this result is compared item by item with the annotation data in the sample load-bearing distribution set, and by calculating the sum of squared differences between the predicted value and the annotation value, the accuracy of the current analysis process is evaluated.

[0082] Furthermore, according to the error evaluation result, the key parameters in the analysis process (such as the threshold setting for image feature extraction, the regional weight distribution coefficient, etc.) are dynamically adjusted. For example, if it is found that the weight prediction error in a certain area is large, the weight of this area in the analysis process is specifically increased to enhance the sensitivity to the image features of this area. By repeatedly iteratively adjusting the parameters, the prediction accuracy of the analysis process for the load-bearing distribution in different scenarios is gradually improved.

[0083] In the verification stage, the reserved verification samples in the data of each branch are used to conduct a systematic performance test on the analysis process that has completed parameter training.

[0084] Specifically, the verification samples are input into the analysis process group by group, and the calculated load-bearing distribution prediction result is obtained and compared with the corresponding real load-bearing data of the samples. By calculating the mean absolute difference between the predicted value and the real value, and the square root of the sum of squared differences of all differences and other quantitative indicators, the prediction accuracy of the analysis process is measured.

[0085] Among them, the mean absolute error is used to measure the average magnitude of the deviation between the predicted value and the true value. The root mean square error further amplifies the influence weight of larger errors by taking the square root of the sum of squared errors, so as to more objectively reflect the overall error level of the analysis process.

[0086] Further, if the deviation index calculated in the above steps exceeds the pre-set error tolerance range, the operating parameters of the analysis process are optimized and adjusted. For example, appropriately modify the parameter step size during data processing, or increase the number of samples processed at one time. By continuously optimizing the parameters and increasing the number of training times until the prediction deviation of the analysis process on the validation samples converges stably within the acceptable error range, the configuration of the load identification device is completed.

[0087] In the embodiment of the present application, after obtaining the pre-configured load identification device including multiple load identification branches, according to the load identification coefficient corresponding to the currently collected monitoring image, a corresponding number of load identification branches are randomly selected in proportion.

[0088] Specifically, the load identification coefficient is mapped to the branch selection ratio. If the load identification coefficient is 0.6, 60% of the branches are randomly selected from multiple load identification branches to participate in the current load distribution identification.

[0089] Further, the real-time collected monitoring images are synchronously input into the selected load identification branches, and each branch independently performs steps such as image processing, feature extraction, weight estimation, and area mapping, and outputs the branch load distribution results respectively. These results contain the load weight data at multiple preset positions in the elevator.

[0090] Further, arithmetic averaging is performed on the load distribution results output by each branch to obtain the load distribution.

[0091] Specifically, for each preset position in the elevator, the load prediction values of all participating branches for this position are added and averaged, and finally a comprehensive load distribution result including the accurate load weights of each position is obtained.

[0092] Among them, the result is presented in the form of a numerical matrix, and each element in the matrix corresponds to the load weight at a specific position in the elevator, providing a quantitative basis for subsequent applications such as load status evaluation and tilt risk warning.

[0093] Exemplarily, assume that the load-bearing recognizer includes 5 parallel load-bearing recognition branches. When analyzing a certain monitored image, 3 of the branches (branch A, branch B, and branch C) are randomly selected to participate in the calculation. Taking the position numbered P1 in the upper left corner of the elevator car as an example, branch A predicts that the load at this position is 25 kg, branch B predicts 28 kg, and branch C predicts 23 kg. Through arithmetic averaging, the final load value at the P1 position is 25.33 kg ((25 + 28 + 23) ÷ 3 = 25.33 kg).

[0094] Similarly, the same calculation is performed on other preset positions in the car (such as the center "P5", the lower right corner "P10", etc.). After summarizing the average load values of all positions, a 10×10 numerical matrix is formed.

[0095] Among them, the element in the first row and the first column of the matrix corresponds to 25.33 kg at the "P1" position, the fifth row and the fifth column corresponds to the calculation result at the "P5" position, and so on. This matrix fully presents the load-bearing distribution state of the entire car. Subsequently, by analyzing the changes in the numerical values of each element in the matrix, the inclination hidden danger caused by uneven weight distribution in the car can be quickly located, providing data support for the safe operation of the elevator.

[0096] S300: According to the load-bearing distribution, predict and obtain real-time influence inclination parameters, and combine with the cumulative influence inclination parameters to calculate and obtain influence inclination parameters; In the embodiment of the present application, during the operation monitoring of the construction elevator, in order to solve the problem that the inclination state evaluation does not match the actual load distribution, it is necessary to comprehensively analyze the real-time load influence and the historical cumulative effect.

[0097] Specifically, based on the obtained load-bearing distribution data, through a preset mechanical model and displacement algorithm, predict and obtain real-time influence inclination parameters including the real-time inclination direction and the real-time inclination angle.

[0098] Furthermore, the real-time influence inclination parameters and the cumulative influence inclination parameters are fused and calculated. Specifically, through the vector synthesis algorithm, the angles of the real-time inclination direction and the historical cumulative inclination direction are synthesized, and at the same time, the method of cumulative summation is used to dynamically fuse the real-time inclination angle and the cumulative inclination angle. Finally, the influence inclination parameters including the cumulative inclination direction and the cumulative inclination angle are obtained. This parameter comprehensively reflects the inclination trend and degree of the elevator under the combined action of the current load and the historical load.

[0099] The step S300 in the method provided by the embodiment of the present application includes: Input the load distribution into a real-time impact tilt analyzer to output real-time impact tilt parameters. Among them, the real-time impact tilt analyzer is trained using a sample load distribution set and a sample real-time impact tilt parameter set. The real-time impact tilt parameters include a real-time tilt direction and a real-time tilt angle. Calculate the impact tilt parameters by cumulative calculation based on the real-time impact tilt parameters and the cumulative impact tilt parameters. Among them, the impact tilt parameters include an accumulated tilt direction and an accumulated tilt angle.

[0100] In the embodiments of the present application, in order to accurately obtain the immediate impact of the current load distribution of the lift on the tilt state, the load distribution is input into a pre-trained real-time impact tilt analyzer, and real-time impact tilt parameters including a real-time tilt direction and a real-time tilt angle are output.

[0101] Among them, the real-time impact tilt analyzer is obtained through machine learning training on a sample load distribution set and a sample real-time impact tilt parameter set, and can dynamically analyze the tilt trend characteristics of the lift under the current working conditions based on the input load distribution data, providing a real-time quantitative basis for subsequent comprehensive evaluation.

[0102] Specifically, the input data is first standardized. That is, the sample load distribution is divided into a 10×10 grid matrix according to the actual physical area of the lift car, and each grid unit corresponds to the load value of a specific area inside the car.

[0103] At the same time, the sample real-time impact tilt parameters are normalized and converted. The tilt angle is uniformly converted to the ratio interval of 0 to 1, and the tilt direction is represented in the form of a unit vector in the Cartesian coordinate system to ensure the consistency and computability of the data format.

[0104] Exemplarily, in a certain monitoring, the actual tilt angle of the lift is 5°. During the normalization conversion, since the maximum designed tilt angle of the lift is 10°, 5° is proportionally converted to 0.5 (5° / 10° = 0.5); if the tilt direction is 10° east by south, with the origin of the Cartesian coordinate system as the reference, the due east direction as the positive X-axis direction, and the due north direction as the positive Y-axis direction, according to the trigonometric function relationship, the X-axis component of the unit vector is calculated as cos(30°)≈0.866, and the Y-axis component is -sin(30°)=-0.5, so the tilt direction is represented as [0.866, -0.5].

[0105] Furthermore, a three-level cascaded image processing structure is used for feature extraction. In the first layer, a 5×5 sized convolution template is used to scan the load distribution matrix to capture the weight distribution characteristics of a larger area.

[0106] Further, the convolution templates of the latter two layers are gradually reduced to 3×3 to focus on extracting local detail features. By increasing the number of feature channels to 64, multi-dimensional extraction of the load-bearing distribution features is achieved.

[0107] Further, in the result output stage, the conversion from feature data to tilt parameters is realized through three layers of data processing modules. Each level of the processing module contains multiple data operation units, which integrate and calculate the feature data step by step. Finally, the X and Y axis components of the tilt direction in the Cartesian coordinate system, as well as the tangent value of the tilt angle, are obtained respectively in the output module.

[0108] During the training process, the mean square error is used as the evaluation index, and certain weight coefficients (such as 0.6 and 0.4) are assigned to the direction error and the angle error respectively. The calculation parameters are optimized through an adaptive parameter adjustment algorithm, and combined with the random data perturbation test simulating different working conditions, the stability and adaptability of the system are improved.

[0109] Further, after multiple rounds of iterative cross-validation tests, if the angle prediction error of the analyzer is controlled within a predefined threshold (such as 0.3 degrees), and the direction prediction error is also controlled within a predefined threshold (such as 2 degrees), then a real-time impact tilt analyzer is successfully constructed.

[0110] Further, by inputting the load-bearing distribution into the pre-trained real-time impact tilt analyzer, the real-time impact tilt parameters of the real-time tilt direction and the real-time tilt angle can be output and obtained, which can meet the requirements of engineering practical applications.

[0111] Further, after obtaining the real-time impact tilt parameters and the cumulative impact tilt parameters, an accumulative calculation method is used for comprehensive calculation to obtain the impact tilt parameters, that is, the accumulative tilt direction and the accumulative tilt angle are obtained respectively.

[0112] Specifically, for the tilt direction, the unit vector corresponding to the real-time tilt direction and the unit vector corresponding to the cumulative tilt direction are vectorially added to obtain the unit vector of the accumulative tilt direction, and the direction angle of this vector is the accumulative tilt direction.

[0113] Exemplarily, if the real-time tilt direction is 30° south of east (unit vector is [0.866, -0.5], where the X-axis component is cos(30°) ≈ 0.866 and the Y-axis component is -sin(30°) = -0.5), and the cumulative tilt direction is 20° north of east (unit vector is [0.940, 0.342], X-axis component is cos(20°) ≈ 0.940, Y-axis component is sin(20°) = 0.342), then the vector sum is [1.806, -0.158] (0.866 + 0.940 = 1.806, -0.5 + 0.342 = -0.158), that is, the unit vector of the cumulative direction is approximately [1.806, -0.158], and the corresponding direction is about 5° south of east.

[0114] For the tilt angle, it is necessary to consider the immediate action of the real-time load and the cumulative effect of the historical load, and calculate through the formula "cumulative tilt angle = real-time tilt angle + cumulative tilt angle".

[0115] Exemplarily, if the real-time tilt angle is 2.5° and the cumulative tilt angle is 1.2°, then the cumulative tilt angle = 2.5 + 1.2 = 3.7°.

[0116] The obtained influence tilt parameters through comprehensive analysis include the cumulative tilt direction and the cumulative tilt angle, which can comprehensively reflect the comprehensive tilt state of the lift under the combined action of the current and historical loads, and provide accurate data support for safety warning.

[0117] S400: According to the influence tilt parameters, perform compensation level control analysis on the lift, obtain compensation control parameters, use the cumulative influence tilt parameters for error correction, obtain the compensation control parameter interval, and perform compensation level control on the lift.

[0118] In the embodiments of the present application, during the operation monitoring of the construction lift, in order to solve the operation safety problem caused by tilt, it is necessary to perform precise compensation level control on the lift according to the influence tilt parameters.

[0119] Specifically, according to the obtained influence tilt parameters, and in combination with the structural parameters and operation characteristics of the lift, perform compensation level control analysis on the lift.

[0120] Furthermore, considering the errors in the calculation process and the complexity of the actual working conditions, use the cumulative influence tilt parameters for error correction. By analyzing the deviation between the historical tilt data and the actual compensation effect, correct and adjust the compensation control parameters, and finally obtain the compensation control parameter interval.

[0121] This interval covers reasonable values of control parameters within different error ranges, providing both an operating boundary for the automated control system and a reference for manual intervention by control personnel, thereby achieving effective compensation level control of the lift and ensuring its stable and safe operation.

[0122] Step S400 in the method provided by the embodiments of this application includes: Generating compensation control parameters for compensating level control of the lift according to the influencing tilt parameters; Processing to obtain an error correction coefficient based on the load-bearing identification coefficient calculated from the cumulative influencing tilt parameters, and performing error correction on the compensation control parameters to obtain a compensation control parameter interval; Performing compensation level control on the lift according to the compensation control parameter interval.

[0123] In the embodiments of this application, during the operation monitoring of the construction lift, to ensure its horizontal stable state, it is necessary to generate compensation control parameters for compensating level control of the lift according to the influencing tilt parameters to achieve precise regulation.

[0124] Specifically, according to the cumulative tilt direction and cumulative tilt angle included in the influencing tilt parameters, combined with the mechanical structure parameters of the lift (such as the distance between support legs, the height of the column, and the stroke range of the hydraulic system) and the load-bearing characteristics (rated load capacity, center of gravity distribution law), calculations are carried out through a preset compensation control algorithm.

[0125] Among them, this algorithm is based on the static equilibrium equation, maps the tilt direction to the offset in the X and Y axis directions, and uses trigonometric function relationships (such as the tangent function) to convert the tilt angle into the height difference that needs to be adjusted at each support point.

[0126] Exemplarily, when the detected cumulative tilt angle is 2°, and the known lateral distance between support legs is 3m, through the formula "height difference = distance × tan(tilt angle)", it is calculated that the height difference that needs to be adjusted at the two support points is approximately 105mm (3000×tan2° = 105). At the same time, according to the tilt direction, the lifting relationship of each support point is determined. For example, when tilted east-southeast, the height of the southeast support point needs to be lowered and the height of the northwest support point needs to be raised.

[0127] The compensation control parameters generated through the calculation process of the above steps cover key information such as the adjustment amplitude (specific lifting height value) and adjustment direction (raising or lowering) of each support point or lifting mechanism of the lift, providing an accurate quantitative basis for subsequent compensation control operations, thereby achieving effective adjustment of the horizontal state of the lift.

[0128] Furthermore, after obtaining the preliminary compensation control parameters, due to certain errors in the practical application of image recognition technology, there may be deviations between the measured results of the tilt parameters and the actual situation. To ensure the accuracy and reliability of the compensation control, error correction needs to be carried out based on the cumulative impact on the tilt parameters.

[0129] Specifically, based on the load-bearing recognition coefficient calculated according to the cumulative impact on the tilt parameters in the foregoing content, error correction is carried out on the compensation control parameters to obtain the compensation control parameter range. For example, if the load-bearing recognition coefficient is 0.4, then calculate the compensation control parameter *(1 ± 0.4) as the compensation control parameter range. For instance, 105mm *(1 ± 0.4) is 63 - 147mm. Within this range, the cage height is adjusted in the opposite direction of the tilt direction to make the cage level.

[0130] The finally formed compensation control parameter range provides a clear operation basis for the control personnel. The control personnel can flexibly select appropriate control parameters within this range according to the actual situation to perform compensation control operations, thereby effectively improving the accuracy and safety of the horizontal control of the construction elevator.

[0131] Furthermore, according to the compensation control parameter range, combined with the real-time operating state of the elevator, the optimal control parameter is dynamically selected within the parameter range to drive the actuator to perform compensation level control.

[0132] Specifically, the current height, tilt angle, and load data of each support point of the elevator collected in real time can be used to compare and analyze these data with the theoretical values within the compensation control parameter range, so as to determine the best adjustment strategy under the current working conditions.

[0133] Specifically, if it is monitored that the elevator tilts towards the southeast direction and the tilt angle is close to the upper limit of the parameter range, the hydraulic control system will be preferentially triggered. In the direction of "lowering the height of the southeast support point and raising the height of the northwest support point", based on the median value of the parameter range, the adjustment amplitude is dynamically adjusted in combination with the real-time load (for example, in the heavy load condition, the adjustment amplitude is taken as 1.2 times the median value of the range to enhance the compensation effect).

[0134] Furthermore, the change rate of the tilt angle during the synchronous monitoring and adjustment process will be monitored. If abnormal fluctuations are found (such as changing more than 0.5° per second), the adjustment will be automatically paused and a safety warning will be triggered to prevent potential safety hazards caused by equipment overload or sudden structural stress changes.

[0135] Furthermore, during the execution of the compensation control, in order to make the control more accurate, a closed-loop feedback method will be used to continuously optimize the control parameters.

[0136] Specifically, after each adjustment operation is completed, the deviation between the actual inclination angle and the target value will be compared. If the deviation exceeds the allowable range (such as ±0.2°), then according to the deviation direction and magnitude, the adjustment amplitude for the next time will be finely adjusted within the compensation control parameter range (such as reducing or enlarging the adjustment amount by a proportional coefficient of 0.8) until the horizontal state of the lift meets the safety operation standard.

[0137] Through this method of dynamic adaptive adjustment, not only can the accuracy of compensation control be ensured, but also the problem of adjustment lag that may be caused by single fixed parameter adjustment can be effectively avoided, and the intelligent and safe control of the horizontal state of the construction lift can be effectively realized.

[0138] Through the above specific implementation manners, the embodiments of the present application achieve the following technical effects: An intelligent construction lift control method provided by an embodiment of the present application is of great significance for improving the intelligent level of construction machinery. First, by collecting the historical load-bearing distribution sequence of the lift and using machine learning or statistical analysis algorithms to carry out the cumulative influence analysis of inclination, the cumulative influence inclination parameters including direction and angle are obtained, solving the problem that traditional control lacks historical trend reference and providing data support for precise control; secondly, during operation, by collecting and monitoring images in real time, combining with the cumulative inclination parameters to calculate the load-bearing recognition coefficient, and analyzing and integrating the results through a multi-branch load-bearing recognizer to obtain the real-time load-bearing distribution, improving the data accuracy; then inputting the load-bearing distribution into the real-time influence inclination analyzer to predict the real-time parameters, and generating comprehensive influence inclination parameters after fusing with the cumulative parameters, comprehensively reflecting the inclination state under the action of current and historical loads and avoiding the one-sidedness of a single data dimension; finally, generating preliminary compensation control parameters, obtaining a parameter range after correcting by using the cumulative parameters to construct an error correction model, dynamically selecting parameters and combining with closed-loop feedback to optimize the control, solving the problems of adjustment lag and large error of traditional fixed parameter adjustment, and ensuring the stability of the horizontal state of the lift.

[0139] The intelligent construction lift control method provided by the embodiment of the present application realizes the precise analysis and intelligent compensation of the inclination state, effectively solves the inclination problem caused by uneven load-bearing, improves the operation stability and adjustment accuracy of the equipment, reduces potential safety hazards, ensures construction safety and efficiency, and provides an innovative solution for the reliable operation of the lift in a complex environment.

[0140] Embodiment 2, as shown in the appendix Figure 2 Based on the inventive concept of an intelligent construction lift control method provided in Embodiment 1, the present application further provides an intelligent construction lift control system, specifically including: A historical data inclination analysis module 01, configured to collect a historical load-bearing distribution sequence according to the historical operation records of the lift, and perform a cumulative influence analysis of the lift inclination to obtain cumulative influence inclination parameters; The real-time load-bearing identification and decision-making module 02 is used to collect the monitoring images inside the construction elevator during its operation, make real-time load-bearing identification and decision according to the cumulative influence inclination parameters, perform load-bearing identification on the monitoring images, and obtain the load-bearing distribution; The inclination parameter calculation module 03 is used to predict and obtain the real-time influence inclination parameters according to the load-bearing distribution, and calculate the influence inclination parameters by combining the cumulative influence inclination parameters; The compensation level control module 04 is used to perform compensation level control analysis on the elevator according to the influence inclination parameters, obtain the compensation control parameters, perform error correction using the cumulative influence inclination parameters, obtain the compensation control parameter interval, and perform compensation level control on the elevator.

[0141] In one embodiment, the historical data inclination analysis module 01 is further used for: Collect the monitoring images inside the construction elevator during its operation; According to the historical operation records of the elevator, collect the historical load-bearing distribution detected during the operation of the elevator within the historical time, and obtain the historical load-bearing distribution sequence; Perform elevator inclination cumulative influence analysis according to the historical load-bearing distribution sequence, and obtain the cumulative influence inclination parameters, where the cumulative influence inclination parameters include the cumulative inclination direction and the cumulative inclination angle.

[0142] In one embodiment, the real-time load-bearing identification and decision-making module 02 is further used for: Collect the monitoring images inside the construction elevator during its operation; Calculate the ratio to the historical maximum cumulative inclination angle according to the cumulative inclination angle in the cumulative influence inclination parameters, and make a decision as the load-bearing identification coefficient; Obtain a load-bearing identifier including multiple load-bearing identification branches configured in advance, and randomly select the load-bearing identification branches accounting for the ratio of the load-bearing identification coefficient; Input the monitoring images into the selected load-bearing identification branches, identify and output multiple branch load-bearing distributions, and obtain the load-bearing distribution through arithmetic averaging, where the load-bearing distribution includes the load weights at multiple positions inside the elevator.

[0143] In one embodiment, the inclination parameter calculation module 03 is further used for: Input the load-bearing distribution into the real-time influence inclination analyzer, and output and obtain the real-time influence inclination parameters, where the real-time influence inclination analyzer is obtained by training with a sample load-bearing distribution set and a sample real-time influence inclination parameter set, and the real-time influence inclination parameters include the real-time inclination direction and the real-time inclination angle; Based on the real-time influence tilt parameter and the cumulative influence tilt parameter, the influence tilt parameter is obtained through cumulative calculation, wherein the influence tilt parameter includes an accumulated tilt direction and an accumulated tilt angle.

[0144] In one embodiment, the compensation level control module 04 is further configured to: Generate a compensation control parameter for compensating and leveling the elevator according to the influence tilt parameter; Process the load-bearing identification coefficient obtained by calculating according to the cumulative influence tilt parameter to obtain an error correction coefficient, and perform error correction on the compensation control parameter to obtain a compensation control parameter range; Perform compensation and leveling control on the elevator according to the compensation control parameter range.

[0145] It should be noted that the above sequence of embodiments of the present application is only for description and does not represent the superiority or inferiority of the embodiments. And the above describes specific embodiments of this specification. Additionally, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0146] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included within the protection scope of the present application.

[0147] This specification and the drawings are only exemplary descriptions of the present application and are considered to have covered any and all modifications, variations, combinations, or equivalents within the scope of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application.

Claims

1. An intelligent construction hoist control method, characterized in that The method includes: Collecting a historical load-bearing distribution sequence according to the historical operation record of the lift, performing an analysis of the cumulative influence of the lift tilt, and obtaining a cumulative influence tilt parameter; During the operation of the construction lift, collecting the monitoring images inside the lift, making a real-time load-bearing identification decision according to the cumulative influence tilt parameter, performing load-bearing identification on the monitoring images, and obtaining a load-bearing distribution; Predicting a real-time influence tilt parameter according to the load-bearing distribution, combining with the cumulative influence tilt parameter, and calculating an influence tilt parameter; Performing a compensation level control analysis on the lift according to the influence tilt parameter to obtain a compensation control parameter, using the cumulative influence tilt parameter for error correction to obtain a compensation control parameter interval, and performing a compensation level control on the lift.

2. The intelligent construction elevator control method according to claim 1, characterized in that During the operation of the construction lift, collecting the monitoring images inside the lift, and according to the historical operation record of the lift, collecting a historical load-bearing distribution sequence, performing an analysis of the cumulative influence of the lift tilt, and obtaining a cumulative influence tilt parameter, including: During the operation of the construction lift, collecting the monitoring images inside the lift; According to the historical operation record of the lift, collecting the historical load-bearing distribution detected during the operation of the lift within the historical time to obtain a historical load-bearing distribution sequence; Performing an analysis of the cumulative influence of the lift tilt according to the historical load-bearing distribution sequence to obtain a cumulative influence tilt parameter, where the cumulative influence tilt parameter includes a cumulative tilt direction and a cumulative tilt angle.

3. The intelligent construction elevator control method according to claim 2, wherein Performing an analysis of the cumulative influence of the lift tilt according to the historical load-bearing distribution sequence to obtain a cumulative influence tilt parameter, including: Collecting a set of sample load-bearing distribution sequences according to the operation and maintenance historical data of the construction lift, and collecting the cumulative tilt direction and cumulative tilt angle of the lift tilt under different sample load-bearing distribution sequences, and labeling to obtain a set of sample cumulative influence tilt parameters; Constructing a tilt cumulative influence analyzer for analyzing the cumulative influence of the lift tilt, where the tilt cumulative influence analyzer is constructed by machine learning, the input feature is the load-bearing distribution sequence, and the output feature is the cumulative influence tilt parameter; Using the set of sample load-bearing distribution sequences and the set of sample cumulative influence tilt parameters to perform supervised training on the tilt cumulative influence analyzer until convergence; Inputting the historical load-bearing distribution sequence into the converged tilt cumulative influence analyzer, and outputting to obtain a cumulative influence tilt parameter.

4. The intelligent construction elevator control method according to claim 1, characterized in that, During the operation of the construction lift, collecting the monitoring images inside the lift, making a real-time load-bearing identification decision according to the cumulative influence tilt parameter, performing load-bearing identification on the monitoring images, and obtaining a load-bearing distribution, including: During the operation of the construction lift, collecting the monitoring images inside the lift; Calculating the ratio to the historical maximum cumulative tilt angle according to the cumulative tilt angle in the cumulative influence tilt parameter, and making a decision as a load-bearing identification coefficient; Obtaining a load-bearing identifier including a plurality of load-bearing identification branches configured in advance, and randomly selecting the load-bearing identification branch with the proportion of the load-bearing identification coefficient; Input the monitored image into the selected load-bearing recognition branch, recognize and output the load-bearing distributions of multiple branches, and obtain the load-bearing distribution through arithmetic averaging. The load-bearing distribution includes the load weights at multiple positions inside the elevator.

5. The intelligent construction elevator control method according to claim 4, characterized in that, The configuration steps of the load-bearing recognizer include: Collect a set of sample monitored images according to the operation data record of the construction elevator, and perform load-bearing distribution annotation based on different sample monitored images to obtain a set of sample load-bearing distributions, where the annotation is performed according to the number, weight, and position of people and goods in the sample monitored images; Perform multiple divisions on the set of sample monitored images and the set of sample load-bearing distributions to obtain multiple load-bearing recognition branch data, where there is overlapping data between every two load-bearing recognition branch data; Use a convolutional neural network to construct multiple load-bearing recognition branches, where the input data of each load-bearing recognition branch is a monitored image and the output data is a load-bearing distribution; Use multiple load-bearing recognition branch data respectively to perform supervised training and verification on multiple load-bearing recognition branches. After convergence, complete the configuration to obtain a load-bearing recognizer.

6. The intelligent construction elevator control method according to claim 1, characterized in that Based on the load-bearing distribution, predict to obtain real-time impact tilt parameters, and combine with the cumulative impact tilt parameters to calculate the impact tilt parameters, including: Input the load-bearing distribution into a real-time impact tilt analyzer to output real-time impact tilt parameters. The real-time impact tilt analyzer is trained using the set of sample load-bearing distributions and the set of sample real-time impact tilt parameters. The real-time impact tilt parameters include a real-time tilt direction and a real-time tilt angle; Accumulatively calculate the impact tilt parameters based on the real-time impact tilt parameters and the cumulative impact tilt parameters. The impact tilt parameters include an accumulative tilt direction and an accumulative tilt angle.

7. The intelligent construction elevator control method according to claim 1, characterized in that Based on the impact tilt parameters, perform compensation level control analysis on the elevator to obtain compensation control parameters, and perform error correction using the cumulative impact tilt parameters to obtain a compensation control parameter range, and perform compensation level control on the elevator, including: Generate compensation control parameters for compensating the level control of the elevator according to the impact tilt parameters; Process to obtain an error correction coefficient based on the load-bearing recognition coefficient calculated from the cumulative impact tilt parameters, and perform error correction on the compensation control parameters to obtain a compensation control parameter range; Perform compensation level control on the elevator according to the compensation control parameter range.

8. An intelligent construction elevator control system, characterized in that, The system is used to execute the intelligent construction elevator control method according to any one of claims 1-7. The system includes: A historical data tilt analysis module, used to collect a historical load-bearing distribution sequence according to the historical operation record of the elevator, perform an analysis of the cumulative impact of the elevator tilt, and obtain cumulative impact tilt parameters; A real-time load-bearing recognition decision module, used to collect a monitored image inside the construction elevator during the operation of the construction elevator, make a real-time load-bearing recognition decision according to the cumulative impact tilt parameters, and perform load-bearing recognition on the monitored image to obtain a load-bearing distribution; A tilt parameter calculation module, used to predict real-time impact tilt parameters based on the load-bearing distribution, and combine with the cumulative impact tilt parameters to calculate the impact tilt parameters; The compensation level control module is used to perform compensation level control analysis on the lift according to the influencing inclination parameter, obtain compensation control parameters, perform error correction using the cumulative influencing inclination parameter, obtain a compensation control parameter interval, and perform compensation level control on the lift.

9. The intelligent construction elevator control system according to claim 8, wherein, The historical data inclination analysis module is further used for: During the operation of the construction lift, collect the monitoring images inside the lift; According to the historical operation records of the lift, collect the historical load-bearing distribution detected during the operation of the lift within the historical time, and obtain a historical load-bearing distribution sequence; Perform cumulative influence analysis of lift inclination according to the historical load-bearing distribution sequence, and obtain a cumulative influencing inclination parameter, where the cumulative influencing inclination parameter includes a cumulative inclination direction and a cumulative inclination angle.

10. The intelligent construction elevator control system according to claim 8, wherein The real-time load-bearing identification decision module is further used for: During the operation of the construction lift, collect the monitoring images inside the lift; Calculate the ratio with the historical maximum cumulative inclination angle according to the cumulative inclination angle in the cumulative influencing inclination parameter, and make a decision as the load-bearing identification coefficient; Obtain a load-bearing identifier including multiple load-bearing identification branches configured in advance, and randomly select the load-bearing identification branch with the proportion of the load-bearing identification coefficient; Input the monitoring image into the selected load-bearing identification branch, identify and output multiple branch load-bearing distributions, and obtain the load-bearing distribution through arithmetic averaging, where the load-bearing distribution includes the load weights at multiple positions inside the lift.

Images