Intelligent trolley monitoring method and system based on cloud platform

Abstract

The invention relates to an intelligent trolley monitoring method and system based on a cloud platform, and the method comprises the steps: collecting multi-source data of a trolley in a construction process, and the multi-source data comprises sensor data and video monitoring data; preprocessing the multi-dimensional data through edge computing equipment, and uploading the preprocessed data to a cloud platform for centralized processing and storage; the preprocessed data are analyzed, the construction state of the trolley is recognized, the construction state analysis result of the trolley is generated, and the construction state comprises the trolley operation state, the concrete pouring state and the lining quality; and dynamically adjusting construction parameters of the trolley according to a construction state analysis result of the trolley. The construction method has the effect of improving the construction quality.

Description

Technical Field

[0001] The present application relates to the technical field of tunnel construction, and in particular, to a trolley intelligent monitoring method and system based on a cloud platform. Background Technique

[0002] During the construction of underground structures such as tunnels and culverts, concrete lining is a key link to ensure the structural stability and durability. As an important construction equipment for supporting the formwork and assisting in concrete pouring, the operating state, formwork positioning accuracy, concrete vibration quality, etc. of the trolley directly affect the forming effect of the lining structure and the overall construction quality.

[0003] In the prior art, the concrete pouring and trolley construction processes mainly rely on manual operation and empirical judgment for control and monitoring. For example, operators need to judge based on experience whether the pouring is full, whether the formwork is deformed, and whether the vibration is sufficient. At the same time, the on-site status is collected through manual inspection, which has great subjectivity and lag.

[0004] The above prior art solutions have the following defects: Since the concrete pouring process is usually inside the formwork, construction workers cannot directly observe the state of the concrete and cannot accurately judge whether the pouring is sufficient, resulting in quality problems being difficult to discover and handle in a timely manner. Therefore, there is room for improvement. Summary of the Invention

[0005] In order to improve the construction quality, the present application provides a trolley intelligent monitoring method and system based on a cloud platform.

[0006] The first invention object of the present application is achieved through the following technical solutions: A trolley intelligent monitoring method based on a cloud platform, the trolley intelligent monitoring method based on a cloud platform includes: Collect multi-source data during the construction of the trolley, the multi-source data including sensor data and video monitoring data; Preprocess the multi-dimensional data through an edge computing device, and upload the preprocessed data to the cloud platform for centralized processing and storage; Analyze the preprocessed data, identify the construction state of the trolley, and generate an analysis result of the construction state of the trolley, the construction state including the trolley operating state, concrete pouring state, and lining quality; Dynamically adjust the construction parameters of the trolley according to the analysis result of the construction state of the trolley.

[0007] By adopting the above technical solutions, by collecting multi-source data of the trolley during construction, multi-dimensional information including sensor data and video monitoring data can be obtained, realizing a comprehensive perception of the construction site status, thus avoiding monitoring blind spots caused by relying solely on a single data source and improving the coverage and perception accuracy of construction monitoring; by preprocessing the multi-dimensional data through edge computing devices and uploading the preprocessed data to the cloud platform for centralized processing and storage, the real-time performance and response speed of data processing can be improved, reducing the cloud load, thereby enhancing the processing efficiency and stability of the system; by analyzing the preprocessed data to identify the construction status of the trolley and generating a construction status analysis result, real-time status determination of the entire construction process can be achieved, thus promptly discovering problems such as construction deviations and quality risks and ensuring the lining construction quality; by dynamically adjusting construction parameters according to the construction status analysis result, a closed-loop regulation system of monitoring-identifying-controlling can be formed, thereby realizing the intelligent and adaptive regulation of the concrete lining trolley construction and improving the construction quality and efficiency.

[0008] In one example, the present application can be further configured as: the multi-source data of the trolley during construction includes: Detecting the concrete pouring pressure through a pressure sensor and determining the lining integrity; Monitoring the concrete flow rate using an ultrasonic flowmeter to obtain the concrete pouring flow rate; Detecting the concrete pouring temperature using an infrared temperature sensor; Real-time monitoring of the radial distance between the trolley formwork of the trolley and the concrete through a laser displacement sensor to obtain formwork displacement data; Monitoring the working status of the vibrator of the trolley through a MEMS vibration sensor to obtain vibration parameters; Collecting video data of the construction process of the trolley using an industrial camera to obtain the video monitoring data.

[0009] By adopting the above technical solutions, by using a pressure sensor to detect the concrete pouring pressure and determine the integrity of the lining, it is possible to judge whether there are risks such as hollowing and leakage during the lining process, and thus take timely measures such as supplementary pouring to ensure the structural density; by using an ultrasonic flowmeter to monitor the concrete flow rate, it is possible to grasp the pouring rate in real time, prevent structural non-uniformity caused by too fast or too slow pouring, and thus ensure the uniformity of pouring; by using an infrared temperature sensor to detect the concrete temperature when it enters the formwork, it is possible to achieve temperature control monitoring, prevent early cracks caused by too large a temperature difference, and thus improve the consistency of concrete strength development; by using a laser displacement sensor to continuously monitor the radial distance between the formwork and the concrete, it is possible to evaluate the displacement or deformation degree of the formwork, and thus timely correct the formwork positioning to ensure the lining dimension accuracy; by using a MEMS vibration sensor to monitor the working state of the vibrator, it is possible to judge whether the vibration is sufficient, and thus effectively avoid cavities and structural damage caused by missed vibration or over-vibration; by using an industrial camera to collect construction videos, it is possible to visually master the construction operation behavior, and thus provide an intuitive basis for quality traceability and safety supervision.

[0010] In one example, the present application can be further configured as: the preprocessing of the multi-dimensional data by the edge computing device and the uploading of the preprocessed data to the cloud platform for centralized processing and storage include: Adopting a data filtering algorithm to perform noise reduction processing on the sensor data, and performing correlation calculation on the noise-reduced data to generate a preliminary evaluation result of the construction state; Performing object detection on the video monitoring data to identify construction anomalies; Combining the sensing data and the video monitoring data, performing a preliminary statistics on the construction state of the trolley to obtain the preprocessed data and mark potential risk points.

[0011] By adopting the above technical solutions, by using a data filtering algorithm to perform noise reduction processing on the sensor data and performing correlation calculation, it is possible to eliminate invalid fluctuations or interference information in the data, and thus improve the accuracy and usability of the monitoring data; by performing object detection on the video monitoring data to identify construction anomalies, it is possible to identify visual anomalies such as formwork offset and concrete splashing from the image level, and thus make up for the deficiency of the sensor data in the visible range; by combining the sensing data and the video monitoring data to perform preliminary statistics and mark potential risk points, it is possible to realize the fusion judgment of multi-modal perception data, and thus enhance the system's ability to identify abnormal states and provide a risk basis for subsequent analysis.

[0012] In one example, the present application can be further configured as: the performing of correlation calculation on the noise-reduced data to generate a preliminary evaluation result of the construction state includes: Perform time synchronization on the data after noise reduction processing, and use the interpolation algorithm to fill in the missing data to obtain the corrected data; According to the sensor layout positions of the trolley, construct a spatial mapping model of the construction area of the trolley, perform spatial projection on the corrected data, adopt a weighted calculation method to calculate the weight contribution values of each construction parameter, obtain the weighted calculation result, and dynamically adjust the weights of each construction parameter according to the construction environment, concrete setting characteristics and trolley state of the trolley; Perform correlation analysis on the video monitoring data and the weighted calculation result to generate a preliminary evaluation result of the construction state.

[0013] By adopting the above technical solution, by performing time synchronization on the data after noise reduction processing and using the interpolation algorithm to fill in the missing data, the time axis of each data source can be unified and the data continuity can be restored, thereby improving the overall integrity of the data and the timeliness of comparative analysis; by constructing a spatial mapping model according to the sensor layout positions and performing spatial projection, the spatial correlation between the collected data and the physical structure of the trolley can be realized, thereby accurately judging the specific location where the abnormality occurs and improving the positioning accuracy; by adopting the weighted calculation method to calculate the weight contribution values of the construction parameters and dynamically adjusting the weights of each parameter according to the construction environment, concrete setting characteristics and trolley state, the dynamic adaptability of the data evaluation model can be realized, thereby maintaining the accuracy and reliability of the analysis under different construction conditions; by performing correlation analysis on the weighted calculation result and the video monitoring data to generate a preliminary evaluation result, a mapping relationship can be established between the structured data and the visual information, thereby improving the credibility and accuracy of the overall evaluation and judgment.

[0014] In one example of this application, it can be further configured that: the target detection of the video monitoring data and the identification of construction abnormal situations include: Use the YOLO algorithm to identify image abnormalities at the construction site; Combine voice recognition to analyze the voice commands of construction workers and judge whether the voice commands conform to the operation process; Adopt lidar scanning technology to construct a three-dimensional point cloud model of the construction environment where the trolley is located, and detect whether the trolley deviates from the predetermined trajectory or physical interference occurs.

[0015] By adopting the above technical solutions, by using the YOLO algorithm to identify anomalies in the construction site images, events such as concrete splashing and formwork offset can be quickly located in the video images, thus improving the intelligent level and efficiency of visual monitoring; by combining speech recognition to analyze the voice commands of construction workers to determine whether they conform to the operation process, the manual operation process can be intelligently audited, thus assisting in judging whether the construction process is standard and helping to improve the automation degree of on-site management; by using lidar scanning to construct a three-dimensional point cloud model and detecting whether the trolley deviates from the predetermined trajectory or interferes, the operation path of the equipment can be accurately monitored, thus ensuring the safety and construction accuracy of the trolley in a limited space.

[0016] In one example, the present application can be further configured as: analyzing the preprocessed data to identify the construction state of the trolley and generating the construction state analysis result of the trolley includes: Extracting features from the preprocessed data to obtain key parameters related to the construction state, where the key parameters include the uniformity of concrete pouring, vibration effect, integrity of the lining structure, and environmental stability; Inputting the key parameters into a pre-trained decision tree algorithm for pattern recognition to determine the construction state of the trolley and generating the construction state analysis result, where the construction state includes the trolley operation state, concrete pouring state, and lining quality, and the construction state analysis result is used to drive the real-time optimization control of construction parameters.

[0017] By adopting the above technical solutions, by extracting features from the preprocessed data to obtain key parameters such as the uniformity of concrete pouring, vibration effect, integrity of the lining structure, and environmental stability, the most critical quality indicators in the construction process can be comprehensively extracted, thus providing rich basic features for subsequent construction state judgment; by inputting the key parameters into a pre-trained decision tree algorithm for pattern recognition to determine the construction state, it can quickly determine whether it is in a normal, abnormal, or warning state through the trained model, thus improving the recognition efficiency and reducing the cost of manual intervention; by using the construction state analysis result to drive the real-time optimization control of construction parameters, a closed-loop feedback of analysis - decision - control can be achieved, thus significantly improving the intelligent regulation ability of the system and the overall construction quality.

[0018] In one example, the present application can be further configured as: the method for intelligent monitoring of a trolley based on a cloud platform further includes: Synchronizing the multi-source data collected during the construction process of the trolley with the video monitoring data and classifying and storing them in the database of the cloud platform according to time periods, construction locations, and event types; When an abnormal state or construction deviation is detected, marking the video segments and monitoring data corresponding to the time period to form a traceable construction record.

[0019] By adopting the above technical solutions, by synchronizing the multi-source data collected during the construction process with the video surveillance data and storing them classified by time period, construction location, and event type, a systematic data archive of the entire construction process can be constructed, thereby providing a complete basis for later quality management and problem traceability; by marking the corresponding video segments and monitoring data when detecting abnormal states or construction deviations to form a traceable record, automatic event retention and classification filing can be achieved, thereby improving the transparency of construction and the traceability of quality responsibilities.

[0020] The second above-mentioned inventive object of the present application is achieved by the following technical solutions: A trolley intelligent monitoring system based on a cloud platform, the trolley intelligent monitoring system based on a cloud platform includes: An acquisition module for acquiring multi-source data of the trolley during the construction process, the multi-source data including sensor data and video surveillance data; A preprocessing module for preprocessing the multi-dimensional data through an edge computing device and uploading the preprocessed data to the cloud platform for centralized processing and storage; An analysis module for analyzing the preprocessed data, identifying the construction state of the trolley, and generating an analysis result of the construction state of the trolley, the construction state including the trolley running state, the concrete pouring state, and the lining quality; An adjustment module for dynamically adjusting the construction parameters of the trolley according to the analysis result of the construction state of the trolley.

[0021] By adopting the above technical solutions, by using the acquisition module to acquire multi-source data during the construction process, the full-scale perception information of the construction site can be obtained, thereby providing a data basis for subsequent intelligent processing; by using the preprocessing module to preprocess the data through edge computing and upload it to the cloud platform, the real-time performance of on-site data processing can be improved and the network load can be reduced, thereby ensuring the data transmission and analysis efficiency; by using the analysis module to identify the state of the preprocessed data and generate an analysis result, the concrete pouring state and the lining quality can be accurately monitored, thereby realizing the real-time control of the construction process; by using the adjustment module to dynamically adjust the construction parameters according to the analysis result, the intelligent closed-loop control of the construction process can be achieved, thereby effectively improving the construction automation level and the stability of the formed quality.

[0022] In summary, the present application includes the following beneficial technical effects: 1. By collecting multi-source data during the construction process of the trolley, multi-dimensional information including sensor data and video surveillance data can be obtained, enabling a comprehensive perception of the construction site status, thus avoiding monitoring blind spots caused by relying solely on a single data source and improving the coverage and perception accuracy of construction monitoring. By preprocessing the multi-dimensional data through edge computing devices and uploading the preprocessed data to the cloud platform for centralized processing and storage, the real-time performance and response speed of data processing can be improved, reducing the cloud load, and thereby enhancing the processing efficiency and stability of the system; 2. By analyzing the preprocessed data to identify the construction status of the trolley and generating a construction status analysis result, real-time status determination of the entire construction process can be achieved, thus promptly detecting problems such as construction deviations and quality risks and ensuring the quality of lining construction. By dynamically adjusting construction parameters according to the construction status analysis result, a closed-loop regulation system of monitoring-identifying-controlling can be formed, thereby realizing the intelligent and adaptive regulation of the concrete lining trolley construction and improving construction quality and efficiency. Description of the Drawings

[0023] Figure 1 is a flowchart of a trolley intelligent monitoring method based on a cloud platform in an embodiment of the present application; Figure 2 is an implementation flowchart of step S10 in a trolley intelligent monitoring method based on a cloud platform in an embodiment of the present application; Figure 3 is an implementation flowchart of step S20 in a trolley intelligent monitoring method based on a cloud platform in an embodiment of the present application; Figure 4 is an implementation flowchart of step S21 in a trolley intelligent monitoring method based on a cloud platform in an embodiment of the present application; Figure 5 is an implementation flowchart of step S22 in a trolley intelligent monitoring method based on a cloud platform in an embodiment of the present application; Figure 6 is an implementation flowchart of step S30 in a trolley intelligent monitoring method based on a cloud platform in an embodiment of the present application; Figure 7 is an implementation flowchart of a trolley intelligent monitoring method based on a cloud platform in an embodiment of the present application; Figure 8 is a principle block diagram of a trolley intelligent monitoring system based on a cloud platform in an embodiment of the present application. Detailed Description of the Embodiment

[0024] The following further details the present application with reference to the accompanying drawings.

[0025] In one embodiment, as Figure 1As shown in the figure, the present application discloses an intelligent monitoring method for a trolley based on a cloud platform, which specifically includes the following steps: S10: Collect multi-source data during the construction of the trolley. The multi-source data includes sensor data and video monitoring data.

[0026] Specifically, before the trolley advances to the designated lining construction section of the tunnel, the system acquisition command is triggered through a preset data acquisition scheduling mechanism. The scheduling interface will call various front-end acquisition components to start the data acquisition task, poll and access the sensor nodes deployed at each key part of the trolley, and obtain the real-time status information at the current time point. The sensor data includes structural pressure signals, fluid flow rates, material temperatures, formwork displacements, vibration responses, etc. The data is buffered at a standard sampling frequency and unified encoding format. At the same time, through the video monitoring module integrated with the trolley body, multi-channel image data covering the periphery of the formwork, the vibration area, and the operation area is collected. The image data is pre-embedded with timestamps at the acquisition end for subsequent synchronous comparison. All the collected data is cached in the edge node temporary storage area for preprocessing to ensure the integrity and continuity of the construction data.

[0027] S20: Preprocess the multi-dimensional data through the edge computing device, and upload the preprocessed data to the cloud platform for centralized processing and storage.

[0028] Specifically, after the edge computing device receives the complete data packet, it first performs a preliminary validity check on the sensor data, eliminates invalid data points caused by transmission jitter or sampling errors, and at the same time executes signal denoising algorithms such as moving average filtering and median filtering to weaken the occasional outliers caused by environmental interference. Secondly, the sensor data from different sources is aligned and interpolated according to a unified time axis to fill the time domain gaps caused by network packet loss or synchronization deviation. The image data is frame-compressed using the H.264 compression standard and key frames or working condition-related frames are extracted to reduce transmission redundancy, and an index association table is constructed between the sensing data and the image frames. Subsequently, it is packaged and uploaded to the specified receiving service interface of the cloud platform in the form of JSON or binary data stream. SSL encryption is used during the upload process to ensure data security, and finally, centralized management of multi-dimensional construction data in the cloud is realized.

[0029] S30: Analyze the preprocessed data, identify the construction state of the trolley, and generate an analysis result of the construction state of the trolley. The construction state includes the running state of the trolley, the concrete pouring state, and the lining quality.

[0030] Specifically, when the cloud platform receives the uplink data, the system will call the state analysis engine that combines the rule base and the statistical model to extract features from the data in multiple dimensions such as the pressure curve, flow rate fluctuation, temperature change, and template position change of the sensor, identify whether the concrete pouring is in a continuous and stable state, whether there is deformation or incomplete fitting of the template, and at the same time combine the trolley position data to judge whether there are abnormal operations such as non-linear displacement and sudden stop, forming a multi-dimensional construction state vector. After clustering analysis and comparison with historical construction parameters, this vector generates a set of structured results representing the current construction stage state, including the operation stage classification (such as standby, pouring, form removal), the pouring completion index, the preliminary quality assessment level, and possible risk warnings for subsequent viewing and response by the control terminal and supervision personnel.

[0031] S40: Dynamically adjust the construction parameters of the trolley according to the construction state analysis result of the trolley.

[0032] Specifically, after the system generates the construction state analysis result, the control logic will perform deviation analysis against the currently set target parameters. If there is a situation where the deviation significantly exceeds the threshold, it will trigger the dynamic control strategy module to reschedule the key parameters such as the hydraulic position of the template, the concrete pumping flow rate, and the vibration frequency. The control strategy is set by combining the parameter feedback mechanism and the priority rule. For example, when the vibration uniformity is insufficient, the vibration power is preferentially increased rather than modifying the flow rate. The adjustment instruction packet is pushed to the local control unit of the trolley to drive it to modify the execution parameters, and the feedback data is continuously collected during the control feedback action to verify the adjustment effect, ensuring that the adjustment operation is immediate, effective, and closed-loop. At the same time, this adjustment action will be recorded in the construction log for subsequent review.

[0033] By adopting the above technical solutions, by collecting multi-source data during the construction process of the trolley, multi-dimensional information including sensor data and video monitoring data can be obtained, realizing a comprehensive perception of the construction site state, thus avoiding the monitoring blind area caused by relying solely on a single data source and improving the coverage and perception accuracy of construction monitoring; by preprocessing the multi-dimensional data through the edge computing device and uploading the preprocessed data to the cloud platform for centralized processing and storage, the real-time performance and response speed of data processing can be improved, reducing the cloud load, thereby improving the processing efficiency and stability of the system; by analyzing the preprocessed data to identify the construction state of the trolley and generating the construction state analysis result, the real-time state determination of the entire construction process can be realized, thus timely discovering problems such as construction deviation and quality risk and ensuring the lining construction quality; by dynamically adjusting the construction parameters according to the construction state analysis result, a closed-loop adjustment system of monitoring-identifying-controlling can be formed, thus realizing the intelligent and adaptive control of the concrete lining trolley construction and improving the construction quality and efficiency.

[0034] In one embodiment, as Figure 2 shown, in step S10, that is, collecting multi-source data of the trolley during construction, specifically including: S11: Detecting the concrete pouring pressure through a pressure sensor and determining the integrity degree of the lining.

[0035] Specifically, highly sensitive pressure sensors are fixedly installed at multiple positions of the trolley formwork structure, and the sensors are arranged in contact with the formwork steel surface to directly sense the acting force applied during the concrete injection process. When the sensors work, they continuously read the instantaneous pressure values per unit time at a high frequency (such as above 50 Hz), and monitor the pressure curve during the entire pouring process in real time. By comparing the pressure distribution trends of different formwork parts, it is judged whether the concrete is evenly filled. If it is found that the pressure peak is insufficient or the local pressure fluctuation is abnormal, it may reflect that this area is not fully filled or there is a cavity. The system can calculate the integrity score or risk level of the current lining based on this, providing a basis for on-site quality recheck and parameter adjustment.

[0036] S12: Monitoring the concrete flow rate with an ultrasonic flowmeter to obtain the concrete pouring flow rate.

[0037] Specifically, a non-contact ultrasonic flowmeter is set in the middle section of the concrete conveying pipeline. The actual flow rate of the concrete in the pipeline is calculated by the time difference of the ultrasonic signal reflection during transmission and reception. This process does not interfere with the concrete flow and can provide real-time feedback. By continuously measuring and combining with the pipeline cross-sectional area parameter, the flow rate value per unit time is calculated in real time. The system can judge whether there is pumping blockage, over-pouring or cut-off based on the flow trend. In addition, combined with the pouring time, this data can be used to deduce the cumulative pouring volume, providing a quantitative basis for the pouring progress control.

[0038] S13: Detecting the concrete temperature when entering the formwork with an infrared temperature sensor.

[0039] Specifically, the infrared temperature sensor is installed in a non-contact manner near the concrete outlet, continuously scanning and measuring the temperature of the concrete surface before entering the formwork. The system converts the collected thermal radiation value into temperature data. By analyzing the continuous temperature change curve, it can be used to identify in real time whether there is abnormal temperature control in the current concrete material, such as premature setting caused by too fast hydration heat reaction, or a decrease in fluidity in a low-temperature environment, etc., so as to give an early warning prompt for the applicability of the concrete when entering the formwork.

[0040] S14: Real-time monitoring the radial distance between the trolley formwork of the trolley and the concrete with a laser displacement sensor to obtain the formwork displacement data.

[0041] Specifically, laser displacement sensors are arranged at multiple edge points of the outer frame of the template. The sensor directs the laser beam to the concrete surface, and infers the actual distance between the current template and the concrete through the time difference and reflection angle of the laser beam return. The system synthesizes the return values ​​of multiple measuring points and constructs a template space posture model. If it is found that the edge or one side of the template continuously deviates from the design value, it is considered that the template is tilted or rebounded, which prompts the need for recalibration or fixation and reinforcement.

[0042] S15: Monitor the working state of the vibrator of the trolley through the MEMS vibration sensor to obtain vibration parameters.

[0043] Specifically, a miniature MEMS vibration sensor is fastened to the vibrator structure to measure in real time the acceleration changes and frequency response characteristics during the construction process. The system compares these vibration waveforms with the standard vibration model. If the fluctuation amplitude is insufficient or the spectrum is abnormal, it may indicate a vibrator failure, insufficient vibration or unstable operation. In addition, the system can count the vibration time and interval, thereby inferring the adequacy and standardization of the vibration operation, providing data support to ensure the compaction effect.

[0044] S16: Use an industrial camera to collect video data of the trolley construction process to obtain video monitoring data.

[0045] Specifically, multi-angle industrial cameras are installed at key locations such as the trolley operating platform, formwork joints, and pouring ports. The cameras have automatic focusing and low-light imaging capabilities, can adapt to the dim environment in the tunnel, continuously capture images of the entire construction process, and output image frames at a fixed frame rate. The system can set the recording cycle, resolution, and image encoding parameters in the background, transmit the video data to the data processing module through the edge node, and add timestamp information for synchronous comparison with the sensor data. The image includes elements such as personnel operation, equipment movement, and concrete status, which can be used to assist target recognition and behavior analysis, and provide image data support for the system to build a complete multi-source perception model.

[0046] During the implementation of the present invention, the video monitoring system conducts real-time monitoring and video recording of the entire secondary lining concrete pouring process, and fully covers and shoots key construction sites by deploying multiple cameras at multiple angles to ensure that there are no blind spots in the monitoring area and improve the level of on-site visualization. The cameras used have high-definition imaging capabilities and integrated night vision functions, and can clearly present image details even in insufficiently lit environments such as tunnels, effectively ensuring image quality and recognition accuracy. At the same time, the matching LED display screen has the characteristics of small size and light weight, which is convenient for flexible installation and deployment in the limited space of the trolley, does not affect construction operations and equipment operation, and further improves the system's integration and on-site adaptability.

[0047] In one embodiment, ifFigure 3 As shown, in step S20, the multi-dimensional data is preprocessed by the edge computing device, and the preprocessed data is uploaded to the cloud platform for centralized processing and storage, specifically including: S21: The data filtering algorithm is used to denoise the sensor data, and the correlated calculation is performed on the denoised data to generate a preliminary evaluation result of the construction state.

[0048] Specifically, after receiving the original sensor data, first, a filter queue is established for each data channel respectively. According to the data characteristics, denoising algorithms such as moving average filtering, median filtering, or exponential weighted smoothing are selected to remove high-frequency noise or discrete outliers caused by equipment jitter, electromagnetic interference, or poor contact. Subsequently, cross-channel correlation processing is performed on the denoised data to extract the synchronous change trends and interaction characteristics between the data. For example, through correlation analysis, it is identified whether the sudden change in vibration frequency is accompanied by phenomena such as an increase in formwork displacement or a decrease in pressure. These interaction patterns are transformed into a preliminary evaluation reference for construction state factors, thereby realizing a preliminary judgment of the current pouring state, formwork stability, or equipment operation smoothness and outputting an initial result vector. S22: Object detection is performed on the video surveillance data to identify construction anomalies.

[0049] Specifically, after the video data compression and upload are completed, a video segment with key frame characteristics of the construction is selected and the object detection model is called for image recognition processing. The recognition algorithm is based on the trained video construction scene dataset. The bounding box information of the main objects such as formwork, concrete nozzle, operator, vibrator, etc. in the image is extracted through the feature extraction network and classified and labeled. By comparing the spatial position relationships and pose states between the objects, it is judged whether there are image abnormal behaviors such as formwork opening, perfusion interruption, personnel leaving the post, vibrator not placed, etc. The recognition result will be accompanied by a confidence score and an image frame number, and at the same time, the frame is marked with an abnormal type code as a reference frame that may have construction deviation, providing an image layer input basis for subsequent data fusion analysis.

[0050] S23: Combining the sensing data and the video surveillance data, a preliminary statistics of the construction state of the trolley is performed to obtain the preprocessed data and mark potential risk points.

[0051] Specifically, after completing the noise reduction of sensing data and image target detection, the two types of data are aligned based on the same timestamp system. The abnormal behaviors identified in the image frames are respectively mapped to the sensor data windows at the corresponding moments. By constructing an event feature matrix, the image risk factors and sensor behavior indicators are jointly judged. The abnormal degrees of various signals are normalized by using a weighted statistical method to generate a comprehensive status score. If the score exceeds the preset risk threshold, the corresponding time period is marked as a potential risk time window. At the same time, the corresponding data segment is tagged and output as a preprocessed fusion data packet, which contains the data source identifier, timestamp, preliminary status evaluation result, and risk level annotation, and is used to enter the subsequent analysis and recognition module.

[0052] In one embodiment, as Figure 4 shown, in step S21, that is, the correlated calculation is performed on the noise-reduced data to generate a preliminary evaluation result of the construction status, specifically including: S211: Synchronize the time of the noise-reduced data and use the interpolation algorithm to fill in the missing data to obtain the corrected data.

[0053] Specifically, for all the sensor data that has completed noise reduction processing, a unified time axis alignment operation is performed according to its original sampling time. The global time step is set and a standard sampling point sequence is established according to the main time axis. The sampling time points of each sensor are matched one by one, and the channels that do not hit the sampling points are subjected to time interpolation processing. The interpolation method selects methods such as linear interpolation, spline interpolation, or Lagrange interpolation according to the data stability and fluctuation characteristics to complement the missing data points, so that each sensor channel has continuous and comparable numerical samples in the same time series, thereby obtaining a multi-channel fusion data set with a consistent time series structure, preparing data for spatial projection and weighted calculation.

[0054] S212: According to the sensor layout position of the trolley, construct a spatial mapping model of the trolley construction area, perform spatial projection on the corrected data, use a weighted calculation method to calculate the weight contribution values of each construction parameter, obtain the weighted calculation result, and dynamically adjust the weights of each construction parameter according to the construction environment where the trolley is located, the concrete setting characteristics, and the trolley status.

[0055] Specifically, a spatial mapping table is established based on the actual installation positions of each sensor and the schematic diagram of the trolley formwork structure. Information such as the spatial coordinates corresponding to each sensor, the formwork associated surface, and the construction process impact factors is recorded in the table. Subsequently, the time series data of each sensor is projected onto the corresponding area in the 3D construction model according to the spatial coordinates, and a data distribution map of the construction area is established. Combining the trend consistency of multi-point data, spatial weight values are assigned to different areas. On this basis, the comprehensive contribution values of various construction parameters are calculated according to the preset weight fusion model to generate an intermediate weighted calculation result. During the fusion process, the parameter weight ratio is dynamically adjusted according to factors such as the temperature and humidity at the construction site, the trolley attitude, the construction rhythm, and the initial setting time of the concrete. For example, the proportion of temperature data is increased in a low-temperature environment, and the sensitivity of displacement parameters is increased when formwork shaking is detected, ensuring that the weighted result has stronger environmental adaptability and judgment accuracy.

[0056] S213: Correlate and analyze the video surveillance data with the weighted calculation result to generate a preliminary evaluation result of the construction status.

[0057] Specifically, the construction behavior recognition result extracted based on image object detection is correlated and compared with the comprehensive parameter score from the weighted calculation model. By establishing a logical mapping model between the abnormal type index and the weighted parameter dimension, such as jointly correlating the formwork opening phenomenon with the decrease in formwork pressure value and displacement offset, a composite event model is constructed for linkage matching analysis to determine whether a specific image anomaly has corresponding support from sensor data. If the two are highly consistent, the abnormal confidence level of the event is increased. Conversely, if the image anomaly lacks data support, it is marked as suspicious or a secondary risk. Finally, the construction status evaluation result is output, which includes the status classification label, comprehensive health score, main risk type, and recommended adjustment direction at the current stage, providing a decision-making basis for the subsequent control module.

[0058] In one embodiment, as Figure 5 shown, in step S22, that is, performing object detection on the video surveillance data to identify construction anomalies, specifically including: S221: Use the YOLO algorithm to identify image anomalies at the construction site.

[0059] Specifically, after preprocessing the video data at the construction site and extracting key frames, the video frame images are input into a preset YOLO object detection model. The YOLO model performs multi-scale feature extraction on the images according to its internal convolutional neural network structure and generates candidate boxes, and real-time detects and classifies objects in the construction scene such as formwork, concrete discharge outlets, construction workers, vibrators, etc. Subsequently, the detection results are compared with the standard construction posture model to identify whether there are abnormal behaviors or state deviations in the images. For example, situations such as detected formwork edge offsets, improper positions of construction workers, or excessive material accumulation are identified. The identified abnormal image frames are tagged with feature labels and their corresponding timestamps, position annotations, and abnormal type numbers are recorded, providing an image evidence basis for subsequent risk analysis.

[0060] S222: Analyze the voice commands of construction workers in combination with speech recognition, and judge whether the voice commands conform to the operation process.

[0061] Specifically, after the voice signals picked up at the construction site are collected in real time through a microphone array, an end-to-end speech recognition model is used for speech content transcription. This recognition model supports the mixed recognition of Mandarin and dialects and can perform signal-to-noise ratio optimization processing according to the noise characteristics of the construction site. The recognized speech text is matched with keywords and compared with the semantic structure of the construction operation instruction library. For example, the instruction library contains standardized operation passwords such as "start perfusion", "stop vibration", "formwork closing", etc. It is judged whether the construction worker's password is issued at the correct time and the content is standardized through semantic matching. At the same time, the recognized instructions are compared with the construction behavior data of this time period. If the recognized voice is inconsistent with the current equipment state or process state, it is recorded as a suspected violation instruction event and used as an auxiliary judgment basis for subsequent behavior auditing and quality traceability.

[0062] S213: Use lidar scanning technology to construct a three-dimensional point cloud model of the construction environment where the trolley is located, and detect whether the trolley deviates from the predetermined trajectory or physical interference occurs.

[0063] Specifically, the lidar is installed on the top or front end of the trolley and continuously scans the construction tunnel environment in a 360° rotation or directional manner. By emitting laser beams and receiving reflected signals to calculate the distance and reflection angle, the position data of each structural surface in the construction space is obtained in real time. The scanning results generate high-density three-dimensional point cloud data according to time slices and synchronously superimpose the trolley position information. Through three-dimensional registration with a preset tunnel construction BIM model or standard trajectory line, it is compared and judged whether there are problems such as abnormal distances between the current trolley outline and surrounding structures or attitude tilts. If it is found that the trolley deviates from the trajectory, approaches the side wall, or there is a risk of interference with the hoisting member, a warning message is generated and the offset and risk type are output, which are used to assist trolley movement adjustment and trigger safety intervention strategies.

[0064] In one embodiment, as Figure 6 shown, in step S30, the preprocessed data is analyzed to identify the construction state of the trolley and generate the construction state analysis result of the trolley, which specifically includes: S31: Extract features from the preprocessed data to obtain key parameters related to the construction state. The key parameters include the uniformity of concrete pouring, the vibration effect, the integrity of the lining structure, and the environmental stability.

[0065] Specifically, after noise reduction, time synchronization, and spatial mapping of the original multi-source data, feature parameters closely related to the construction state are extracted from each type of data. Specifically, the rate of change and the mean value of the pressure in the concrete pressure curve are extracted to represent the pouring uniformity, the effective amplitude, the frequency spectrum distribution, and the working duration in the vibration data are extracted to reflect the vibration effect, the peak value, the stability interval, and the maximum offset amplitude of the formwork displacement data are extracted to judge the stability of the lining structure, and the average value and the standard deviation of parameters such as environmental temperature, humidity, and vibration are collected to evaluate the environmental stability during construction. All feature parameters are classified and packaged according to the construction stage and a feature vector is constructed for the input of the subsequent state recognition model.

[0066] S32: Input the key parameters into a pre-trained decision tree algorithm for pattern recognition to determine the construction state of the trolley and generate the construction state analysis result. The construction state includes the trolley operation state, the concrete pouring state, and the lining quality. The construction state analysis result is used to drive the real-time optimization control of the construction parameters.

[0067] Specifically, the constructed construction feature vector is input into an offline-trained decision tree classification model. The model makes layer-by-layer branching judgments on the feature parameters through a tree structure, makes logical condition judgments at each node, and gradually infers the construction stage state of the current trolley. For example, it determines whether it is in the initial pouring period, the ending period, or the formwork adjustment period. At the same time, it combines the concrete state parameters and the formwork structure features to judge the quality grade of the current lining formation, and outputs a structured result including the construction state label, the risk assessment grade, and the recommended control strategy. This result will be used as the input of the parameter scheduling module to trigger the real-time adjustment of subsequent construction parameters, the optimization of thresholds, or the alarm control strategy to achieve a state-driven intelligent construction process.

[0068] To improve the accuracy of status determination and the real-time performance of responses, the system also performs semantic-layer cross-verification on the analysis results by combining video surveillance data corresponding to key parameters in specific time periods. Information such as the template posture, concrete status, and construction worker behavior in video images is used to assist in judging the credibility and consistency of the analysis results, ensuring that the status determination results are more robust and practical. The finally generated construction status analysis results will be used to drive the dynamic adjustment strategy of construction parameters, realizing the intelligent closed-loop management of the trolley operation control.

[0069] During the process of pre-training the decision tree algorithm, historical construction data samples from multiple tunnel construction sites are first collected. The samples include sensor data and construction status labels formed under different construction conditions, different lining sections, and different trolley working conditions. After the sample data undergoes unified preprocessing steps, including missing value filling, outlier removal, standard normalization processing, etc., key parameters are extracted according to feature dimensions, including multi-dimensional parameters such as pressure, vibration, displacement, and temperature during the concrete pouring process as input features, and the construction stage classification (such as operation, pouring, formwork removal) and quality grade (such as qualified, deviation, abnormal) are used as output labels. The information gain or Gini coefficient is used as the splitting criterion to construct the splitting nodes of the decision tree, and hyperparameters such as the tree depth and minimum sample split number are optimized through cross-validation to avoid overfitting problems. After the trained decision tree model reaches a high accuracy on the validation set, it is solidified into a model parameter file and deployed to the cloud platform for identifying and classifying the construction status of the real-time collected feature vectors during actual construction.

[0070] In one embodiment, as Figure 7 shown, this intelligent trolley monitoring method based on the cloud platform further includes: S50: Synchronize the multi-source data collected during the trolley construction process with the video surveillance data, and classify and store them in the database of the cloud platform according to time periods, construction locations, and event types.

[0071] Specifically, after all data preprocessing is completed at the edge, all the sensor data and video images uploaded to the cloud are aligned and matched according to a unified timestamp to construct a multi-modal data sequence on a unified time axis. Then, according to the construction location identifier, the data in each time period is attributed to the corresponding template area or trolley position block, and event type classification is performed according to the event labels contained in the data, such as "template fine-tuning", "pouring in progress", "abnormal warning", etc. The processed structured data is stored in the distributed database of the cloud platform. Different classification dimensions correspond to different index paths and retrieval interfaces. The database supports multi-condition retrieval callbacks according to time, event, location, etc., ensuring that the required construction data segments and video materials can be efficiently located during subsequent scheduling, analysis, playback, or auditing.

[0072] S60: When an abnormal state or construction deviation is detected, mark the video segments and monitoring data corresponding to the corresponding time period to form a traceable construction record.

[0073] Specifically, when the analysis module determines that there are abnormal states, parameter overruns, or identifies construction deviation behaviors during the construction process, it immediately locates the time period corresponding to the abnormal event and extracts the complete sensor data window and synchronized video frames within this time period, marks the abnormal event, construction location, and abnormal type description for this segment of data, and simultaneously establishes an abnormal event index table to bind and store this abnormal event with information such as the data storage path, timestamp, and project number. Subsequently, the abnormal tracking module can quickly retrieve the corresponding construction records according to the abnormal type, location, or time range, replay the on-site images, and view all the data environments when the abnormality occurred, providing original record support for quality traceability, liability determination, and construction improvement. It should be understood that the magnitudes of the sequence numbers of the steps in the above embodiments do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0074] In one embodiment, a trolley intelligent monitoring system based on a cloud platform is provided. This trolley intelligent monitoring system based on a cloud platform corresponds one-to-one with the trolley intelligent monitoring method based on a cloud platform in the above embodiment. As Figure 8 shown, this trolley intelligent monitoring system based on a cloud platform includes a collection module, a preprocessing module, an analysis module, and an adjustment module. The detailed descriptions of each functional module are as follows: The collection module is used to collect multi-source data during the construction process of the trolley. The multi-source data includes sensor data and video monitoring data; The preprocessing module is used to preprocess the multi-dimensional data through edge computing devices and upload the preprocessed data to the cloud platform for centralized processing and storage; The analysis module is used to analyze the preprocessed data, identify the construction state of the trolley, and generate an analysis result of the construction state of the trolley. The construction state includes the trolley running state, concrete pouring state, and lining quality; The adjustment module is used to dynamically adjust the construction parameters of the trolley according to the analysis result of the construction state of the trolley.

[0075] Optionally, the collection module includes: The pressure detection sub-module is used to detect the concrete pouring pressure through a pressure sensor and determine the integrity of the lining; The flow rate monitoring sub-module is used to monitor the concrete flow rate using an ultrasonic flowmeter to obtain the concrete pouring flow rate; The temperature detection sub-module is used to detect the concrete pouring temperature using an infrared temperature sensor; The laser detection sub-module is used to monitor the radial distance between the trolley formwork and the concrete of the trolley in real time through a laser displacement sensor, and obtain the formwork displacement data; The vibration monitoring sub-module is used to monitor the working state of the vibrator of the trolley through a MEMS vibration sensor, and obtain the vibration parameters; The video acquisition sub-module is used to collect video data of the trolley construction process by using an industrial camera, and obtain the video monitoring data.

[0076] Optionally, the preprocessing module includes: The preliminary evaluation sub-module is used to perform noise reduction processing on the sensor data by using a data filtering algorithm, and perform correlation calculation on the data after noise reduction processing, and generate a preliminary evaluation result of the construction state; The anomaly recognition sub-module is used to perform target detection on the video monitoring data to identify construction anomalies; The preliminary statistics sub-module is used to combine the sensing data and the video monitoring data to perform preliminary statistics on the construction state of the trolley, obtain the data after preprocessing, and mark potential risk points.

[0077] Optionally, the preliminary evaluation sub-module includes: The calibration unit is used to synchronize the time of the data after noise reduction processing, and use the interpolation algorithm to fill in the missing data to obtain the calibrated data; The calculation unit is used to construct a spatial mapping model of the trolley construction area according to the sensor layout position of the trolley, perform spatial projection on the calibrated data, use a weighted calculation method to calculate the weight contribution value of each construction parameter, obtain the weighted calculation result, and dynamically adjust the weights of each construction parameter according to the construction environment where the trolley is located, the concrete setting characteristics, and the trolley state; The correlation analysis unit is used to perform correlation analysis on the video monitoring data and the weighted calculation result to generate a preliminary evaluation result of the construction state.

[0078] Optionally, the anomaly recognition sub-module includes: The algorithm recognition unit is used to identify image anomalies at the construction site by using the YOLO algorithm; The voice analysis unit is used to combine voice recognition to analyze the voice commands of construction workers and judge whether the voice commands conform to the operation process; The 3D construction unit is used to construct a 3D point cloud model of the construction environment where the trolley is located by using lidar scanning technology to detect whether the trolley deviates from the predetermined trajectory or physical interference occurs.

[0079] Optionally, the analysis module includes: The feature extraction sub-module further extracts features from the preprocessed data to obtain key parameters related to the construction status. The key parameters include the uniformity of concrete pouring, the vibration effect, the integrity of the lining structure, and the environmental stability. The pattern recognition sub-module is used to input the key parameters into a pre-trained decision tree algorithm for pattern recognition to determine the construction status of the trolley and generate a construction status analysis result. The construction status includes the trolley operation status, the concrete pouring status, and the lining quality. The construction status analysis result is used to drive the real-time optimization control of the construction parameters.

[0080] Optionally, the trolley intelligent monitoring system based on the cloud platform further includes: The storage module is used to synchronize the multi-source data collected during the trolley construction process with the video monitoring data and classify and store them in the database of the cloud platform according to the time period, construction location, and event type. The marker backtracking module is used to mark the video segments and monitoring data corresponding to the time period when an abnormal state or construction deviation is detected to form a traceable construction record.

[0081] For the specific limitations of a trolley intelligent monitoring system based on the cloud platform, reference can be made to the limitations of a trolley intelligent monitoring method based on the cloud platform in the above text, which will not be elaborated here. Each module in the above trolley intelligent monitoring system based on the cloud platform can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0082] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the above division of each functional unit and module is used as an example. In actual applications, the above functions can be allocated to different functional units and modules according to needs, that is, the internal structure of the system is divided into different functional units or modules to complete all or part of the functions described above.

[0083] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A trolley intelligent monitoring method based on a cloud platform, characterized in that: The cloud platform-based trolley intelligent monitoring method includes: Collect multi-source data of the trolley during the construction process, wherein the multi-source data includes sensor data and video monitoring data; Preprocessing the multidimensional data through edge computing devices, and uploading the preprocessed data to a cloud platform for centralized processing and storage; Analyze the preprocessed data, identify the construction status of the trolley, and generate a construction status analysis result of the trolley, wherein the construction status includes the trolley operation status, concrete pouring status, and lining quality; According to the analysis result of the construction status of the trolley, the construction parameters of the trolley are dynamically adjusted.

2. According to the cloud platform-based intelligent monitoring method for trolleys according to claim 1, it is characterized in that: The multi-source data collected by the trolley during the construction process includes: The concrete pouring pressure is detected by the pressure sensor, and the integrity of the lining is determined; An ultrasonic flow meter is used to monitor the concrete flow rate and obtain the concrete pouring flow rate; Use infrared temperature sensor to detect the temperature of concrete entering the mold; The radial distance between the trolley template and the concrete is monitored in real time by a laser displacement sensor to obtain template displacement data; Monitor the working state of the vibrator of the trolley by using a MEMS vibration sensor to obtain vibration parameters; An industrial camera is used to collect video data of the trolley construction process to obtain the video monitoring data.

3. The method for intelligent monitoring of a trolley based on a cloud platform according to claim 2, characterized in that: Preprocessing the multidimensional data by edge computing equipment and uploading the preprocessed data to a cloud platform for centralized processing and storage includes: Using a data filtering algorithm to perform noise reduction processing on the sensor data, and performing correlation calculation on the noise-reduced data to generate a preliminary evaluation result of the construction status; Performing target detection on the video surveillance data to identify abnormal construction conditions; Combined with the sensor data and the video surveillance data, preliminary statistics are made on the construction status of the trolley to obtain the preprocessed data and mark potential risk points.

4. The method for intelligent monitoring of a trolley based on a cloud platform according to claim 3 is characterized in that: The performing of correlation calculation on the noise-reduced data to generate a preliminary evaluation result of the construction status includes: Performing time synchronization on the noise-reduced data and filling in missing data using an interpolation algorithm to obtain corrected data; According to the sensor arrangement position of the trolley, a spatial mapping model of the trolley construction area is constructed, the corrected data is spatially projected, a weighted calculation method is used to calculate the weight contribution value of each construction parameter, and a weighted calculation result is obtained, and the weight of each construction parameter is dynamically adjusted according to the construction environment of the trolley, the concrete setting characteristics and the trolley state; The video surveillance data is correlated with the weighted calculation result to generate a preliminary evaluation result of the construction status.

5. The method for intelligent monitoring of a trolley based on a cloud platform according to claim 3 is characterized in that: The performing target detection on the video surveillance data to identify abnormal construction conditions includes: Use the YOLO algorithm to identify image anomalies at the construction site; Analyze the voice instructions of the construction personnel in combination with voice recognition to determine whether the voice instructions comply with the operating procedures; A three-dimensional point cloud model of the construction environment in which the trolley is located is constructed using laser radar scanning technology to detect whether the trolley deviates from the predetermined trajectory or physical interference occurs.

6. The method for intelligent monitoring of a trolley based on a cloud platform according to claim 1, characterized in that: The analyzing the pre-processed data, identifying the construction status of the trolley, and generating the construction status analysis result of the trolley includes: Performing feature extraction on the preprocessed data to obtain key parameters related to the construction status, wherein the key parameters include concrete pouring uniformity, vibration effect, lining structure integrity and environmental stability; The key parameters are input into a pre-trained decision tree algorithm for pattern recognition to determine the construction status of the trolley and generate the construction status analysis results, which include the trolley operation status, concrete pouring status and lining quality. The construction status analysis results are used to drive real-time optimization control of construction parameters.

7. The method for intelligent monitoring of a trolley based on a cloud platform according to claim 1, characterized in that: The cloud platform-based trolley intelligent monitoring method further includes: The multi-source data collected during the construction process of the trolley are synchronized with the video surveillance data, and are classified and stored in the database of the cloud platform according to time period, construction location and event type; When an abnormal condition or construction deviation is detected, the video clips and monitoring data of the corresponding time period are marked to form a traceable construction record.

8. A trolley intelligent monitoring system based on a cloud platform, characterized in that: The trolley intelligent monitoring system based on the cloud platform includes: A collection module is used to collect multi-source data of the trolley during the construction process, wherein the multi-source data includes sensor data and video monitoring data; A preprocessing module, used to preprocess the multidimensional data through an edge computing device, and upload the preprocessed data to a cloud platform for centralized processing and storage; An analysis module, used to analyze the preprocessed data, identify the construction status of the trolley, and generate an analysis result of the construction status of the trolley, wherein the construction status includes the trolley operation status, concrete pouring status, and lining quality; The adjustment module is used to dynamically adjust the construction parameters of the trolley according to the analysis results of the construction status of the trolley.

Images