A foundation pit deformation visual monitoring and early warning system and method

Abstract

The application discloses a kind of foundation pit deformation visual monitoring early warning system and method, it is related to foundation pit visual monitoring technical field, in the surface of foundation pit to be monitored covers electronic skin;Electronic skin includes composite film layer, strain sensing layer, pressure signal layer, environmental perception layer, power supply unit and surface functional layer;Multispectral camera array is deployed in the periphery of foundation pit to be monitored;In surface functional layer, fluorescent two-dimensional code mark point is printed.The electronic skin of the application can directly and continuously monitor the small strain and soil pressure change on the surface of foundation pit through strain sensing layer and pressure signal layer, and provide high-precision mechanical data.Multispectral camera array provides visible light image for identifying mark point to measure displacement, also provides thermal image for monitoring temperature field change, captures potential anomaly difficult to be found by naked eye or single sensor.Electronic skin covers the surface of foundation pit, continuous, distributed sensing for the whole monitoring area, avoids local deformation that traditional discrete measuring point can miss.

Description

Technical Field

[0001] This invention relates to the field of visual monitoring technology for foundation pits, and in particular to a visual monitoring and early warning system and method for foundation pit deformation. Background Technology

[0002] Foundation pit engineering is an indispensable and crucial component of modern construction projects, and its safety directly impacts the overall quality and stability of the project. Foundation pit deformation monitoring, a key means of ensuring the safety of foundation pit projects, has long relied on traditional measurement methods and sensor technology. However, with the continuous expansion of construction project scale, foundation pit deformation monitoring faces multiple challenges, including increasing accuracy requirements, improving monitoring efficiency, and enhancing real-time performance.

[0003] The industry currently faces numerous pain points in deformation monitoring, especially in the monitoring of deformation in foundation pits and tunnels, where the available monitoring methods are relatively limited. Common general-purpose monitoring methods first create a 3D model of the structure being monitored, and then output the detection results using the 3D information. This process involves a large amount of data, slow processing speed, and relatively low accuracy. Furthermore, the complex algorithms result in low real-time monitoring performance.

[0004] Specifically, traditional monitoring methods have the following limitations:

[0005] Manual monitoring is inefficient: Traditional deformation monitoring mainly relies on manual labor, which is not only time-consuming and labor-intensive, but also has a limited monitoring frequency, making it difficult to detect and deal with deformation problems in a timely manner.

[0006] Insufficient data accuracy: Some monitoring methods and instruments are not accurate enough to capture minute deformation changes, resulting in inaccurate monitoring results and affecting subsequent analysis and decision-making.

[0007] Limited monitoring range: For large or complex civil engineering structures, traditional monitoring methods often cannot cover the entire area, resulting in monitoring blind spots and increasing safety hazards.

[0008] Data processing is complex: The large amount of data generated by deformation monitoring needs to be analyzed and processed, but traditional data processing methods are cumbersome and inefficient, making it difficult to meet the needs of real-time monitoring. Summary of the Invention

[0009] To address the aforementioned technical problems, this invention provides a visual monitoring and early warning system and method for foundation pit deformation. The technical solution adopted is as follows:

[0010] A method for visual monitoring and early warning of foundation pit deformation includes the following steps:

[0011] Step 1: Cover the surface of the foundation pit to be monitored with an electronic skin; the electronic skin includes a composite thin film layer, a strain sensing layer, a pressure signal layer, an environmental sensing layer, a power supply unit, and a surface functional layer; deploy a multispectral camera array around the foundation pit to be monitored; print fluorescent QR code markers on the surface functional layer;

[0012] Step 2: The strain sensing layer and pressure signal layer on the electronic skin monitor the strain and soil pressure changes on the surface of the foundation pit, generating resistance change signals and capacitance change signals; the environmental sensing layer collects environmental temperature and humidity data; the multispectral camera array synchronously captures images of the foundation pit, thermal images, and marked point images at a set frequency.

[0013] Step 3: The computer monitoring server collects electronic skin data at set intervals. The electronic skin data includes resistance change signal data, capacitance change signal data, and environmental temperature and humidity data. Simultaneously, it collects visual data from a multispectral camera array, which includes pit images, thermal images, and marked point images.

[0014] Step 4: The computer monitoring server synchronizes the electronic skin data timestamps with the multispectral camera array visual data at the same frame rate.

[0015] Step 5: Calculate the local strain value based on the resistance change signal data; calculate the pressure distribution on the surface of the foundation pit based on the capacitance change signal data; identify fluorescent QR code markers using image processing algorithms, establish a spatial coordinate system, calculate the displacement field on the surface of the foundation pit using image matching algorithms, and analyze the temperature distribution and abnormal areas in the thermal infrared image.

[0016] Step 6: Fuse the strain data from the electronic skin monitoring with the displacement data from the vision system;

[0017] Step 7: Use a risk assessment algorithm based on fused data to output the deformation risk level of the fluorescent QR code markers.

[0018] By adopting the above technical solution, the electronic skin can directly and continuously monitor the minute strain and soil pressure changes on the surface of the foundation pit through the strain sensing layer and the pressure signal layer, providing high-precision mechanical data.

[0019] Multispectral camera arrays not only provide visible light images for identifying marker points and measuring displacement, but also provide thermal images for monitoring temperature field changes and capturing potential anomalies that are difficult to detect with the naked eye or a single sensor (such as leakage or temperature anomalies caused by stress concentration).

[0020] The environmental sensing layer supplements temperature and humidity information, which helps to correct the impact of environmental factors on sensor readings and material properties, thereby improving the accuracy of monitoring data.

[0021] The electronic skin covering the surface of the pit enables continuous and distributed sensing of the entire monitoring area, avoiding the local deformation that may be missed by traditional discrete measuring points.

[0022] By combining a multispectral camera array with high-frequency shooting, the system can achieve near real-time data acquisition and promptly detect deformation trends and sudden anomalies.

[0023] By fusing the intrinsic mechanical data (strain, pressure) provided by the electronic skin with the extrinsic geometric and physical data (displacement, temperature) provided by the multispectral camera, the state of the foundation pit can be cross-verified from multiple dimensions.

[0024] For example, an increase in strain in a certain area should correspond to a change in displacement or pressure in that area. Data fusion can eliminate the limitations or false alarms of a single sensor, providing a more comprehensive and reliable assessment of the foundation pit's condition.

[0025] The system automates the entire process from data collection and processing to analysis and early warning, reducing manual intervention and improving monitoring efficiency and consistency.

[0026] Risk assessment algorithms based on fused data can automatically calculate the deformation risk level of each marked point (representing the area), providing a quantitative basis for engineering decisions and achieving an intelligent leap from monitoring to early warning.

[0027] The electronic skin uses flexible materials, which can fit well against the surface of the pit and adapt to certain deformations.

[0028] Fluorescent QR code markers are visible under specific lighting conditions. Combined with image processing algorithms, they can be reliably identified even in complex backgrounds. Furthermore, the QR code itself is unique, making it easy to track long-term changes at specific points.

[0029] Timestamp and frame rate synchronization ensures accurate temporal correspondence between data from different sources, providing a foundation for subsequent precise analysis and event backtracking.

[0030] Multispectral data (visible light, heat) combined with strain, pressure and other data can generate richer visualization products, such as displacement cloud maps, strain cloud maps, pressure distribution maps, temperature anomaly maps, etc., enabling engineers to intuitively understand the complex state of the foundation pit.

[0031] By integrating advanced electronic skin sensing technology and multispectral vision technology, comprehensive, accurate, real-time, and intelligent monitoring of foundation pit deformation has been achieved, significantly improving the safety level of foundation pit engineering and possessing significant engineering application value.

[0032] Optionally, step 8 is also included, which sets a deformation risk level threshold and triggers an alarm when the deformation risk level exceeds the threshold.

[0033] By adopting the above technical solution, the alarm mechanism ensures that the deformation of the foundation pit can be identified and addressed before it reaches a critical state that may cause safety problems. Compared with traditional methods, which may suffer from problems such as delayed analysis of monitoring data and delays in manual judgment, this automated threshold-triggered alarm greatly shortens the time from monitoring to response and improves the efficiency of risk management.

[0034] Optionally, the composite film layer is a polyimide-silicone composite film, the strain sensing layer is a silver nanowire resistive mesh, the pressure signal layer is a flexible capacitive pressure sensor array, the environmental sensing layer is a flexible temperature and humidity sensor array, the power supply units supply power to the strain sensing layer, the pressure signal layer and the environmental sensing layer respectively, the surface functional layer is a nano-antireflective film, and fluorescent QR code marking points are printed using inkjet printing and ultraviolet curing processes.

[0035] By employing the above technical solutions, both polyimide (PI) and silicone exhibit excellent flexibility, enabling them to closely adhere to the surface of the foundation pit. Even with some surface irregularities, good contact between the sensor layer and the monitored object is ensured, reducing measurement errors. PI possesses excellent high-temperature resistance, low-temperature resistance, and chemical corrosion resistance, while silicone offers good waterproof and moisture-proof properties. This combined use enhances the long-term stability and reliability of the electronic skin in complex and harsh outdoor foundation pit environments (such as exposure to sunlight and rain, and temperature fluctuations).

[0036] PI provides good mechanical strength and rigid support, while silicone provides flexibility and sealing. The composite structure ensures flexibility while also having a certain degree of tensile and tear resistance, making it less prone to damage during installation or when the foundation pit undergoes minor deformation.

[0037] The silver nanowire resistance grid uses silver nanowires (AgNWs) with extremely low resistivity and excellent conductivity. The resulting grid structure is highly sensitive to minute deformations and can accurately capture the tensile and compressive strains on the surface of the pit (reflected by changes in resistance), thus achieving high-precision strain measurement.

[0038] Mesh structures enable large-area, uniform strain sensing, facilitating the acquisition of continuous strain field distribution information. AgNWs are easily fabricated on flexible substrates, perfectly matching the flexible properties of composite thin film layers.

[0039] Sensor array structures can provide spatial pressure distribution information, not just single-point pressure. They can identify areas of concentrated pressure or trends of change, providing richer data for analyzing soil stress states. Capacitive sensors typically have good linear output and long-term stability. Flexible substrates allow sensor arrays to conform to curved surfaces, adapting to changes in the shape of the excavation pit surface, and are less prone to measurement interference from additional stress caused by their own stiffness.

[0040] Simultaneous monitoring of temperature and humidity allows for a more comprehensive capture of the impact of environmental changes on the foundation pit and the performance of the electronic skin itself.

[0041] Nanoscale antireflective coatings can reduce the reflection and scattering of light on the surface of electronic skin, improving the clarity and contrast of images captured by multispectral cameras (especially visible light images), which is beneficial for subsequent marker identification and image analysis.

[0042] Optionally, the formula for calculating the local strain value in step 5 is:

[0043] ;

[0044] in It is a local strain value. It is the relative change in phase resistance, and k is a material constant.

[0045] Optional,

[0046] The formula for the pressure distribution on the surface of the foundation pit in step 5 is:

[0047] ;

[0048] Where P is the calculated pressure value. This is the baseline pressure value. It measures the capacitance value. This is the initial capacitance value.

[0049] Optionally, the specific steps for the image matching algorithm to calculate the displacement field on the surface of the foundation pit are as follows:

[0050] Step 51: Extract SIFT feature points from images of consecutive time phases;

[0051] Step 52: Establish the correspondence between feature points through nearest neighbor matching and consistency check;

[0052] Step 53: Evaluate the matching quality using a random sampling consensus algorithm and remove abnormal matching points;

[0053] Step 54: Calculate the displacement field on the surface of the foundation pit based on the matching point pairs.

[0054] Optionally, in step 6, the step of fusing the strain data from the electronic skin monitoring with the displacement data from the vision system is as follows:

[0055] Step 61: The electronic skin strain data and visual displacement data are fused using a weighted fusion method.

[0056] Step 62: Spatial registration of electronic skin monitoring points and visual monitoring points to establish a correspondence;

[0057] Step 63: Evaluate the uncertainties of the electronic skin data and visual data, respectively;

[0058] Step 64: Assign weights to the two data sources based on the uncertainty index and calculate the fusion result.

[0059] Optionally, the formula for calculating the fusion result in step 64 is:

[0060] ;

[0061] in It's electronic skin data. It is visual displacement data. It is the weight of electronic skin data. It is the weight of visual displacement data. It is the result of fusion.

[0062] Optionally, the specific steps of the risk assessment algorithm based on fused data are as follows:

[0063] The fusion results of strain, displacement, and displacement rate indices are calculated based on the electronic skin strain data and visual displacement data fused from the corresponding fluorescent QR code markers, and the risk level is determined based on the risk level threshold range.

[0064] The fuzzy comprehensive evaluation method is adopted to integrate the risk assessment results of multiple indicators;

[0065] Based on the comprehensive evaluation results, the deformation risk level of the corresponding fluorescent QR code markers is determined.

[0066] A visual monitoring and early warning system for foundation pit deformation is provided, which implements a method for visual monitoring and early warning of foundation pit deformation. The visual monitoring and early warning system includes an electronic skin, a multispectral camera array, a computer monitoring server, and an alarm 4. The electronic skin includes a composite thin film layer, a strain sensing layer, a pressure signal layer, an environmental sensing layer, a power supply unit, and a surface functional layer. The bottom of the composite thin film layer is attached to the surface of the foundation pit to be monitored. The strain sensing layer is laid on the top surface of the composite thin film layer. The pressure signal layer is laid on the strain sensing layer based on the isolation layer. The environmental sensing layer is laid on the pressure signal layer based on the isolation layer. The surface functional layer is laid on the environmental sensing layer. The power supply unit supplies power to the strain sensing layer, the pressure signal layer, and the environmental sensing layer respectively. The multispectral camera array is deployed around the foundation pit to be monitored. The computer monitoring server is communicatively connected to the strain sensing layer, the pressure signal layer, the environmental sensing layer, and the multispectral camera array respectively. Based on the electronic skin data and visual displacement data, the system outputs the deformation risk level of the corresponding fluorescent QR code marker point. If the deformation risk level exceeds the deformation risk level threshold, the alarm 4 is activated.

[0067] In summary, the present invention has at least one of the following beneficial technical effects:

[0068] This invention provides a visual monitoring and early warning system and method for foundation pit deformation. The electronic skin, through a strain sensing layer and a pressure signal layer, can directly and continuously monitor minute strain and soil pressure changes on the foundation pit surface, providing high-precision mechanical data. The multispectral camera array not only provides visible light images for identifying marker points and measuring displacement, but also provides thermal images for monitoring temperature field changes, capturing potential anomalies that are difficult to detect with the naked eye or a single sensor.

[0069] The electronic skin covering the surface of the foundation pit enables continuous, distributed sensing of the entire monitoring area, avoiding the possibility of missing local deformations by traditional discrete measuring points. Combined with a multispectral camera array that captures images at high frequencies, the system can achieve near real-time data acquisition, promptly detecting deformation trends and sudden anomalies.

[0070] By fusing the intrinsic mechanical data provided by electronic skin with the extrinsic geometric and physical data provided by multispectral cameras, the condition of the foundation pit can be cross-verified from multiple dimensions. A risk assessment algorithm based on this fused data can automatically calculate the deformation risk level of each marker point, and an alarm mechanism ensures that the foundation pit deformation is identified and addressed before it reaches a critical state that could trigger safety issues. Compared to traditional methods, which may suffer from delayed analysis of monitoring data and delays in manual judgment, this automated threshold-triggered alarm significantly shortens the time from monitoring to response, improving the efficiency of risk management. Attached Figure Description

[0071] Figure 1 This is a flowchart illustrating a visual monitoring and early warning method for foundation pit deformation according to the present invention.

[0072] Figure 2 This is a schematic diagram of the electrical component connection principle of a visual monitoring and early warning system for foundation pit deformation according to the present invention;

[0073] Figure 3 This is a schematic diagram of the cross-sectional structure of the electronic skin of a visual monitoring and early warning system for foundation pit deformation according to the present invention;

[0074] Explanation of reference numerals in the attached figures: 11. Composite thin film layer; 12. Strain sensing layer; 13. Pressure signal layer; 14. Environmental sensing layer; 15. Power supply unit; 16. Surface functional layer; 2. Multispectral camera array; 3. Computer monitoring server; 4. Alarm. Detailed Implementation

[0075] The present invention will be further described in detail below with reference to the accompanying drawings.

[0076] This invention discloses a visual monitoring and early warning system and method for foundation pit deformation.

[0077] Reference Figures 1-3Example 1: A visual monitoring and early warning method for foundation pit deformation, comprising the following steps:

[0078] Step 1: Cover the surface of the foundation pit to be monitored with an electronic skin; the electronic skin includes a composite thin film layer 11, a strain sensing layer 12, a pressure signal layer 13, an environmental sensing layer 14, a power supply unit 15, and a surface functional layer 16; deploy a multispectral camera array 2 around the foundation pit to be monitored; print fluorescent QR code markers on the surface functional layer 16;

[0079] Step 2: The strain sensing layer 12 and pressure signal layer 13 on the electronic skin monitor the strain and soil pressure changes on the surface of the foundation pit, generating resistance change signals and capacitance change signals; the environmental sensing layer 14 collects environmental temperature and humidity data; the multispectral camera array 2 synchronously captures foundation pit images, thermal images and marker point images at a set frequency.

[0080] Step 3: The computer monitoring server 3 collects electronic skin data at set intervals. The electronic skin data includes resistance change signal data, capacitance change signal data and environmental temperature and humidity data. Simultaneously, it collects visual data from the multispectral camera array 2. The visual data from the multispectral camera array 2 includes pit images, thermal images and marked point images.

[0081] Step 4: The computer monitoring server 3 synchronizes the electronic skin data timestamp with the visual data of the multispectral camera array 2 at the frame rate.

[0082] Step 5: Calculate the local strain value based on the resistance change signal data; calculate the pressure distribution on the surface of the foundation pit based on the capacitance change signal data; identify fluorescent QR code markers using image processing algorithms, establish a spatial coordinate system, calculate the displacement field on the surface of the foundation pit using image matching algorithms, and analyze the temperature distribution and abnormal areas in the thermal infrared image.

[0083] Step 6: Fuse the strain data from the electronic skin monitoring with the displacement data from the vision system;

[0084] Step 7: Use a risk assessment algorithm based on fused data to output the deformation risk level of the fluorescent QR code markers.

[0085] The electronic skin, through the strain sensing layer 12 and the pressure signal layer 13, can directly and continuously monitor minute strain and soil pressure changes on the surface of the foundation pit, providing high-precision mechanical data.

[0086] The multispectral camera array 2 not only provides visible light images for identifying marker points for displacement measurement, but also provides thermal images for monitoring temperature field changes and capturing potential anomalies that are difficult to detect with the naked eye or a single sensor (such as leakage, temperature anomalies caused by stress concentration).

[0087] The environmental sensing layer 14 supplements temperature and humidity information, which helps to correct the impact of environmental factors on sensor readings and material properties, and improves the accuracy of monitoring data.

[0088] The electronic skin covering the surface of the pit enables continuous and distributed sensing of the entire monitoring area, avoiding the local deformation that may be missed by traditional discrete measuring points.

[0089] By combining a multispectral camera array with high-frequency shooting, the system can achieve near real-time data acquisition and promptly detect deformation trends and sudden anomalies.

[0090] By fusing the intrinsic mechanical data (strain, pressure) provided by the electronic skin with the extrinsic geometric and physical data (displacement, temperature) provided by the multispectral camera, the state of the foundation pit can be cross-verified from multiple dimensions.

[0091] For example, an increase in strain in a certain area should correspond to a change in displacement or pressure in that area. Data fusion can eliminate the limitations or false alarms of a single sensor, providing a more comprehensive and reliable assessment of the foundation pit's condition.

[0092] The system automates the entire process from data collection and processing to analysis and early warning, reducing manual intervention and improving monitoring efficiency and consistency.

[0093] Risk assessment algorithms based on fused data can automatically calculate the deformation risk level of each marked point (representing the area), providing a quantitative basis for engineering decisions and achieving an intelligent leap from monitoring to early warning.

[0094] The electronic skin uses flexible materials, which can fit well against the surface of the pit and adapt to certain deformations.

[0095] Fluorescent QR code markers are visible under specific lighting conditions. Combined with image processing algorithms, they can be reliably identified even in complex backgrounds. Furthermore, the QR code itself is unique, making it easy to track long-term changes at specific points.

[0096] Timestamp and frame rate synchronization ensures accurate temporal correspondence between data from different sources, providing a foundation for subsequent precise analysis and event backtracking.

[0097] Multispectral data (visible light, heat) combined with strain, pressure and other data can generate richer visualization products, such as displacement cloud maps, strain cloud maps, pressure distribution maps, temperature anomaly maps, etc., enabling engineers to intuitively understand the complex state of the foundation pit.

[0098] By integrating advanced electronic skin sensing technology and multispectral vision technology, comprehensive, accurate, real-time, and intelligent monitoring of foundation pit deformation has been achieved, significantly improving the safety level of foundation pit engineering and possessing significant engineering application value.

[0099] Example 2 also includes step 8, setting a deformation risk level threshold, and triggering an alarm when the deformation risk level exceeds the deformation risk level threshold.

[0100] The alarm mechanism ensures that foundation pit deformation can be identified and addressed before it reaches a critical state that could lead to safety issues. Compared to traditional methods, which may suffer from delayed analysis of monitoring data and errors in manual judgment, this automated threshold-triggered alarm significantly shortens the time from monitoring to response and improves the efficiency of risk management.

[0101] In Example 3, the composite thin film layer 11 is a polyimide-silicone composite thin film, the strain sensing layer 12 is a silver nanowire resistive mesh, the pressure signal layer 13 is a flexible capacitive pressure sensor array, the environmental sensing layer 14 is a flexible temperature and humidity sensor array, the power supply unit 15 supplies power to the strain sensing layer 12, the pressure signal layer 13 and the environmental sensing layer 14 respectively, and the surface functional layer 16 is a nano-antireflective film, with fluorescent QR code marking points printed using inkjet printing and ultraviolet curing processes.

[0102] Both polyimide (PI) and silicone possess excellent flexibility, allowing them to adhere tightly to the surface of the foundation pit. Even with surface irregularities, they ensure good contact between the sensor layer and the monitored object, reducing measurement errors. PI exhibits excellent resistance to high and low temperatures and chemical corrosion, while silicone provides good waterproof and moisture-proof properties. This composite application enhances the long-term stability and reliability of the electronic skin in complex and harsh outdoor foundation pit environments (such as exposure to sunlight and rain, and temperature fluctuations).

[0103] PI provides good mechanical strength and rigid support, while silicone provides flexibility and sealing. The composite structure ensures flexibility while also having a certain degree of tensile and tear resistance, making it less prone to damage during installation or when the foundation pit undergoes minor deformation.

[0104] The silver nanowire resistance grid uses silver nanowires (AgNWs) with extremely low resistivity and excellent conductivity. The resulting grid structure is highly sensitive to minute deformations and can accurately capture the tensile and compressive strains on the surface of the pit (reflected by changes in resistance), thus achieving high-precision strain measurement.

[0105] Mesh structures enable large-area, uniform strain sensing, facilitating the acquisition of continuous strain field distribution information. AgNWs are easily fabricated on flexible substrates, perfectly matching the flexible properties of composite thin film layers.

[0106] Sensor array structures can provide spatial pressure distribution information, not just single-point pressure. They can identify areas of concentrated pressure or trends of change, providing richer data for analyzing soil stress states. Capacitive sensors typically have good linear output and long-term stability. Flexible substrates allow sensor arrays to conform to curved surfaces, adapting to changes in the shape of the excavation pit surface, and are less prone to measurement interference from additional stress caused by their own stiffness.

[0107] Simultaneous monitoring of temperature and humidity allows for a more comprehensive capture of the impact of environmental changes on the foundation pit and the performance of the electronic skin itself.

[0108] Nanoscale antireflective coatings can reduce the reflection and scattering of light on the surface of electronic skin, improving the clarity and contrast of images captured by multispectral cameras (especially visible light images), which is beneficial for subsequent marker identification and image analysis.

[0109] In Example 4, the formula for calculating the local strain value in step 5 is:

[0110] ;

[0111] in It is a local strain value. It is the relative change in phase resistance, and k is a material constant.

[0112] Example 5, the formula for the pressure distribution on the surface of the foundation pit in step 5 is:

[0113] ;

[0114] Where P is the calculated pressure value. This is the baseline pressure value. It measures the capacitance value. This is the initial capacitance value.

[0115] Example 6: The specific steps of the image matching algorithm for calculating the displacement field on the surface of the foundation pit are as follows:

[0116] Step 51: Extract SIFT feature points from images of consecutive time phases;

[0117] Step 52: Establish the correspondence between feature points through nearest neighbor matching and consistency check;

[0118] Step 53: Evaluate the matching quality using a random sampling consensus algorithm and remove abnormal matching points;

[0119] Step 54: Calculate the displacement field on the surface of the foundation pit based on the matching point pairs.

[0120] In Example 7, step 6, the step of fusing the strain data monitored by the electronic skin with the displacement data monitored by the vision system is as follows:

[0121] Step 61: The electronic skin strain data and visual displacement data are fused using a weighted fusion method.

[0122] Step 62: Spatial registration of electronic skin monitoring points and visual monitoring points to establish a correspondence;

[0123] Step 63: Evaluate the uncertainties of the electronic skin data and visual data, respectively;

[0124] Step 64: Assign weights to the two data sources based on the uncertainty index and calculate the fusion result.

[0125] In Example 8, the formula for calculating the fusion result in step 64 is:

[0126] ;

[0127] in It's electronic skin data. It is visual displacement data. It is the weight of electronic skin data. It is the weight of visual displacement data. It is the result of fusion.

[0128] Example 9: The specific steps of the risk assessment algorithm based on fused data are as follows:

[0129] The fusion results of strain, displacement, and displacement rate indices are calculated based on the electronic skin strain data and visual displacement data fused from the corresponding fluorescent QR code markers, and the risk level is determined based on the risk level threshold range.

[0130] The fuzzy comprehensive evaluation method is adopted to integrate the risk assessment results of multiple indicators;

[0131] Based on the comprehensive evaluation results, the deformation risk level of the corresponding fluorescent QR code markers is determined.

[0132] Example 10: A visual monitoring and early warning system for foundation pit deformation, used to implement a visual monitoring and early warning method for foundation pit deformation. The visual monitoring and early warning system includes an electronic skin, a multispectral camera array 2, a computer monitoring server 3, and an alarm 4. The electronic skin includes a composite thin film layer 11, a strain sensing layer 12, a pressure signal layer 13, an environmental sensing layer 14, a power supply unit 15, and a surface functional layer 16. The bottom of the composite thin film layer 11 is attached to the surface of the foundation pit to be detected. The strain sensing layer 12 is laid on the top surface of the composite thin film layer 11. The pressure signal layer 13 is laid on the strain sensing layer 12 based on the isolation layer. The environmental sensing layer... 14. An isolation layer is laid on the pressure signal layer 13, and a surface functional layer 16 is laid on the environmental sensing layer 14. The power supply unit 15 supplies power to the strain sensing layer 12, the pressure signal layer 13, and the environmental sensing layer 14 respectively. The multispectral camera array 2 is deployed around the pit to be monitored. The computer monitoring server 3 is communicatively connected to the strain sensing layer 12, the pressure signal layer 13, the environmental sensing layer 14, and the multispectral camera array 2 respectively. Based on electronic skin data and visual displacement data, the deformation risk level of the corresponding fluorescent QR code marker point is output. If the deformation risk level exceeds the deformation risk level threshold, the alarm 4 is controlled to sound an alarm.

[0133] The following specific embodiments illustrate the implementation principle of the present invention:

[0134] A new subway station construction project in a city center has a foundation pit depth of approximately 25 meters. The surrounding environment is complex, with existing buildings and roads nearby. To ensure construction safety and the stability of the surrounding environment, it was decided to adopt the foundation pit deformation visual monitoring and early warning system of this invention.

[0135] Electronic Skin Installation: After cleaning the surface of the slope and bottom slab formed by the excavation of the foundation pit, a thin film made of polyimide-silicone composite is laid as a substrate (composite film layer 11). A resistive grid composed of silver nanowires (strain sensing layer 12) is precisely laid on top to sense surface strain. Above the strain sensing layer, a flexible capacitive pressure sensor array (pressure signal layer 13) is laid through an insulating material to monitor soil pressure. Then, a flexible temperature and humidity sensor array (environmental sensing layer 14) is laid through another insulating layer to monitor ambient temperature and humidity. Finally, a nano-antireflective membrane (surface functional layer 16) is covered on top. All sensing layers are powered by an integrated flexible power supply unit (power supply unit 15). Dozens of evenly distributed fluorescent QR code markers are printed on the surface functional layer using inkjet printing and UV curing processes.

[0136] Multispectral camera array deployment: An array of four multispectral cameras (multispectral camera array 2) was installed at a safe distance around the perimeter of the foundation pit. Each camera can simultaneously capture visible light images, thermal infrared images, and focus specifically on fluorescent QR code markers. The cameras synchronously capture images every 5 minutes to ensure data consistency in time.

[0137] Computer monitoring server connection: A high-performance computer monitoring server (computer monitoring server 3) is placed in the monitoring center near the construction site. The server is connected to the data acquisition module in the electronic skin via a wired network, and periodically receives resistance change signals, capacitance change signals, and environmental temperature and humidity data. At the same time, the server also receives image data (excavation pit images, thermal images, and marker point images) from a multispectral camera array via the network.

[0138] Data Acquisition: The electronic skin continuously monitors and sends data to the server. A multispectral camera array captures images at a set frequency and transmits them to the server.

[0139] Data synchronization: The electronic skin data and camera image data received by the server are all timestamped. The server first performs timestamp calibration to ensure that the data from different devices are synchronized in time for subsequent correlation analysis.

[0140] Mechanical data analysis: The server processes resistance change signals and calculates local strain values ​​for each region using known formulas, reflecting the tensile or compressive conditions on the foundation pit surface. Simultaneously, it processes capacitance change signals to calculate the pressure distribution map on the foundation pit surface, identifying areas of concentrated soil pressure.

[0141] Visual data analysis: The server processes the image data. First, image processing algorithms are used to identify and decode fluorescent QR code markers in the visible light image. Based on the positions of these markers, a spatial coordinate system corresponding to the pit surface is established. Then, image matching algorithms such as SIFT feature point extraction and matching, and RANSAC to remove mismatches are used to calculate the displacement of each marker in the continuous image, thereby obtaining the displacement field of the entire pit surface. In addition, the server also analyzes the thermal infrared image to identify areas of abnormal temperature, which may be related to leakage or stress concentration.

[0142] Ambient temperature and humidity data are used to correct strain and pressure data, eliminating or reducing measurement biases caused by environmental factors.

[0143] The server fuses strain data calculated by the electronic skin with displacement data calculated by the vision system. First, spatial registration is performed, mapping sensor points on the electronic skin to the nearest visual markers. Uncertainties in both data sources are evaluated (e.g., based on sensor accuracy, image matching confidence, etc.), and different weights are assigned to the strain and displacement data according to these uncertainties. Finally, a weighted fusion method is used to obtain the comprehensive deformation information for each corresponding region.

[0144] Risk Assessment and Early Warning: Based on the fused data, the server runs a risk assessment algorithm. This algorithm first calculates the strain, displacement, and rate of change of the area represented by each marker point, and compares these values ​​with preset risk level thresholds to initially determine the risk level. Then, a fuzzy comprehensive evaluation method is used to fuse the risk assessment results of these individual indicators to obtain a more comprehensive and objective overall risk level. For example, if the strain and displacement of a certain area both exceed the warning value, and the displacement rate accelerates, then the overall risk level of that area will be rated as "high".

[0145] Alarm and Visualization: The server displays the deformation risk level of each marker point in real time on the monitoring interface, typically using color coding (e.g., green - low risk, yellow - medium risk, red - high risk) on the digital model or image of the foundation pit. When the risk level of one or more marker points exceeds the preset alarm threshold (e.g., reaching "high" risk), the server immediately triggers the alarm (alarm 4), issuing an audible and visual alarm and pushing alarm information to project managers and relevant safety personnel, prompting them to take immediate inspection or reinforcement measures.

[0146] Through this system, the subway foundation pit project achieved continuous, automated, and multi-dimensional monitoring of the entire excavation face, enabling timely detection of potential deformation risks and ensuring the safe and smooth progress of the project.

[0147] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method for visual monitoring and early warning of foundation pit deformation, characterized in that: The method comprises the following steps: Step 1: covering the surface of the foundation pit to be monitored with an electronic skin; the electronic skin comprises a composite film layer (11), a strain sensing layer (12), a pressure signal layer (13), an environment sensing layer (14), a power supply unit (15) and a surface functional layer (16); a multi-spectral camera array (2) is deployed around the foundation pit to be monitored; a fluorescent two-dimensional code marker point is printed on the surface functional layer (16); Step 2: the strain sensing layer (12) and the pressure signal layer (13) on the electronic skin monitor the strain and the pressure change of the soil on the surface of the foundation pit, and generate resistance change signals and capacitance change signals; the environment sensing layer (14) collects environmental temperature and humidity data; the multi-spectral camera array (2) synchronously shoots pit images, thermal images and marker point images according to a set frequency; Step 3: a computer monitoring server (3) collects electronic skin data every set time, the electronic skin data comprising resistance change signal data, capacitance change signal data and environmental temperature and humidity data, and synchronously collects multi-spectral camera array (2) visual data, the multi-spectral camera array (2) visual data comprising pit images, thermal images and marker point images; Step 4: the computer monitoring server (3) synchronizes the frame rate of the time stamp of the electronic skin data and the visual data of the multi-spectral camera array (2); Step 5: calculating local strain values based on the resistance change signal data; calculating the pressure distribution on the surface of the foundation pit based on the capacitance change signal data; identifying the fluorescent two-dimensional code marker point through an image processing algorithm, establishing a spatial coordinate system, calculating the displacement field on the surface of the foundation pit through an image matching algorithm, and analyzing the temperature distribution and abnormal areas in the thermal infrared image; Step 6: fusing the strain data monitored by the electronic skin and the displacement data monitored by the visual system; Step 7: outputting the deformation risk level of the fluorescent two-dimensional code marker point by using a risk assessment algorithm based on the fused data; In step 6, the step of fusing the strain data monitored by the electronic skin and the displacement data monitored by the visual system is: Step 61: fusing the electronic skin strain data and the visual displacement data by using a weighted fusion method: Step 62: spatially registering the electronic skin monitoring points and the visual monitoring points to establish a corresponding relationship; Step 63: respectively evaluating the uncertainty of the electronic skin data and the visual data; Step 64: according to the uncertainty index, assigning weights to the two data sources, and calculating the fusion result; The formula for calculating the fusion result in step 64 is: ； wherein is electronic skin data, is visual displacement data, is a weight of the electronic skin data, is a weight of the visual displacement data, is a fusion result; The specific steps of the risk assessment algorithm based on the fused data are as follows: Based on the fusion results of the electronic skin strain data and the visual displacement data corresponding to the fluorescent two-dimensional code marker point, the fusion results of the strain, displacement and displacement rate indexes are calculated, and the risk level is determined based on the risk level threshold interval; Using a fuzzy comprehensive evaluation method, the multi-index risk assessment results are fused; According to the comprehensive evaluation result, the deformation risk level of the corresponding fluorescent two-dimensional code marker point is determined. 2.The method according to claim 1, characterized in that: Step 8: setting a deformation risk level threshold, and alarming when the deformation risk level exceeds the deformation risk level threshold.

3. The method according to claim 2, characterized in that: The composite film layer (11) is a polyimide silica gel composite film, the strain sensing layer (12) is a silver nanowire resistance grid, the pressure signal layer (13) is a flexible capacitive pressure sensor array, the environment sensing layer (14) is a flexible temperature and humidity sensor array, the power supply unit (15) respectively supplies power to the strain sensing layer (12), the pressure signal layer (13) and the environment sensing layer (14), the surface functional layer (16) is a nano anti-reflection film, and a fluorescent two-dimensional code mark point is printed by using inkjet printing and ultraviolet curing process.

4. The visual monitoring and early warning method for foundation deformation according to claim 3, characterized in that: The formula for calculating the local strain value in step 5 is: ； wherein is the local strain value, is the relative change in phase resistance value, k is a material constant.

5. The visual monitoring and early warning method for foundation deformation according to claim 4, characterized in that: The formula for calculating the pressure distribution on the surface of the foundation in step 5 is: ； where P is a calculated pressure value, is a reference pressure value, is a measured capacitance value, is an initial capacitance value.

6. The method of claim 5, wherein the method further comprises: The specific steps for calculating the displacement field on the surface of the foundation by using the image matching algorithm are as follows: Step 51, SIFT feature points are extracted from the images of continuous time phases; Step 52, the corresponding relationship of the feature points is established through nearest neighbor matching and consistency test; Step 53, the matching quality is evaluated through the random sample consensus algorithm, and the abnormal matching points are removed; Step 54, the displacement field on the surface of the foundation is calculated based on the matching point pairs.

7. A visual monitoring and early warning system for foundation pit deformation, characterized in that: A visual monitoring and early warning method for foundation deformation is implemented according to claim 6, and the visual monitoring and early warning system comprises an electronic skin, a multi-spectral camera array (2), a computer monitoring server (3) and an alarm (4). The electronic skin comprises a composite film layer (11), a strain sensing layer (12), a pressure signal layer (13), an environment sensing layer (14), a power supply unit (15) and a surface functional layer (16). The bottom of the composite film layer (11) is attached to the surface of the foundation to be detected, the strain sensing layer (12) is laid on the top surface of the composite film layer (11), the pressure signal layer (13) is laid on the strain sensing layer (12) based on an isolation layer, the environment sensing layer (14) is laid on the pressure signal layer (13) based on an isolation layer, the surface functional layer (16) is laid on the environment sensing layer (14), and the power supply unit (15) respectively supplies power to the strain sensing layer (12), the pressure signal layer (13) and the environment sensing layer (14). The multi-spectral camera array (2) is arranged around the foundation to be monitored, the computer monitoring server (3) is communicatively connected with the strain sensing layer (12), the pressure signal layer (13), the environment sensing layer (14) and the multi-spectral camera array (2), respectively, a deformation risk level of the corresponding fluorescent two-dimensional code mark point is output based on the electronic skin data and the visual displacement data, and the alarm (4) is controlled to alarm if the deformation risk level exceeds a deformation risk level threshold.

Images