A non-contact monitoring device and method for road surface settlement

Abstract

The application discloses a non-contact monitoring device for road surface settlement, comprising a monitoring device and a plurality of reflective units; the plurality of reflective units are arranged on monitoring points located in a monitoring area and reference points located in a stable area; the monitoring device comprises: an image acquisition device, which is used for photographing the reference points and each shooting position before and during monitoring to obtain an initial reference image and a target image; a controller, which acquires absolute angles of each shooting position when the initial reference image is photographed, judges whether the monitoring points are shielded in the target image according to the initial reference image, and calculates the settlement amount of the monitoring points according to the initial reference image and the effective target image after obtaining the effective target image without shielding of the current shooting position. The absolute angles are switched to the next shooting position until the settlement amount calculation of all the monitoring points is completed. The application realizes real-time monitoring of the road surface settlement under the normal traffic conditions of vehicles, which is non-contact, sub-millimeter and full road section covering.

Description

Technical Field

[0001] This invention relates to the field of road settlement monitoring technology, specifically to a non-contact monitoring device and method for road settlement. Background Technology

[0002] Highways are highly susceptible to structural settlement when traversing areas with mining subsidence or expansive soil subgrades, or when encountering natural disasters such as torrential rain or earthquakes, or when exhibiting obvious road defects or facing disturbances from underpass construction. This can threaten driving safety. Therefore, it is necessary to employ scientific methods to dynamically monitor road surface settlement.

[0003] Current mainstream monitoring technologies have significant limitations:

[0004] (1) Contact measurement

[0005] While measurement robots rely on the deployment of reflecting prisms to ensure measurement accuracy, the placement of prisms on the road surface can easily cause traffic safety hazards, thus limiting their practical application. Leveling requires manual operation and temporary lane closures, resulting in low efficiency. Distributed fiber optic monitoring technology requires the pre-embedding of sensor cables within the road structure layer, leading to high construction costs and complex maintenance. Traditional fixed image-based monitoring technology uses wired infrared lamps as measurement points, which cannot be widely used in the field of highway pavement settlement monitoring due to high power consumption, wiring, and exposure affecting traffic safety. Furthermore, fixed image-based methods have insufficient field of view coverage, requiring multiple units for deployment, resulting in high monitoring costs. Additionally, the measurement accuracy is significantly affected by environmental interference (light / vibration) because it is difficult to set up benchmarks within the field of view.

[0006] (2) Non-contact monitoring

[0007] GNSS technology is limited by satellite signal obstruction, and can only be deployed on both sides of the road, unable to cover the driving surface area, and the cost of a single point device exceeds 10,000 yuan; although InSAR satellite remote sensing can achieve large-scale monitoring, it is affected by atmospheric turbulence and vegetation cover, and the monitoring accuracy is only at the sub-centimeter level, and the data procurement cost is high. Summary of the Invention

[0008] To overcome the problems of existing technologies that are difficult to balance non-contact, high precision, full road coverage, and economy, this invention provides a non-contact monitoring device and method for road surface settlement. Through a monitoring device, non-contact, full road coverage, and real-time monitoring of road surface settlement can be achieved under normal vehicle traffic conditions.

[0009] According to one aspect of the present invention, a non-contact monitoring device for road surface settlement is provided, comprising:

[0010] The monitoring equipment includes a rotating table, a controller, and a light-emitting unit and an image acquisition device mounted on the rotating table;

[0011] Several reflective units are used to be deployed at various monitoring points in the monitoring area and at reference points in the stable area.

[0012] The light emitted by the light-emitting unit is reflected to the image acquisition device;

[0013] The image acquisition device is used to capture initial reference images of the reference point and initial reference images of each shooting position before monitoring, and to capture target images of the reference point and target images of each shooting position during monitoring.

[0014] The controller is used to acquire the absolute angle of each shooting position when shooting the initial reference image, and to determine whether each monitoring point in the target image is occluded based on the initial reference image of the current shooting position when shooting the target image. If it is determined that there are monitoring points occluded in the target image, new target images are continuously shot until all monitoring points in the new target image are unoccluded, so as to obtain a valid target image of the current shooting position. The controller calculates the settlement of all monitoring points covered by the current shooting position based on the initial reference image and target image of the reference point, the initial reference image and valid target image of the current shooting position. The controller switches to the next shooting position based on the absolute angle until the settlement of all monitoring points covered by all shooting positions within a monitoring cycle is calculated.

[0015] The above technical solution uses a monitoring device to capture images of the reference point and each monitoring point before and during monitoring, obtaining unobstructed initial reference images and target images. Based on the unobstructed initial reference image, it determines whether each monitoring point in the target image is obstructed, and obtains the absolute angle of each shooting position when the initial reference image was captured. After obtaining an unobstructed and valid target image of all monitoring points in the monitoring area of ​​the current shooting position, it switches to the next shooting position based on the absolute angle until all monitoring points are covered. Finally, it calculates the settlement of the monitoring points based on the initial reference image and the valid target image, thereby realizing non-contact, full-section coverage real-time monitoring of road surface settlement under normal vehicle traffic conditions.

[0016] In this invention, specifically, reflective units are first deployed at each monitoring point in the settlement area and at a reference point in the stable area. Before monitoring, a rotating platform is rotated to optimize the number of monitoring points observable within the field of view of the image acquisition device. During a vehicle-free interval, an unobstructed initial reference image is captured using the image acquisition device. This process is repeated until all monitoring points are covered, and the controller records the shooting position of each initial reference image. During monitoring, the controller automatically controls the rotating platform to rotate according to the recorded shooting positions, thereby rotating the image acquisition device sequentially to each shooting position. Each time the device rotates to a new shooting position, an unobstructed target image (i.e., a valid target image) must be captured before the rotating platform moves to the next shooting position, until all monitoring points are covered. Finally, the controller calculates the settlement amount of the monitoring points based on the initial reference images and valid target images.

[0017] Furthermore, the rotary table is also used to send the absolute angles generated when rotating to each shooting position to the controller when shooting the initial reference image; when shooting the target image, the controller is used to control the rotation of the rotary table according to the absolute angle, thereby driving the image acquisition device to rotate to each shooting position in sequence.

[0018] Furthermore, the image acquisition device includes a camera and an infrared lens, used to acquire a target image or initial reference image containing only infrared light spots formed by reflection by several reflective units.

[0019] Furthermore, based on the initial reference image and target image of the benchmark point, and the initial reference image and effective target image of the current shooting location, the settlement of all monitoring points covered by the current shooting location is calculated using the following formula:

[0020] ;

[0021] Where S represents the actual settlement at the monitoring point; L represents the camera pixel size; L represents the distance from the monitoring point to the camera; f represents the lens focal length.

[0022] s represents the pixel displacement of the monitoring point. ;

[0023] In the formula, i represents the sequence number of the shooting location, n represents the number of monitoring cycles, and the superscript j represents the sequence number of the monitoring point; This represents the y-coordinate of the j-th monitoring point in the valid target image; This represents the y-coordinate of the j-th monitoring point in the initial reference image; Represents the roll angle at the i-th shooting position; Represents the y-axis coordinate of the reference point; This represents the y-axis coordinate of the reference point in the initial reference image; This indicates the roll angle at the shooting position of the reference point in the captured effective target image.

[0024] Through the above solution, the present invention achieves non-contact, sub-millimeter level, and full-section coverage real-time monitoring of road surface settlement under normal vehicle traffic conditions.

[0025] Furthermore, the monitoring device also includes an angle sensor connected to the image acquisition device, which is used to obtain the roll angle of the image acquisition device's shooting position.

[0026] According to one aspect of the present invention, a non-contact monitoring method for road surface settlement is provided, comprising: deploying reflective units at various monitoring points located in a monitoring area and at reference points located in a stable area; capturing initial reference images of the reference points and initial reference images of each shooting position, and obtaining the absolute angle of each shooting position; capturing a target image of the reference points; capturing a target image of the current shooting position, and determining whether each monitoring point in the target image is obstructed based on the initial reference image of the current shooting position; when it is determined that a monitoring point in the target image is obstructed, continuously capturing new target images until all monitoring points in the new target image are unobstructed, so as to obtain a valid target image of the current shooting position; calculating the settlement amount of all monitoring points covered by the current shooting position based on the initial reference image and target image of the reference points, the initial reference image of the current shooting position, and the valid target image; switching to the next shooting position based on the absolute angle, until the calculation of the settlement amount of all monitoring points covered by all shooting positions within a monitoring cycle is completed.

[0027] Further, determining whether each monitoring point in the target image is occluded based on the initial reference image at the current shooting position includes: analyzing the initial reference image and the target image at the current shooting position respectively using an image processing algorithm to obtain feature values ​​of the initial reference image and the target image; the feature values ​​include the spot area of ​​the monitoring point, spot contour information, sub-pixel level coordinates of the spot center, and spot grayscale gradient direction; determining whether each monitoring point in the target image is occluded based on the feature values ​​of the initial reference image and the target image.

[0028] Furthermore, if a monitoring point in the target image does not meet any of the following criteria, then the monitoring point is determined to be occluded:

[0029] (1) Neighborhood feature boundary integrity: The number of boundary break points is obtained by the difference between the number of pixels in the light spot contour of each monitoring point in the target image and the number of pixels in the light spot contour of the monitoring point in the initial reference image; when the number of boundary break points is greater than the first preset threshold, it is determined that the corresponding monitoring point does not meet the neighborhood feature boundary integrity.

[0030] (2) Local gradient consistency: Calculate the average value of the absolute angle difference between the gray gradient direction and the average gradient direction of each pixel in the spot corresponding to each monitoring point in the target image; when the average value is greater than or equal to the second preset threshold, it is determined that the corresponding monitoring point does not meet the local gradient consistency.

[0031] (3) Geometric features: Calculate the ratio of the area of ​​the spot at each monitoring point in the target image to the area of ​​the spot at the monitoring point in the initial reference image; when the area ratio is less than the third preset threshold, it is determined that the corresponding monitoring point does not meet the geometric features.

[0032] Furthermore, a template matching algorithm is used to establish the correspondence between each monitoring point in the effective target image and each monitoring point in the initial reference image.

[0033] Furthermore, the overall monitoring area, composed of the monitoring areas corresponding to each shooting location, covers all monitoring points, and the monitoring points covered by each monitoring area do not overlap.

[0034] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0035] (1) The measurement of the present invention is fully automatic and does not require manual intervention. It can also automatically process the occluded data, thereby improving the intelligence level of the product.

[0036] (2) When measuring, the road does not need to be closed in advance, and it does not affect the operation and use of the road and the normal passage of vehicles. Compared with the traditional road settlement measurement method, it effectively reduces the safety risks of measurement personnel working on the road.

[0037] (3) The location of the reference point of the present invention is not limited by the site, which can effectively solve the problem of decreased measurement accuracy caused by the inability to set reference points.

[0038] (4) The location of the monitoring equipment of the present invention is not limited by the site, and a set of equipment can cover the entire road section settlement monitoring through a high-precision one-dimensional rotating table, which reduces the monitoring cost.

[0039] (5) The present invention uses a combination of infrared lamps and reflective units, which effectively reduces the power consumption of the device compared with traditional power supply and light-emitting unit road signs, and is more suitable for low power consumption application environments under outdoor solar power supply mode. Attached Figure Description

[0040] Figure 1 This is a schematic diagram of a non-contact monitoring device for road surface settlement provided in an embodiment of the present invention.

[0041] Figure 2 This is a schematic diagram of the monitoring deployment scheme provided in an embodiment of the present invention.

[0042] Figure 3 This is a schematic diagram of the shooting position provided in an embodiment of the present invention.

[0043] Figure 4 This is a schematic diagram of the pixel displacement of the monitoring point before and after deformation, provided in an embodiment of the present invention.

[0044] Figure 5 This is a flowchart of a non-contact monitoring method for road surface settlement provided in an embodiment of the present invention. Detailed Implementation

[0045] The technical solutions of various embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0046] Please refer to Figure 1 This embodiment provides a non-contact monitoring device for road surface settlement. The monitoring device includes monitoring equipment and several reflective units. The monitoring equipment includes a rotating table, a controller, and light-emitting units and an image acquisition device mounted on the rotating table. The reflective units are deployed at various monitoring points located in the monitoring area (i.e., the settlement area) and at reference points located in the stable area, reflecting the light emitted by the light-emitting units to the image acquisition device. The image acquisition device is used to capture initial reference images of the reference points and initial reference images of each shooting position before monitoring. The image acquisition device is also used to capture target images of the reference points and target images of each shooting position during monitoring. The controller is used to acquire the absolute angles of each shooting position when capturing the initial reference image. When capturing the target image, it determines whether each monitoring point in the target image is occluded based on the initial reference image of the current shooting position. If it determines that a monitoring point in the target image is occluded, it continuously captures new target images until all monitoring points in the new target image are unoccluded, thus obtaining a valid target image of the current shooting position. Based on the initial reference image and target image of the reference point, and the initial reference image and valid target image of the current shooting position, it calculates the settlement of all monitoring points covered by the current shooting position. It switches to the next shooting position based on the absolute angle until the settlement of all monitoring points covered by all shooting positions within one monitoring cycle is completed.

[0047] In this embodiment, the light-emitting unit is an infrared lamp with a divergence angle of 7°, and the reflective unit is preferably a reflective road stud. For example... Figure 2 As shown, reflective road studs are installed at various monitoring points and reference points. The reflective road studs at the monitoring points are installed on the lane markings within the settlement area. Their installation specifications and conditions must meet the requirements of Article 10.17 of the "Specifications for the Installation of Highway Traffic Signs and Markings JTGD82-2009". Specifically, the height of the reflective road studs should generally be 10-25mm above the road surface, and the dimensions of the reflective road studs include, but are not limited to, 100mm. 100 20mm. The monitoring equipment is deployed on fixed observation piers outside the settlement area, while the reference point is deployed in a stable area (i.e., outside the settlement area) and closest to the monitoring equipment. Understandably, the monitoring point may be obstructed by vehicles, while the reference point is not. Furthermore, the placement of the reference point in this invention is not limited by the site, effectively solving the problem of decreased measurement accuracy caused by the inability to set up a reference point. The deployment of the monitoring equipment in this invention is also not limited by the site, and a single set of equipment can cover the entire road section's settlement monitoring through a high-precision rotating table, reducing monitoring costs.

[0048] The image acquisition device includes a camera and an infrared lens, used in conjunction with a light-emitting unit. The light-emitting unit emits infrared light towards reflective spikes located at a reference point and various monitoring points. After retroreflection by the reflective spikes, the light is captured and focused by the infrared lens, and finally imaged by the camera to obtain a target image or initial reference image containing only infrared light spots formed by reflections from several reflective units. The infrared lens captures and focuses only infrared light and filters out interference from other wavelengths, resulting in better image quality and facilitating subsequent spot image analysis. Furthermore, the rotating stage sends the absolute angles generated at each shooting position to the controller when capturing the initial reference image; when capturing the target image, the controller controls the rotation of the stage based on the absolute angles, thereby rotating the image acquisition device sequentially to each shooting position. In this embodiment, the infrared lens is a narrowband infrared lens with a bandpass filtering range of 850nm±10nm to avoid interference from ambient light. The camera is an industrial camera with a frame rate of 40fps. The controller is a high-performance industrial computer. The rotary table is a one-dimensional rotary table, rigidly connected to the camera. The angular measurement range of the one-dimensional rotary table is 0~360°, and the repeatability accuracy reaches 0.5''. Understandably, the angular measurement range of the one-dimensional rotary table, the bandpass filtering range of the infrared lens, the frame rate of the camera, and the type of controller can all be set according to actual needs, and are not limited here.

[0049] The monitoring equipment also includes an angle sensor, a 4G module, and a solar panel. The angle sensor is connected to the camera and is used to obtain the camera's roll angle. The solar panel powers the entire monitoring equipment. The 4G module provides a network for data uploading to the monitoring equipment. The angle sensor obtains the camera's roll angle with an accuracy of 0.001°. Furthermore, because this invention uses a combination of infrared lights and reflective road studs, this combination effectively reduces the device's power consumption and is more suitable for low-power applications in outdoor solar-powered environments.

[0050] Please refer to the following: Figure 5 The present invention also provides a non-contact monitoring method for road surface settlement, comprising the following steps (including steps S10-S70):

[0051] Step S10: Install reflective units at each monitoring point in the monitoring area and at the reference point in the stable area.

[0052] In step S10, reflective road studs are installed according to a preset layout plan. The preset layout plan involves placing monitoring points on the lane markings within the monitoring area, setting reference points in a stable area at a certain distance from the road surface settlement influence zone, setting observation piers in a stable area at a certain distance from the road surface settlement influence zone, and installing the monitoring equipment on the observation piers. The monitoring equipment is powered on and then adjusted using the observation piers. In this embodiment, the reference point is set at the initial shooting position (i.e., the starting position) of the monitoring equipment or at any other fixed shooting position, so that the monitoring equipment can quickly rotate to that fixed shooting position to photograph the reference point.

[0053] Step S20: Take initial reference images of the reference point and initial reference images of each shooting position, and obtain the absolute angle of each shooting position.

[0054] In step S20, the initial reference image is captured manually. Manual capture allows the human eye to directly determine whether the captured initial reference image of the reference point and the initial reference images of each shooting position are occluded. In other words, in reality, the initial reference image of the reference point and the initial reference images of each shooting position are unoccluded images.

[0055] like Figure 3As shown, starting from the reference point, the device obtains an initial reference image of the reference point upon power-on. Existing image processing algorithms, including but not limited to grayscale processing, edge detection, and sub-pixel level geometric center recognition and positioning, are used to locate the reference point in the initial reference image, obtaining its feature values. The reference point is presented as a light spot in both the initial reference image and the target image. The feature values ​​include, but are not limited to, the light spot area, light spot contour information, sub-pixel level coordinates of the light spot center, and the light spot grayscale gradient direction. The controller automatically stores the calculated initial reference image feature values ​​in the reference point's data storage folder. The controller establishes a system with the upper left corner of the initial reference image as the origin O, the vertical axis as the y-axis, and the horizontal axis as the x-axis, as shown below. Figure 4 The image coordinate system shown is xoy. The pixel coordinates of the reference point are... The subscript 0 indicates that the sequence number of the current image is 0, and the image with sequence number 0 is the initial reference image.

[0056] like Figure 3 As shown, after capturing the initial reference image of the benchmark point, the operator rotates the one-dimensional rotary table to optimize the number of observable points within the camera's field of view, and records the absolute angle of the rotary table at this point. The location is marked as position 1, and the controller automatically generates a data storage folder for position 1. During a vehicle-free interval, an unobstructed initial reference image is captured at position 1. At this time, existing image processing algorithms, including but not limited to grayscale processing, edge detection, and sub-pixel level geometric center recognition and localization, are used to locate the monitoring points in the initial reference image to obtain the feature values ​​of the monitoring points. The controller automatically stores the calculated initial reference image feature values ​​into the data storage folder at position 1. The pixel coordinates of the monitoring points are... , which represents the pixel coordinates of the j-th monitoring point in the initial reference image. Here, the superscript j is the index of the monitoring point. Repeat the above operation until all remaining monitoring points are covered. The absolute angle of each position i is represented by... The positions 1, 2...i mentioned above represent the camera's shooting positions. The overall monitoring area, composed of the monitoring areas corresponding to each shooting position, covers all monitoring points, and the points covered by each monitoring area do not overlap. The absolute angle refers to the angle at which the camera rotates to each shooting position. , ... .

[0057] Step S30: Capture the target image of the reference point.

[0058] In step S30, the reference point of the present invention is set at the initial shooting position (i.e., the starting position) of the monitoring device or at any other fixed shooting position. Therefore, the monitoring device can quickly rotate and position itself to the fixed shooting position to shoot the reference point. Furthermore, the reference point of the present invention is laid out without obstruction; therefore, there is no need to determine whether the reference point is obstructed.

[0059] Step S40: Capture a target image at the current shooting location, and determine whether each monitoring point in the target image is occluded based on the initial reference image at the current shooting location.

[0060] In step S40, after capturing the target image of the reference point, the controller automatically controls the rotary table to rotate to the first shooting position based on the absolute angle and captures the target image at the current shooting position. Then, it analyzes the target image using existing image processing algorithms, including but not limited to grayscale processing, edge detection, and geometric center sub-pixel level recognition and positioning, to obtain the feature values ​​of the target image. The feature values ​​include, but are not limited to, spot area, spot contour information, spot center sub-pixel level coordinates, and spot grayscale gradient direction. Among them, the spot area, spot contour information, and spot grayscale gradient direction are used to determine whether the monitoring points are occluded. The spot center sub-pixel level coordinates refer to the center coordinate values ​​of the spot obtained by geometric center recognition and positioning. After being determined to be a valid target image, it is used to calculate the image pixel displacement of the monitoring points, thereby obtaining the settlement amount. The controller also reads the feature values ​​of the initial reference image from the data storage folder and uses the feature values ​​of the initial reference image as the reference template. It uses a template matching algorithm to compare and correlate the feature values ​​of the initial reference image with the feature values ​​of the current target image to determine whether each monitoring point in the current target image is occluded.

[0061] If a monitoring point in the target image does not meet any of the following criteria, the monitoring point is determined to be occluded. In other words, if a monitoring point in the target image fails to meet any of the following criteria: neighborhood feature boundary integrity, local gradient consistency, or geometric features, it is determined that a monitoring point in the target image is occluded and needs to be re-captured.

[0062] (1) Neighborhood Feature Boundary Integrity: The number of boundary breakpoints is obtained by comparing the number of pixels in the light spot contour of each monitoring point in the target image with the number of pixels in the light spot contour of the monitoring point in the initial reference image. When the number of boundary breakpoints is greater than a first preset threshold, the corresponding monitoring point is determined to not conform to the neighborhood feature boundary integrity. In this embodiment, the criterion for neighborhood feature boundary integrity is: when the number of boundary breakpoints is greater than 5% of the number of contour points, it is considered to be occluded and does not conform to the neighborhood feature boundary integrity; when the number of boundary breakpoints is less than or equal to 5% of the number of contour points, it is considered not to be occluded and conforms to the neighborhood feature boundary integrity. It can be understood that the first preset threshold of 5% of the number of contour points is only an example, and the first preset threshold can be set according to actual needs, and is not limited here.

[0063] (2) Local gradient consistency: Calculate the average value of the absolute angle difference between the gray-level gradient direction and the average gradient direction of each pixel in the spot corresponding to each monitoring point in the target image. When the average value is greater than or equal to the second preset threshold, the corresponding monitoring point is determined to not meet the local gradient consistency requirement. In this embodiment, for the spot of a certain monitoring point, the average gradient direction of each pixel in the spot is first calculated: ;in, This represents the average gradient direction of each pixel in the light spot, and N represents the total number of pixels in the light spot. This represents the direction of the grayscale gradient of the k-th pixel in the light spot (i.e., the directions of the grayscale gradients of the N pixels in the light spot are respectively...). , , ......, ( Normalized to [0, 2π).

[0064] Furthermore, the local gradient discontinuity (i.e., gradient direction dispersion) is defined as the average of the absolute angle differences between the gray-level gradient directions of each pixel in the light spot and the average gradient direction, as shown in the formula: ,in, Indicates the gradient direction dispersion; This represents the absolute angle difference between the gray-level gradient direction of the k-th pixel in the light spot and the average gradient direction. The criterion for local gradient consistency is: when (When the gradient direction is relatively consistent and the edge continuity is good, it is considered unoccluded and meets the local gradient consistency requirement; when If the gradient direction is inconsistent, it is considered occlusion and does not conform to local gradient consistency. Understandably, the second preset threshold of 5° is merely an example; the second preset threshold can be set according to actual needs and is not limited here.

[0065] (3) Geometric Features: Calculate the ratio of the area of ​​the spot at each monitoring point in the target image to the area of ​​the spot at the monitoring point in the initial reference image. When the area ratio is less than a third preset threshold, the corresponding monitoring point is determined to not meet the geometric features. In this embodiment, the formula for calculating the area ratio is: In the formula, Indicates the area ratio; This represents the area of ​​the light spot at the j-th monitoring point in the m-th target image. Let represent the area of ​​the light spot at the j-th monitoring point in the initial reference image. The criterion for geometric features is: occlusion of the light spot will cause the area of ​​the light spot to decrease, when... When this occurs, the light spot is considered to be blocked, which does not conform to geometric characteristics. At this point, the light spot is considered unobstructed and conforms to geometric characteristics. Understandably, the third preset threshold of 0.98 is merely an example; the third preset threshold can be set according to actual needs and is not limited here.

[0066] Step S50: When it is determined that there are monitoring points that are blocked in the target image, new target images are continuously captured until all monitoring points in the new target images are unblocked, so as to obtain a valid target image of the current shooting position.

[0067] In step S50, it is understood that the target image is an image captured at the current shooting location, and a valid target image refers to a target image in which all monitoring points are not obstructed by vehicles. It is understood that a target image with obstructed monitoring points cannot be used for analyzing the settlement of the road surface at that point and is considered an invalid target image. A target image with all monitoring points unobstructed (i.e., a valid target image) is required for the analysis of road surface settlement. It is understood that this invention does not require prior road closure during measurement, effectively reducing the safety risks to measurement personnel working on vehicular roads.

[0068] When it is determined that all monitoring points in the target image are unobstructed, the target image is considered a valid target image, and the valid target image is automatically saved in the data storage folder at the current location. The data storage folder for each shooting location stores data information such as the absolute angle of the corresponding shooting location, the initial reference image, the corresponding initial reference image feature values, the valid target image, and the corresponding valid target image feature values.

[0069] Step S60: Calculate the settlement of all monitoring points covered by the current shooting position based on the initial reference image and target image of the benchmark point, the initial reference image and effective target image of the current shooting position.

[0070] In step S60, an image coordinate system xoy is established with the upper left corner of the effective target image as the origin o, the vertical axis as the y-axis, and the horizontal axis as the x-axis. The formula for calculating the settlement of monitoring points in the effective target image is: Where S represents the settlement at the monitoring point; The sensor represents the camera pixel size; L represents the distance from the monitoring point to the camera. The distance from each monitoring point to the camera is different and needs to be measured separately using a laser rangefinder before the initial reference image is captured; f represents the lens focal length; and s represents the pixel displacement of the monitoring point. In the formula, i represents the sequence number of the shooting location; j represents the sequence number of the monitoring point; and n represents the number of monitoring cycles. This represents the y-coordinate of the j-th monitoring point in the valid target image; This represents the y-coordinate of the j-th monitoring point in the initial reference image; Represents the roll angle at the i-th shooting position; This represents the y-axis coordinate of a reference point in the target image; This represents the y-axis coordinate of the reference point in the initial reference image; This indicates the roll angle at the shooting location of the reference point. It should be noted that a monitoring cycle includes the acquisition and analysis of the target image at the reference point and the effective target images at each shooting location, as well as the calculation of settlement at the monitoring points.

[0071] It should be noted that in this embodiment, the target image and the initial reference image corresponding to each shooting position are stored and processed separately. A template matching algorithm is used to realize the correspondence between each monitoring point in the effective target image of the current shooting position and each monitoring point in the initial reference image, which is conducive to the unique identification and tracking of the monitoring points.

[0072] It should also be noted that the purpose of setting the reference point in this invention is primarily to eliminate deformation of the foundation at the camera's location, camera settlement, and environmental interference. Foundation deformation is creep deformation. Although the target image at the reference point and the effective target images at various locations in the monitoring area are not acquired at the same time, it can be considered that the foundation deformation is zero during the camera's shooting process within a monitoring cycle. However, monitoring is a long-term process (generally measured in months), and the cumulative deformation of the foundation during long-term monitoring cannot be ignored. This invention incorporates the y-axis coordinates of the reference point in the initial reference image and the reference point in the target image into the pixel displacement calculation formula, thereby eliminating foundation deformation, camera settlement, and environmental interference at the camera's location through the reference point.

[0073] Step S70: Switch to the next shooting position based on the absolute angle until the settlement of all monitoring points covered by all shooting positions within a monitoring cycle is calculated.

[0074] In step S70, after completing the initial reference image acquisition and feature value calculation, the controller's built-in measurement software will determine the measurement based on the absolute angle. The system automatically controls a one-dimensional rotating platform to sequentially rotate to each shooting position, and controls the camera to capture the target image at each shooting position. Steps S40-S60 are repeated until the settlement of all monitoring points covered by all shooting positions within a monitoring cycle is calculated. Steps S30-S70 are then continued to achieve the long-term monitoring goal. In essence, capturing the initial reference image is equivalent to pre-setting the shooting positions before monitoring. During subsequent monitoring, the monitoring equipment obtains the absolute angles and initial reference images of each shooting position, automatically performing real-time monitoring of road settlement based on these information. This eliminates the need for manual intervention and automatically processes obscured data, improving the product's intelligence level.

[0075] In summary, this invention uses a monitoring device to capture images of the reference point and each monitoring point before and during monitoring, obtaining unobstructed initial reference images and target images. Based on the unobstructed initial reference image, it determines whether each monitoring point in the target image is obstructed, and obtains the absolute angle of each shooting position when capturing the initial reference image. After obtaining an unobstructed and valid target image at the current shooting position, it switches to the next shooting position based on the absolute angle until all monitoring points are covered. Finally, it calculates the settlement of the monitoring points based on the initial reference image and the valid target image, thereby achieving non-contact, sub-millimeter level, and full-section coverage real-time monitoring of road surface settlement under normal vehicle traffic conditions.

[0076] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the technical solutions of the embodiments of the present invention.

Claims

1. A non-contact monitoring device for road surface settlement, characterized in that, include: The monitoring equipment includes a rotating table, a controller, and a light-emitting unit and an image acquisition device mounted on the rotating table; Several reflective units are used to be deployed at various monitoring points in the monitoring area and at reference points in the stable area. The light emitted by the light-emitting unit is reflected to the image acquisition device; The image acquisition device is used to capture initial reference images of the reference point and initial reference images of each shooting position before monitoring, and to capture target images of the reference point and target images of each shooting position during monitoring. The controller is used to acquire the absolute angles of each shooting position when capturing the initial reference image, and to determine whether each monitoring point in the target image is occluded based on the initial reference image of the current shooting position when capturing the target image. If it is determined that a monitoring point in the target image is occluded, new target images are continuously captured until all monitoring points in the new target image are unoccluded, thus obtaining a valid target image of the current shooting position. The controller calculates the settlement amount of all monitoring points covered by the current shooting position based on the initial reference image and target image of the reference point, and the initial reference image and valid target image of the current shooting position. It then switches to the next shooting position based on the absolute angle until the settlement amount calculation of all monitoring points covered by all shooting positions within one monitoring cycle is completed. The formula for calculating the settlement amount is: ； Where S represents the actual settlement at the monitoring point; L represents the camera pixel size; L represents the distance from the monitoring point to the camera; f represents the lens focal length. s represents the pixel displacement of the monitoring point. ; In the formula, i represents the sequence number of the shooting location, n represents the number of monitoring cycles, and the superscript j represents the sequence number of the monitoring point; This represents the y-coordinate of the j-th monitoring point in the valid target image; This represents the y-coordinate of the j-th monitoring point in the initial reference image; Represents the roll angle at the i-th shooting position; This represents the y-axis coordinate of a reference point in the target image; This represents the y-axis coordinate of the reference point in the initial reference image; It indicates the roll angle at the shooting position of the reference point in the captured effective target image.

2. The non-contact monitoring device for road surface settlement according to claim 1, characterized in that, The rotary table is also used to send the absolute angles generated when rotating to each shooting position to the controller when shooting the initial reference image; when shooting the target image, the controller is used to control the rotation of the rotary table according to the absolute angle, thereby driving the image acquisition device to rotate to each shooting position in sequence.

3. The non-contact monitoring device for road surface settlement according to claim 1, characterized in that, The image acquisition device includes a camera and an infrared lens, used to acquire a target image or initial reference image containing only infrared light spots formed by reflections by several reflective units.

4. The non-contact monitoring device for road surface settlement according to claim 1, characterized in that, The monitoring device also includes an angle sensor, which is connected to the image acquisition device and is used to obtain the roll angle of the image acquisition device's shooting position.

5. A non-contact monitoring method for road surface settlement, implemented based on a non-contact monitoring device for road surface settlement as described in any one of claims 1 to 4, characterized in that, include: Reflective units are deployed at each monitoring point in the monitoring area and at the reference point in the stable area; Take initial reference images of the reference point and initial reference images of each shooting position, and obtain the absolute angle of each shooting position; Capture target images of the reference point; Capture a target image at the current shooting location, and determine whether each monitoring point in the target image is occluded based on the initial reference image at the current shooting location; When it is determined that a monitoring point in the target image is obstructed, new target images are continuously captured until all monitoring points in the new target images are unobstructed, so as to obtain a valid target image of the current shooting position. Calculate the settlement of all monitoring points covered by the current shooting location based on the initial reference image and target image of the benchmark point, the initial reference image and effective target image of the current shooting location; The process continues by switching to the next shooting position based on the absolute angle, until the settlement of all monitoring points covered by all shooting positions within a monitoring cycle is calculated.

6. The non-contact monitoring method for road surface settlement according to claim 5, characterized in that, Determining whether each monitoring point in the target image is occluded based on the initial reference image at the current shooting location includes: Image processing algorithms are used to analyze the initial reference image and the target image at the current shooting location to obtain the feature values ​​of the initial reference image and the target image. The feature values ​​include the spot area of ​​the monitoring point, the spot contour information, the sub-pixel level coordinates of the spot center, and the gray-level gradient direction of the spot. Based on the feature values ​​of the initial reference image and the target image, it is determined whether each monitoring point in the target image is occluded.

7. The non-contact monitoring method for road surface settlement according to claim 6, characterized in that, If a monitoring point in the target image does not meet any of the following criteria, the monitoring point is determined to be occluded: (1) Neighborhood feature boundary integrity: The number of boundary break points is obtained by the difference between the number of pixels in the light spot contour of each monitoring point in the target image and the number of pixels in the light spot contour of the monitoring point in the initial reference image; when the number of boundary break points is greater than the first preset threshold, it is determined that the corresponding monitoring point does not meet the neighborhood feature boundary integrity. (2) Local gradient consistency: Calculate the average value of the absolute angle difference between the gray gradient direction and the average gradient direction of each pixel in the spot corresponding to each monitoring point in the target image; when the average value is greater than or equal to the second preset threshold, it is determined that the corresponding monitoring point does not meet the local gradient consistency. (3) Geometric features: Calculate the ratio of the area of ​​the spot at each monitoring point in the target image to the area of ​​the spot at the monitoring point in the initial reference image; when the area ratio is less than the third preset threshold, it is determined that the corresponding monitoring point does not meet the geometric features.

8. The non-contact monitoring method for road surface settlement according to claim 5, characterized in that, A template matching algorithm is used to establish the correspondence between each monitoring point in the effective target image and each monitoring point in the initial reference image.

9. The non-contact monitoring method for road surface settlement according to claim 5, characterized in that, The overall monitoring area, composed of the monitoring areas corresponding to each shooting location, covers all monitoring points, and the monitoring points covered by each monitoring area do not overlap.

Images