CN116815835A

CN116815835A - Deep horizontal displacement monitoring system and monitoring method based on machine vision

Info

Publication number: CN116815835A
Application number: CN202311082474.0A
Authority: CN
Inventors: 陈静; 万世平; 陈耀洲; 章芯; 曹祥寅
Original assignee: Sichuan Hezhong Precision Technology Co ltd
Current assignee: Sichuan Hezhong Precision Technology Co ltd
Priority date: 2023-08-28
Filing date: 2023-08-28
Publication date: 2023-09-29
Anticipated expiration: 2043-08-28
Also published as: CN116815835B

Abstract

The invention relates to the field of foundation pit monitoring, in particular to a deep horizontal displacement monitoring system and a monitoring method based on machine vision; the deep horizontal displacement monitoring system based on machine vision comprises an inclinometer pipe and measuring joints, wherein the inclinometer pipe is arranged in a foundation pit, a plurality of measuring joints connected in series are arranged in the inclinometer pipe, a camera is arranged at the lower end of each measuring joint, and a plurality of LED lamps are arranged at the upper end of each measuring joint; in the monitoring method, the LED lamps of the lower measuring section are shot by the camera of the upper measuring section, the shot images are uploaded to the data acquisition unit, and the horizontal displacement is obtained after processing and calculation. The beneficial effects achieved by the invention are as follows: the method provides a new thought for monitoring the horizontal displacement of the foundation pit in the inclinometer pipe, has low cost and high measurement accuracy, and is suitable for practical measurement of foundation pit engineering.

Description

Deep horizontal displacement monitoring system and monitoring method based on machine vision

Technical Field

The invention relates to the technical field of deep horizontal displacement monitoring in foundation pit engineering, in particular to a deep horizontal displacement monitoring system and a monitoring method based on machine vision.

Background

The current common measuring equipment for monitoring deep horizontal displacement is a sliding inclinometer. The sliding inclinometer is provided with a pair of pulleys at the upper part and the lower part, the upper wheel distance and the lower wheel distance are 500mm, and the working principle is based on gravity and electronic technology, and the inclination angle of an object is determined by measuring the change of the acceleration of gravity.

Slide inclinometers are typically comprised of one or more accelerometers and a microprocessor. An accelerometer is a sensor capable of measuring the acceleration of an object, which can detect changes in acceleration of the object in different directions. The microprocessor is responsible for processing the data of the accelerometer and converting the data into an inclination angle, and according to the difference value of the two inclination angles of the probe at the same position, the horizontal displacement value of the structure where the probe is positioned can be calculated.

The method for monitoring the deep horizontal displacement of the soil body around the supporting structure or the foundation pit adopts a method of embedding the inclinometer in the supporting structure or the soil body, and continuously measures the horizontal displacement variation at each depth by the sliding inclinometer, wherein the variation can reflect the deep displacement variation of the supporting structure.

When the sliding inclinometer is used for measuring the deep horizontal displacement of a supporting structure or soil body, the sliding inclinometer has the following defects: 1. the technical staff is required to go to the field to operate, the measurement frequency is low, and the real-time measurement requirement cannot be met; 2. because the installation of the inclinometer tube can not be completely guaranteed, the chute of the inclinometer tube can be distorted, and the position of the inclinometer tube where the probe is positioned can be changed and sometimes can be separated from the chute during each measurement, so that the monitoring precision of deep horizontal displacement can be influenced.

In order to solve the above-mentioned shortcomings of the sliding inclinometer: some companies now develop sliding inclinometers, and use a lifting robot fixed at the top opening of the inclinometer to lift the inclinometer, so that the problem of manual measurement is solved, the measurement frequency is improved, and the measurement accuracy is not particularly ideal. The other solution is that a plurality of inclinometers are connected in series and are arranged in the inclinometer pipe at one time, lifting operation is not carried out on the inclinometers, and the positions of all probes are kept motionless for a long time.

Based on the above, the company designs a novel inclinometer, improves the inclinometer and a measuring method, largely discards an inclination sensor on which the traditional inclinometer depends, changes the machine vision principle, and realizes accurate measurement on the basis of ensuring low cost.

Disclosure of Invention

The invention aims to overcome the defects of the prior art, and provides a deep horizontal displacement monitoring system and a monitoring method based on machine vision, which solve the problems of low cost and high precision.

It should be noted that: when the sliding inclinometer is used for measuring in a sliding manner in the inclinometer pipe, measurement inaccuracy can be caused whether the sliding inclinometer is lifted manually or lifted by a lifting robot; specifically, each time the point of measurement changes slightly, for example, 10m from the bottom of the pit, there is an error in the sliding inclinometer lifting to this position, in fact, at a position (possibly slightly higher, possibly slightly lower) adjacent to the 10m height. Thus, the measured data is not 10m data; after the measurement of each measuring section is finished, the drawn horizontal displacement and depth H change curve along with time cannot accurately reflect the real situation.

The measurement accuracy can be ensured by using a mode of connecting a plurality of probes (inclinometers) in series and then placing the probes (inclinometers) in a foundation pit to keep the probes stationary so as to perform measurement, but the cost is very high; the method is suitable for experiments and is not suitable for actual engineering measurement. Specifically, the probe measurement is realized by a built-in chip, the price of one chip meeting the precision requirement is 700-800 yuan, and the price of one probe is 3000-4000 yuan; if multiple probes are connected in series, the cost is prohibitive.

(first aspect)

The embodiment of the disclosure provides a deep horizontal displacement monitoring system based on machine vision.

In the embodiment of the disclosure, a deep horizontal displacement monitoring system based on machine vision comprises an inclinometer pipe and a measuring section; the plurality of measuring joints are connected through the hose to form a connecting string, and the connecting string is arranged in the inclinometer pipe for measurement;

wherein, a chute is axially arranged at the pipe wall of the inclinometer pipe; burying an inclinometer pipe in a supporting structure, wherein the supporting structure is positioned in a foundation pit, and the radial direction of a chute of the inclinometer pipe is perpendicular to the empty surface of the foundation pit;

the measuring section is provided with a camera at the lower end and a plurality of LED lamps at the upper end, and rollers are arranged at two sides of the measuring section; the camera is a digital camera; in the measuring section, the plurality of LED lamps comprise LED target lamps and auxiliary LED lamps, and the LED lamps are positioned on the same plane and are not positioned on the same straight line in the plane;

When assembled, the components are as follows: a plurality of measuring joints are connected through hoses to form a connecting string; the connecting string is placed in the inclinometer pipe, and the roller is matched with the chute; the data lines and the power lines of the measuring sections are connected with the data buses outside the measuring sections, the data acquisition units are arranged outside the supporting structure and are responsible for controlling the working of the measuring sections, receiving images, processing the images and uploading the images to the cloud back stage;

the measurement is as follows: the measuring section positioned at the upper position shoots the LED lamp of the measuring section adjacent to the measuring section positioned at the lower position through the camera, and the shot image is uploaded to the data acquisition unit; and calculating coordinates of LEDs in the image, and calculating the space displacement of each measuring section after calculation according to the coordinates of the LED lamp images of each measuring section at different times, and calculating the space displacement of each measuring section of the connecting string to obtain the deep displacement of the support structure.

For example, according to an embodiment of the present disclosure, the measuring section includes a metal housing, a camera, a data transmission interface, a plurality of LED lights; a camera is arranged in the lower end of the metal shell, and an LED target lamp and an auxiliary LED lamp are arranged in the radial section of the upper end of the metal shell; the position connecting lines of the LED lamps are not in a straight line, and the distance between the LED lamps needs to be accurately measured; the metal shell is provided with a data transmission interface electrically connected with the camera, and the data acquisition unit is in butt joint with the data transmission interface through a cable; and two sides of the metal shell are provided with one end of the torsion spring which is fixed on the outer cylindrical surface of the metal shell, and the other end of the torsion spring is provided with a roller.

For example, according to embodiments of the present disclosure, the camera has a CCM/CMOS sensor located in the center of the two-sided scroll wheel wiring.

For example, according to embodiments of the present disclosure, the CCD/CMOS sensor of the camera is rectangular, one side of the rectangle being parallel to the two-sided pulley wiring.

For example, according to embodiments of the present disclosure, the CCD sensor of the camera is disposed horizontally, which is no more than 10mm from the plane in which the LED lamp is located.

For example, according to embodiments of the present disclosure, the spacing between the measurement nodes is 500mm or 1000mm. In the connection string, the hose is connected with the measuring joint in a sealing way.

For example, according to an embodiment of the present disclosure, the data acquisition unit is located in the data acquisition unit DTU, and the data acquisition unit DTU is placed at a wellhead position of the foundation pit; the data acquisition unit is responsible for controlling the work of each unit section, receiving the image data of each measurement section and analyzing and calculating the received data; or the data acquisition unit is responsible for controlling each measuring section to work, receiving images, processing the images and uploading the images to the cloud platform, and the cloud platform analyzes, calculates and processes the data.

(second aspect)

The embodiment of the disclosure provides a deep horizontal displacement monitoring system and a monitoring method based on machine vision, comprising the following steps:

S1, sliding a traditional inclinometer down into an inclinometer pipe along a sliding chute to perform initial measurement, and obtaining initial spatial position data of the inclinometer pipe; if the geometrical space shape of the inclinometer pipe is not considered, the step can be omitted;

selecting one stable end of the inclinometer as a starting point, and respectively arranging a first measuring section, a second measuring section and an nth measuring section from one end of the starting point of the measuring tube to the other end of the starting point of the measuring tube; determination of correspondence by conventional inclinometersInitial spatial coordinates of the LED target lamp of each measurement section; wherein the initial spatial coordinates of the first measurement node are (X ₁ ，Y ₁ ，H ₁ ），X ₁ 、Y ₁ LED target lamp plane coordinate representing 1 st measuring section, H ₁ The height of the LED target lamp is 1 st measuring section, and X is calculated because the first measuring section is the starting point ₁ =0，Y ₁ =0; wherein the initial spatial coordinates of the nth measurement node are (X _n ，Y _n ，H _n ），X _n 、Y _n Representing the plane coordinates of the LED target lamp of the nth measuring section, H _n The height of the LED target lamp is the nth measuring section;

it should be noted that, the coordinates and measurement corresponding to the step S1 have no substantial effect on the subsequent steps, and the step S1 mainly includes checking the degree of distortion of the inclinometer tube before the measurement (the procedure is also performed in other measurement modes);

s2, installing a deep displacement inclinometer system in the inclinometer pipe;

Lifting the uppermost end of the whole connecting string by using a pull rope, and placing the connecting string in the inclinometer pipe, so that the connecting string slides to the bottom of the inclinometer pipe along a chute by using a roller, the bottom of the inclinometer pipe is only provided with an LED target lamp, and each measuring section cannot move up and down in the whole measuring period; after the measurement sections are installed, the measurement sections are connected with an external data acquisition unit DTU;

selecting one stable end of the inclinometer pipe as a starting position, wherein the bottom of the inclinometer pipe can be selected, and the pipe orifice at the top of the inclinometer pipe can also be selected;

s3, shooting;

starting each measuring section camera to shoot for the first time, enabling the camera to shoot an LED target lamp of the last measuring section to obtain an image of the LED target lamp, wherein the image is an initial observation image or a last observation image, and transmitting the shot image to a data acquisition unit;

after a period of time, for example, after one day, starting each measuring section camera to shoot for the second time, shooting the adjacent LED target lamp below the measuring section camera through the camera to obtain an image of the LED target lamp, wherein the image is the current observation image, and transmitting the shot image to the data acquisition unit;

the position of the inclinometer is kept unchanged, cameras of all measuring sections are started to shoot according to a specified time interval, an LED target lamp of one measuring section is shot through the cameras, an image of the LED target lamp is obtained, and the shot image is transmitted to a data acquisition unit;

For processing the images to be simple and regularly circulated, the image shot by each measuring section at the first time is defined as an initial observation image (or the image shot at the last time relative to the observation image at the second time); the latest image shot by each measuring section is the current observation image, and the next new image shot by each measuring section is the last observation image; the last observation image shot by each measuring section of the second time is a first observation image;

after the latest observation image is uploaded each time, the data acquisition unit is used for analyzing and processing the current observation image and the change quantity of the last observation image of the LED target lamp, calculating the current horizontal offset and calculating the current accumulated horizontal displacement.

For example, according to the embodiment of the present disclosure, in step S3, after each uploading of the latest observation image, the data acquisition unit analyzes and processes the current observation image and the previous change amount of the observation image of the LED target lamp, calculates the current horizontal offset, and calculates the current accumulated horizontal offset, which means:

A. calculating coefficients;

a1, observing an image after the first shooting, and measuring coordinate values of the centers of the LED lamps in an image coordinate system;

providing an accurate distance between the LED target lamp and the second auxiliary LED lamp as d1 and an accurate distance between the LED target lamp and the third auxiliary LED lamp as d2 in each measuring section, wherein d1 and d2 are provided by equipment manufacturers, and the accuracy is better than 0.1mm;

When observing the image, obtaining image coordinates corresponding to the corresponding LED target lamp, the second auxiliary LED lamp and the third auxiliary LED lamp in the first measurement section to the nth measurement section in the first shooting; wherein, the image coordinates corresponding to the LED target lamp, the second auxiliary LED lamp and the third auxiliary LED lamp in the nth measurement section are (x) _n1-1 ,y _n1-1 ）、（x _n2-1 ,y _n2-1 ）、（x _n3-1 ,y _n3-1 ) In the subscript of the image coordinates, the first number represents the corresponding measurement section, the second number represents the LED target lamp if it is 1, the second auxiliary lamp if it is 2, the third auxiliary lamp if it is 3, and the third number represents the number of times of shooting;

a2, observing images after the second shooting, and measuring coordinate values of the centers of the LED lamps in an image coordinate system; calculating the coefficient of the actual distance corresponding to the single pixel near the LED target lamp;

when observing the image, obtaining image coordinates corresponding to the corresponding LED target lamp, the second auxiliary LED lamp and the third auxiliary LED lamp in the first measurement section to the nth measurement section in the second shooting; wherein, the image coordinates corresponding to the LED target lamp, the second auxiliary LED lamp and the third auxiliary LED lamp in the nth measurement section are (x) _n1-2 ,y _n1-2 ）、（x _n2-2 ,y _n2-2 ）、（x _n3-2 ,y _n3- 2) In the subscript of the image coordinates, the first number represents the corresponding measurement section, the second number represents the LED target lamp if it is 1, the second auxiliary lamp if it is 2, the third auxiliary lamp if it is 3, and the third number represents the number of times of shooting;

When calculating the coefficient of the actual distance corresponding to the single pixel near the LED target lamp, the coefficient of the actual distance corresponding to the single pixel near the LED target lamp of each measuring section in the second shooting is referred to, and the coefficient of the actual distance corresponding to the single pixel near the LED target lamp in the second shooting of the first measuring section to the nth measuring section is included; the coefficient= (d 1/horizontal distance of Led target lamp to second auxiliary Led lamp in image obtained at this time of shooting of corresponding measuring section+d2/horizontal distance of Led target lamp to third auxiliary Led lamp in image obtained at this time of shooting of corresponding measuring section)/2;

a3, sequentially obtaining image coordinates of the Led target lamp, the second auxiliary Led lamp and the third auxiliary Led lamp of each measurement section during the third shooting to the ith shooting in a mode of A2;

calculating to obtain a coefficient of the actual distance corresponding to a single pixel near the LED target lamp from the first measuring section to the nth measuring section after each shooting when shooting from the third shooting to the ith shooting in the A2 mode; wherein, the coefficient of the actual distance corresponding to a single pixel near the LED target lamp in the nth measurement section of the ith shot= (d 1/horizontal distance from the LED target lamp to the second auxiliary LED lamp in the image obtained by the nth measurement section at the time of the i shot+d2/horizontal distance from the LED target lamp to the third auxiliary LED lamp in the image obtained by the nth measurement section at the time of the i shot)/2;

B. After the second shooting and observation, calculating to obtain the horizontal offset of the second shooting of each measuring section relative to the first shooting; calculating to obtain the horizontal displacement of the second shooting of each measuring section relative to the datum point (0, 0);

the horizontal offset of the second shooting of each measuring section relative to the first shooting is calculated, and the horizontal offset is calculated: the displacement offset in the X direction relative to the previous time in the corresponding measurement section= (the X coordinate of the Led target lamp at the second time of the measurement section-the X coordinate of the Led target lamp at the first time of the measurement section) the coefficient of the actual distance corresponding to the single pixel near the Led target lamp in the measurement section at the second time of the imaging, the displacement offset in the Y direction relative to the previous time of the measurement section= (the Y coordinate of the Led target lamp at the second time of the measurement section-the Y coordinate of the Led target lamp at the first time of the measurement section) the coefficient of the actual distance corresponding to the single pixel near the Led target lamp in the measurement section at the second time of the imaging;

wherein, calculate and get the horizontal displacement amount of every measurement festival second shooting relative to datum point (0, 0), when calculating: the displacement offset amount of the corresponding measurement section in the X direction with respect to the reference point at the time of the second shooting = the displacement offset amount of the measurement section in the X direction with respect to the previous time at the time of the second shooting + the displacement offset amount of the corresponding measurement section in the X direction with respect to the reference point at the time of the first shooting; the displacement offset amount of the corresponding measurement section in the Y direction with respect to the reference point at the time of the second shooting = the displacement offset amount of the measurement section in the Y direction with respect to the previous time at the time of the second shooting + the displacement offset amount of the corresponding measurement section in the Y direction with respect to the reference point at the time of the first shooting;

C. According to the mode of the step B, calculating and obtaining the horizontal displacement offset of each measuring section relative to the last shooting in each shooting from the third shooting to the ith shooting; and calculating the horizontal displacement of each measuring node relative to the reference point during each shooting.

In the scheme, the bottommost part of the foundation pit is preferably used as an initial point, and the method is suitable for areas, such as adults, where the bottom of the foundation pit is not easy to deform. Of course, the position of the top of the foundation pit can be selected as an initial point, and the latest coordinates of the top of the foundation pit can be measured in a traditional measuring mode.

In this scheme, the setting mode of camera under, LED lamp is last has been selected in a measurement section, of course also can select the mode of camera under, LED lamp according to actual need, and the latter is the LED lamp of the upper measurement section of upwards shooting through the camera of the measurement section of below.

For easy understanding, the core innovation points of the scheme are described:

1. providing a new idea of monitoring deep horizontal displacement in the inclinometer pipe;

the traditional thought of measuring the foundation pit in the inclinometer is that the inclinometer is arranged in the inclinometer, when the foundation pit is deformed, the inclinometer is inclined, and the horizontal displacement value is calculated by the operation of a trigonometric function sin or cos according to the angle difference value before and after the inclination;

The scheme is that the inclinometer is replaced by the measuring section, and after a plurality of measuring sections are connected through hoses to form a connecting string, the connecting string is arranged in the inclinometer pipe; shooting the LED lamp of the measuring section at the lower position through a camera in the measuring section at the upper position, so as to obtain an image with the LED lamp, and after a coordinate system is established, obtaining the coordinate value of the LED lamp in the image; when the coordinate values of the LEDs are changed in the next measurement, horizontal displacement is indicated, and the coordinate values of the LEDs are converted according to the coefficients to obtain horizontal displacement offset (the coefficients are calculated according to the coordinate values of a plurality of LED lamps); accumulating the horizontal displacement offset of each measuring point to obtain the accumulated deep horizontal displacement of each measuring section;

in addition, when the foundation pit is deformed, the measuring section bends and torque along with the inclinometer pipe, so that the plane of a sensor of a camera of the upper measuring section and the plane of an LED lamp of the lower measuring section are not in a horizontal plane, the planes of the two planes are not parallel, and the horizontal displacement cannot be well reflected during shooting; the measuring section is designed to be heavy and light, so that the sensor plane of the camera of the upper measuring section and the plane of the LED lamp of the lower measuring section are both positioned on the horizontal plane, and the shot image well reflects the horizontal displacement; when the coordinate values of the LED lamps are converted into actual horizontal displacement through coefficients, the coefficients are calculated through a plurality of LED coordinate values, and the accuracy of the converted actual horizontal displacement is ensured;

2. The method has high measurement precision and low cost;

a. the measurement accuracy is effectively improved;

in the traditional measurement mode of sliding in the inclinometer, for example, in the measurement, the whole inclinometer is placed in the inclinometer in the first day of measurement, and one group of data is measured every 0.5m lower; when the inclinometer falls to the bottom, the whole inclinometer is lifted, and the measurement is performed again, and the total measurement is required to be performed for 3 times; when measuring on the second day, referring to the measuring mode of the first day to measure; then the first measurement on the first day needs to drop 0.5m to measure 1 group of data, the height of the dropped 0.5m cannot be accurately ensured, and errors exist; the first day needs 3 times of measurement, and the point positions of the 3 times of measurement cannot be guaranteed to be at the same height, so that errors exist; in the 3 measurements on the second day, the point measured on the second day and the point measured on the first day cannot be guaranteed to have errors in the same position; in short, in the sliding measurement mode, even if a point with a certain height is to be measured, the same position of each measurement point cannot be ensured;

when the measurement is carried out in the scheme, a plurality of measurement sections are connected through the hose to form a connection string, the connection string is placed in the inclinometer pipe and kept motionless, the measurement is carried out at the same point each time, and only analysis and calculation are needed during the measurement, so that the accuracy of each measurement can be ensured; during measurement, the connecting string is arranged in the inclinometer pipe, the position of the connecting string in the inclinometer pipe is kept unchanged in the measurement process, and corresponding information marks are acquired through the data acquisition unit during each measurement; besides the simple and convenient operation, the position of the connecting string in the measuring tube is not changed, the obtained data is accurate, so that the horizontal displacement and depth H change curve drawn subsequently along with time is accurate enough, and the reference can be provided for the construction department well;

b. The cost is low;

the measurement mode that a plurality of probes are connected in series and then kept in an inclinometer pipe is used in the market, so that the same point can be ensured to be measured each time, and the precision is ideal; however, the core of the probe measurement is a chip inside the probe, and the cost of the chip is high, so that the total cost of a single probe is not very good; therefore, the total cost of the plurality of probes connected in series is extremely high;

in the scheme, the measuring section is used for replacing the probe, and the mode of chip processing is replaced by the image processing of the measuring section, so that the cost is saved.

The invention has the following advantages:

(1) The method provides a new idea for monitoring the horizontal displacement of the foundation pit in the inclinometer pipe; according to the scheme, the image is analyzed through the grids by shooting the image, so that accurate position coordinates are obtained; as long as the grid is small enough, enough position precision can be ensured, so that the obtained coordinates of the LED lamp are accurate; very accurate horizontal displacement detection can be obtained after calculation; when the horizontal position exceeds the threshold value, the foundation pit deformation is excessively large, and an alarm can be sent to remind relevant technicians;

(2) The method has the advantages of high measurement precision and low cost, and is beneficial to industrial actual measurement.

Drawings

FIG. 1 is a schematic diagram of the structure of the present invention;

FIG. 2 is a schematic view of the structure at the upper portion of FIG. 1;

FIG. 3 is a schematic view of a measuring section;

FIG. 4 is a graphical illustration of multiple foundation pit monitoring;

FIG. 5 is a graphical illustration of the amount of horizontal displacement at a measurement location at a pit 1 m;

FIG. 6 is a graphical illustration of the amount of horizontal displacement, accumulated deep level displacement, of a plurality of foundation pits at a height position;

in the figure: 1-inclinometer, 2-measuring section, 3-camera, 4-LED lamp, 5-gyro wheel, 6-hose, 7-metal casing, 8-torsional spring.

Detailed Description

The present invention will be further described with reference to the accompanying drawings, but the scope of the present invention is not limited to the following.

The invention provides a deep horizontal displacement monitoring system based on machine vision, which comprises an inclinometer pipe, a measuring section and a data acquisition unit; the pipe wall of the inclinometer pipe is provided with a chute along the axial direction, the inclinometer pipe is buried in a supporting structure, the supporting structure is positioned in the foundation pit, and the radial direction of the chute of the inclinometer pipe is perpendicular to the free surface of the foundation pit; the lower end of the measuring section is provided with a camera, the upper end of the measuring section is provided with an LED lamp, and both sides of the measuring section are provided with rollers; a plurality of measuring joints are connected through hoses to form a connecting string; the connecting string is placed in the inclinometer pipe, and the roller is matched with the chute; the data acquisition unit is used for controlling the work of each measuring node, receiving images and processing the images.

And calculating coordinates of LEDs in the image, and calculating horizontal displacement of each measuring section after calculation according to the coordinates of the LED lamp images of each measuring section at different times, and calculating the space displacement of each measuring section of the connecting string to obtain the deep displacement of the support structure.

The invention provides a monitoring method of a deep horizontal displacement monitoring system based on machine vision, wherein in the monitoring method, a measuring section positioned at an upper position shoots an LED lamp of a measuring section adjacent to a lower position through a camera, and a shot image is uploaded to a data acquisition unit; establishing a horizontal coordinate system through the obtained image, and obtaining the coordinates of the LEDs in the image; calculating the horizontal displacement of each measuring section according to the LED lamp image coordinates of each measuring section in front and back two different times; and calculating the horizontal displacement of each measuring section of the connecting string to obtain the deep displacement of the supporting structure.

When measuring/monitoring, whole connection cluster is arranged in the inclinometer pipe all the time, need not to pull from top to bottom, compares in the inclinometer and needs the measurement mode of constantly pulling in the inclinometer pipe, and this scheme measurement accuracy is higher. According to the scheme, when measuring/monitoring, the LED lamp below is shot through the camera, the coordinates of the LED lamp are obtained through image analysis, so that whether the foundation pit is deformed or not is judged, and compared with a measuring mode of a probe, the scheme is lower in cost. The scheme has the advantages of high precision and low cost.

The hose is on the one hand to link each measurement section, on the other hand in order to avoid debris to exist between camera, the LED lamp, guarantees to obtain the image of wasing. The hose is preferably a PVC plastic hose, i.e. has a certain hardness but is easier to deform.

Fig. 1 is a schematic structural view of a connection string suspended in an inclinometer pipe according to an embodiment of the present invention, and fig. 2 is a schematic structural view at an upper portion of fig. 1. In the embodiment of the invention, as shown in fig. 1, each measuring section 2 is connected through a hose 6 to form a connecting string, and the connecting string is hung in an inclinometer pipe 1 by a pull rope; the rollers on the two sides of the measuring section 2 slide with the sliding grooves of the inclinometer pipe 1. If the foundation pit deforms, the supporting bracket is changed, so that the inclinometer pipe is inclined; the LED lamp 4 in the lower measuring section 2 is shot by the camera 3 in the upper measuring section 2, and if the inclinometer tube is subjected to inclination change, the position of the LED lamp 4 in the shot image can be changed, so that the horizontal displacement variation of each height position in the foundation pit is monitored.

Fig. 3 is a schematic structural diagram of a measuring node according to an embodiment of the present invention. As shown in fig. 3, the measuring section 2 provided by the embodiment of the invention comprises a metal shell 7, a camera 3 and a plurality of LED lamps 4; rollers 5 are arranged on two sides of the metal shell 7; a camera 3 is arranged in the lower end of the metal shell 7, and the camera 3 is a digital camera; a plurality of LED lamps 4 are arranged in the upper end of the metal shell 7 along the radial section, and each LED lamp comprises an LED target lamp and two auxiliary LED lamps; the metal shell 7 is provided with a data transmission interface (drawn in the figure), the data transmission interface is electrically connected with the camera 3 and the LED lamp, and the data transmission interface is also connected with the data acquisition unit through a plug-in cable.

For example, the data acquisition unit is located within a data acquisition unit DTU that is placed at the wellhead of the foundation pit.

For example, the data acquisition unit is responsible for controlling the camera, receiving the image of the camera, and supplying power to the camera and the LED lamp; in addition, the image is also uploaded to the cloud platform. And the operation of the monitoring method is realized through a cloud platform. Such a data acquisition unit is relatively inexpensive.

Or, for example, the data acquisition unit is responsible for controlling the camera, receiving the image of the camera, and supplying power to the camera and the LED lamp; in addition, it is also responsible for analyzing the operation monitoring method. Such data acquisition units are relatively expensive.

For example, cameras have CCD/CMOS sensors that facilitate capturing illuminated LED lights. At the position of

For example, at both sides of the metal housing 7: one end of the torsion spring 8 is provided with a roller 5, and the other end of the torsion spring 8 is welded and fixed on the metal shell 7; the rollers 5 on the left and right sides are symmetrical to each other. And the sensor of the camera is located at the center of the line connecting the two rollers 5. Under the elastic action of the torsion spring 8, the measuring section 2 is adaptively positioned at the center of the inclinometer pipe 1. So that the camera 3 of the measuring section 2 positioned at the upper position photographs the LED lamps of the measuring section 2 positioned at the central position of the inclinometer 1 at the lower position.

For example, the sensor of the camera is rectangular, which is horizontally disposed; one side of the sensor of the camera is parallel to the line connecting the pulleys 5 on both sides. The installation of the position and the angle of the sensor of the camera influence the angle of the image shot by the camera; when one side of the sensor of the camera is parallel to the line connecting the pulleys 5 on both sides, the horizontal coordinate system established by using the image has a rectangular X/Y axis as the vertical side, i.e. one axis is parallel to the line connecting the two rollers, and the calculation step of reducing the angle is reduced in the subsequent calculation process.

For example, the LED target lamp and the two auxiliary LED lamps on the measuring section 2 are positioned on the same cross section but are not positioned on the same straight line. When calculating the horizontal displacement, the coordinates of the three LED lamps are needed, and the horizontal displacement has both displacement in the X-axis direction and displacement in the Y-axis direction, so that the three are not positioned on the same straight line for better response to the horizontal displacement. In addition, three LED lamps are not located on the same straight line, and measurement accuracy can be improved. Specifically, in the monitoring process, when the inclinometer pipe 1 is inclined, the sections where the three LED lamps are positioned are also inclined adaptively; assuming that three LED lamps 4 are designed on a straight line, if the inclinometer 1 is inclined along the vertical plane where the straight line is located, one LED lamp is located at the highest point and one LED lamp is located at the lowest point, while the camera 3 shoots the horizontal coordinates of the LED lamps, if the height difference between the two LED lamps is too large, the larger the horizontal position error generated after projection to the horizontal plane is. Specifically, the first design is that three LED lamps 4 are positioned on the radial straight line segment of a circle and are respectively positioned at the two end points and the center of the radial straight line segment; the design II is that three LED lamps 4 are respectively positioned at the edge of a circle, and the connecting lines of the three LED lamps 4 are equilateral triangles; then, when the flip locus of the circle is located in the vertical plane where the radial straight line segment is located, the maximum position error generated by designing an LED in the horizontal plane is larger than that of the design two.

For example, when a plurality of LED lamps 4 are mounted on the measuring section, the positional accuracy of each LED lamp 4 is better than 0.1mm. Since the horizontal displacement offset is analyzed and calculated, the intervals between the three LED lamps 4 are used, the positional accuracy of the LED lamps 4 themselves must be ensured to ensure the interval accuracy between the LED lamps 4.

For example, the CCD sensor of the camera is located at a position not exceeding 10mm from the position of the LED lamp 4 (not visible in the figure, which is only illustrative). Since the horizontal displacement amount is measured at the time of detection, the height value between the CCD sensor of the camera and the LED lamp 4 cannot be changed excessively by the inclination of the inclinometer pipe 1, and therefore the height difference between the CCD sensor and the LED lamp 4 is defined.

For example, the spacing between measuring knots 2 is 500mm or 1000mm. When a measurement is performed after common probes are connected in series, the interval between the probes is 500mm in order to reduce errors. The post-measurement precision of the scheme is higher, and the distance between the two measuring sections 2 can be 500mm or even 1000mm. In the connection string, the hose is connected with the measuring joint in a sealing way.

At least one embodiment of the present invention provides a monitoring method of a deep horizontal displacement monitoring system based on machine vision, including: s1, sliding a traditional inclinometer down into an inclinometer pipe along a sliding chute to perform initial measurement, and obtaining initial spatial position data of the inclinometer pipe; if the geometrical space shape of the inclinometer pipe is not considered during installation, the step is omitted;

Selecting one stable end of the inclinometer pipe as a starting position, namely selecting the bottom of the inclinometer pipe or selecting the pipe orifice at the top of the inclinometer pipe; in the specific embodiment, the bottom of the inclinometer tube is selected as a starting point, and the initial space coordinates of the LED target lamps of the corresponding measuring sections are measured by a traditional inclinometer; wherein the initial spatial coordinates of the first measurement node are (X ₁ ，Y ₁ ，H ₁ ），X ₁ 、Y ₁ LED target lamp plane coordinate representing 1 st measuring section, H ₁ The height of the LED target lamp is 1 st measuring section, and X is calculated because the first measuring section is the starting point ₁ =0，Y ₁ =0; wherein the initial spatial coordinates of the nth measurement node are (X _n ，Y _n ，H _n ），X _n 、Y _n Representing the plane coordinates of the LED target lamp of the nth measuring section, H _n The height of the LED target lamp is the nth measuring section;

Selecting one stable end of the inclinometer pipe as a starting position, selecting the bottom of the inclinometer pipe or selecting the pipe orifice at the top of the inclinometer pipe;

s3, shooting;

starting each measuring section camera to shoot for the first time, enabling the camera to shoot the adjacent LED target lamps below the measuring section cameras to obtain images of the LED target lamps, wherein the images are initial observation images or last observation images, and transmitting the shot images to the data acquisition unit;

the position of the inclinometer is kept unchanged, each measuring section camera is started to shoot according to a specified time interval, the adjacent LED target lamps below the measuring section cameras are shot through the cameras, the images of the LED target lamps are obtained, and the shot images are transmitted to the data acquisition unit;

After each time of uploading the latest observation image, analyzing and processing the change amounts of the current observation image and the last observation image of the LED target lamp through the data acquisition unit, calculating the current horizontal offset (namely calculating the horizontal displacement offset of the corresponding measuring section after the current measurement relative to the previous measurement), and calculating the current accumulated horizontal displacement (namely calculating the accumulated horizontal displacement of the corresponding measuring section after the current measurement according to the offset);

and S4, if the horizontal offset of a certain measuring section after a certain measurement or the accumulated horizontal displacement after a certain measurement exceeds a set threshold value, alarming.

In the monitoring method provided in the embodiment of the present disclosure, in step S3, after each uploading of the latest observation image, the data acquisition unit analyzes and processes the current observation image and the previous observation image of the LED target lamp, calculates the current horizontal offset, and calculates the current accumulated horizontal displacement, which means:

A. calculating coefficients;

a1, observing an image after the first shooting, and measuring coordinate values of each LED lamp in an image coordinate system; calculating the coefficient of the actual distance corresponding to the single pixel near the LED target lamp;

specifically, when the shot picture is enlarged, a certain pixel grid is defined as the origin of coordinates, in this embodiment, the first pixel grid at the upper left corner in the picture is taken as the origin of coordinates, and the horizontal line and the vertical line of the pixel are taken as X, Y coordinates, so that the image coordinates corresponding to the LED target lamp, the second auxiliary LED lamp and the third auxiliary LED lamp in each of the first measurement section to the nth measurement section can be obtained;

in the first measurement section, the coordinates of the images corresponding to the LED target lamp, the second auxiliary LED lamp and the third auxiliary LED lamp in the first shot image are (x) _11-1 ,y _11-1 ）、（x _12-1 ,y _12-1 ）、（x _13-1 ,y _13-1 ) The method comprises the steps of carrying out a first treatment on the surface of the In the second measurement section, the coordinates of the images corresponding to the LED target lamp, the second auxiliary LED lamp and the third auxiliary LED lamp in the first shot image are (x) _21-1 ,y2 _1-1 ）、（x _22-1 ,y _22-1 ）、（x _23-1 ,y2 _3-1 ) The method comprises the steps of carrying out a first treatment on the surface of the Obtaining image coordinates corresponding to the LED target lamp, the second auxiliary LED lamp and the third auxiliary LED lamp in the first shot image in the third measurement section to the n-1 measurement section one by one; in the nth measurement section, the coordinates of the images corresponding to the LED target lamp, the second auxiliary LED lamp and the third auxiliary LED lamp in the first shot image are (x) _n1-1 ,y _n1-1 ）、（x _n2-1 ,y _n2-1 ）、（x _n3-1 ,y _n3-1 ) The method comprises the steps of carrying out a first treatment on the surface of the In the subscript of the image coordinates, the first number indicates the corresponding measurement section, and the second number indicates the LED target lamp if it is 1, if it is2 represents a second auxiliary lamp, if 3 represents a third auxiliary lamp, and the third number represents the shooting times;

specifically, according to the image coordinates of each measuring section in the first shooting, calculating to obtain the coefficient of the actual distance corresponding to a single pixel near the LED target lamp of each measuring section; the coefficient calculation mode is (d 1/horizontal distance from Led target lamp to second auxiliary Led lamp in the image obtained by the corresponding measuring section in the shooting time +d2/horizontal distance from Led target lamp to third auxiliary Led lamp in the image obtained by the corresponding measuring section in the shooting time)/2;

wherein the coefficient of the actual distance corresponding to a single pixel near the corresponding LED target lamp at the first shooting of the first measuring section

The method comprises the steps of carrying out a first treatment on the surface of the Obtaining coefficients of actual distances corresponding to single pixels near corresponding LED target lamps in the second measurement section to the n-1 measurement section one by one in the first shooting; coefficient of single pixel near LED target lamp corresponding to actual distance in nth measurement section

The method comprises the steps of carrying out a first treatment on the surface of the In the subscript of the corresponding coefficient K, the first number represents the measurement section and the second number represents the number of times of shooting;

In the step A1, the calculation of the coefficients of the actual distances corresponding to the number of pixels in the vicinity of the LED target lamp in the corresponding measurement section is not practical, and the coefficients are not used in the subsequent measurement; these coefficients are written out because the system automatically calculates these coefficients when running through the program, but the corresponding coefficients of step A1 are waste coefficients (rest not used);

specifically, through the second shooting, when the shot picture is amplified, the first pixel grid at the upper left corner in the picture is still used as the origin of coordinates, and the image coordinates corresponding to the LED target lamp, the second auxiliary LED lamp and the third auxiliary LED lamp corresponding to each measuring section from the first measuring section to the nth measuring section during the second shooting are obtained;

wherein, the coordinates of the images corresponding to the LED target lamp, the second auxiliary LED lamp and the third auxiliary LED lamp in the second shot image of the first measuring section are (x) _11-2 ,y _11-2 ）、（x _12-2 ,y _12-2 ）、（x _13-2 ,y _13-2 ) The method comprises the steps of carrying out a first treatment on the surface of the The coefficient of the actual distance corresponding to the single pixel near the LED target lamp during the second shooting of the first measuring section

Obtaining image coordinates corresponding to the LED target lamp, the second auxiliary LED lamp and the third auxiliary LED lamp in the second shot image of the second measuring section to the n-1 measuring section one by one; calculating and obtaining the coefficient of the actual distance corresponding to a single pixel near the LED target lamp in the second shooting of the corresponding measuring section;

in the nth measurement section, the coordinates of the images corresponding to the 3 LED lamps in the second shot image are (x) _n1-2 ,y _n1-2 ）、（x _n2-2 ,y _n2-2 ）、（x _n3-2 ,y _n3-2 ) The method comprises the steps of carrying out a first treatment on the surface of the The coefficient of the actual distance corresponding to a single pixel near the LED target lamp at the time of the 2 nd shooting of the nth measuring section

；

A3, measuring coordinate values of the centers of the LED lamps in an image coordinate system of the LED lamps according to an i-th observed image (i=3, 4.);

in the first measurement section, the coordinates of the image corresponding to the LED target lamp, the second auxiliary LED lamp and the third auxiliary LED lamp in the ith shot image are (x) _11-i ,y _11-i ）、（x _12-i ,y _12-i ）、（x _13-1 ,y _13-1 ) The method comprises the steps of carrying out a first treatment on the surface of the First testCoefficient of single pixel near LED target lamp corresponding to actual distance in ith shooting of measuring section

Obtaining image coordinates corresponding to the LED target lamp, the second auxiliary LED lamp and the third auxiliary LED lamp in the ith shot image in the second measurement section to the n-1 measurement section; calculating and obtaining a coefficient of the actual distance corresponding to a single pixel near the LED target lamp in ith shooting of the corresponding measuring section;

In the nth measurement section, the coordinates of the image corresponding to the 3 LED lamps in the ith shot image are (x) _n1-i ,y _n1-i ）、（x _n2-i ,y _n2-i ）、（x _n3-i ,y _n3-i ) The method comprises the steps of carrying out a first treatment on the surface of the The coefficient of the actual distance corresponding to a single pixel near the LED target lamp at the ith shooting of the nth measuring section

；

B. Calculating the horizontal offset of the current time relative to the first time after the second observation, and the horizontal displacement of the current time relative to the datum point (0, 0);

after the second observation is finished, calculating the current deep level displacement measurement of each measuring section according to the image coordinate change of the LED target lamp;

specifically, in the first measurement section, the image coordinates of the LED target lamp observed for the second time are (x _11-2 ,y _11-2 ) The image coordinates of the LED target lamp of the first observation (primary observation) are (x) _11-1 ,y _11-1 ) Conversion coefficient of K _1-2 The offset in the X direction from the previous time is (X) _11-2 -x _11-1 )*K _1-2 The amount of shift in the Y direction from the previous time is (Y _11-2 -y _11-1 )*K _1-2 The displacement amount in the X direction relative to the reference point is Deltax _1-2 =(x _11-2 -x _11-1 )*K _1-2 +0, the displacement in the Y direction with respect to the reference point is Δy _1-2 =(y _11-2 -y _11-1 )*K _1-2 ；

In the second measurement section, the image coordinates of the LED target lamp observed for the second time are (x _21-2 ,y _21-2 ) The image coordinates of the LED target lamp of the first observation (primary observation) are (x) _21-1 ,y _21-1 ) Conversion coefficient of K _1-2 The offset in the X direction from the previous time is (X) _21-2 -x _21-1 )*K _2-2 The amount of shift in the Y direction from the previous time is (Y _21-2 -y _21-1 )*K _2-2 The displacement amount in the X direction relative to the reference point is Deltax _2-2 =(x _21-2 -x _21-1 )*K _2-2 +Δx _1-2 The displacement amount in the Y direction relative to the reference point is deltay _2-2 =(y _21-2 -y _21-1 )*K _2-2 +Δy _1-2 ；

According to the mode, corresponding data after the second observation are calculated in the third measurement section to the nth measurement section;

in the nth measurement section, the image coordinates of the LED target lamp observed for the second time are (x _n1-2 ,y _n1-2 ) The image coordinates of the LED target lamp of the first observation (primary observation) are (x) _n1-1 ,y _n1-1 ) Conversion coefficient of K _1-n The offset in the X direction from the previous time is (X) _n1-2 -x _n1-1 )*K _1-n The amount of shift in the Y direction from the previous time is (Y _n1-2 -y _n1-1 )*K _1-n The displacement amount in the X direction relative to the reference point is Deltax _n-2 =(x _n1-2 -x _n1-1 )*K _1-n +Δx _（n-1）-2 The displacement amount in the Y direction relative to the reference point is deltay _n-2 =(y _n1-2 -y _n1-1 )*K _1-n +Δy _（n-1）-2 ；

C. Calculating the horizontal offset relative to the previous time after the ith observation, and the horizontal displacement relative to the datum point after the ith observation;

after the i-th observation is finished, calculating the current deep horizontal displacement of each measuring section according to the image coordinate change of the LED target lamp;

specifically, in the first measurement section, the image coordinates of the i-th observed LED target lamp are (x _11-i ,y _11-i ) The image coordinates of the i-1 th observation (last observation) of the LED target lamp are (x) _11-i-1 ,y _11-i-1 ) Conversion coefficient of K _1-i The offset in the X direction from the previous time is (X _11-i -x _11-i-1 )*K _1-i The amount of shift in the Y direction from the previous time is (Y _11-i -y _11-i-1 )*K _1-i The displacement amount in the X direction relative to the reference point is Deltax _1-i =(x _11-i -x _11-i-1 )*K _1-i +Δx _1-（i-1） The displacement amount in the Y direction relative to the reference point is deltay _1-i =(y _11-i -y _11-i-1 )*K _1-i +Δy _1-（i-1）；

In the second measurement section, the image coordinates of the i-th observed LED target lamp are (x _21-i ,y _21-i ) The image coordinates of the i-1 th observation (last observation) of the LED target lamp are (x) _21-i-1 ,y _21-i-1 ) Conversion coefficient of K _2-i The offset in the X direction from the previous time is (X) _21-i -x _21-i-1 )*K _2-i The amount of shift in the Y direction from the previous time is (Y _21-i -y _21-i-1 )*K _2-i The displacement amount in the X direction relative to the reference point is Deltax _2-i =(x _21-i -x _21-i-1 )*K _2-i +Δx _2-（i-1） The displacement amount in the Y direction relative to the reference point is deltay _2-i =(y _21-i -y _21-i-1 )*K _2-i +Δy _2-（i-1）；

According to the mode, calculating the data observed for the ith time in the third measurement section to the nth measurement section;

in the nth measurement section, the image coordinates of the i-th observed LED target lamp are (x _n1-i ,y _n1-i ) The image coordinates of the i-1 th observation (last observation) of the LED target lamp are (x) _n1-i-1 ,y _n1-i-1 ) Conversion coefficient of K _n-i The offset in the X direction from the previous time is (X) _n1-i -x _n1-i-1 )*K _n-i The amount of shift in the Y direction from the previous time is (Y _n1-i -y _n1-i-1 )*K _n-i The displacement amount in the X direction relative to the reference point is Deltax _n-i =(x _n1-i -x _n1-i-1 )*K _n-i +Δx _n-（i-1） The displacement amount in the Y direction relative to the reference point is deltay _n-i =(y _n1-i -y _n1-i-1 )*K _n-i +Δy _n-（i-i）；

D. Each measurement section draws a horizontal displacement graph (drawing means is identical to the conventional drawing means) based on the horizontal displacement amount measured each time and the accumulated horizontal displacement amount.

It is desirable to use the point at the top left corner as the origin of coordinates according to the grid of pixels when the image is set up in a coordinate system. During the measurement, there are three forms: the upper measuring section is motionless, the lower measuring section is motionless (i.e. camera is motionless while the LED lamp is motionless), the upper measuring section is motionless (i.e. camera is motionless while the LED lamp is motionless), and the upper measuring section and the lower measuring section synchronously move or are all motionless; in either way, a measure of belief can be achieved.

In practical application, the embodiment is as shown in fig. 4 to 6: in fig. 4, a plurality of inclinations are provided on the wall circumference side in a rectangular foundation pit, each inclinations having an inclinations pipe; FIG. 5 is a graph showing the measured joint horizontal displacement of a certain inclinometer at a height of 1 m; FIG. 6 shows the horizontal displacement and horizontal accumulation displacement of inclinations at a certain height of the inclinations at the inclinations positions Z15, Z14, Z13, Z12-1, Z12, Z17, Z18, Z19 and Z19-1, and the inclinations with light gray color indicate that severe displacement occurs and alarm is needed.

The foregoing examples represent only preferred embodiments, which are described in more detail and are not to be construed as limiting the scope of the invention. It should be noted that variations and modifications can be made by those skilled in the art without departing from the spirit of the invention, which falls within the scope of the invention.

Claims

1. A deep horizontal displacement monitoring system based on machine vision is characterized in that: comprising the following steps:

the pipe wall of the inclinometer pipe is provided with a chute along the axial direction; burying an inclinometer pipe in a supporting structure, wherein the supporting structure is a component part of a foundation pit, and the radial direction of a chute of the inclinometer pipe is perpendicular to the empty face of the foundation pit;

the lower end of the measuring section is provided with a camera, the upper end of the measuring section is provided with an LED lamp, and two sides of the measuring section are provided with symmetrical rollers; the camera is a digital camera;

the data acquisition unit is used for collecting data;

when assembled, the components are as follows: a plurality of measuring joints are connected through hoses to form a connecting string; the connecting string is placed in the inclinometer pipe, and the roller is matched with the chute; the data acquisition unit is connected with each measuring power saving, and controls each measuring section to work and collects image data of each measuring section;

the measurement is as follows: the measuring section positioned at the upper position shoots the LED lamps of the measuring section adjacent to the lower position through the camera, the shot image is uploaded to the data acquisition unit, the image coordinates of the LED lamps are obtained through the image, and then the horizontal displacement of each measuring section is obtained through calculation; and then the accumulated horizontal displacement is obtained through calculation, and the accumulated horizontal displacement can reflect the horizontal variation of the supporting structure.

2. The machine vision based deep level shift monitoring system of claim 1, wherein: in the measuring section, the LED lamp comprises an LED target lamp and two auxiliary LED lamps, wherein the LED lamps are positioned in the same plane and are not positioned on the same straight line.

3. The machine vision based deep level shift monitoring system of claim 1, wherein: in the measuring section, the camera is provided with a CCD/CMOS sensor;

torsion springs are fixed on two sides of the measuring section, and idler wheels are arranged at the outer ends of the torsion springs;

the CCD/CMOS sensor is positioned at the center of the connecting line of the rollers at the two sides.

4. A machine vision based deep level shift monitoring system as claimed in claim 3, wherein: in the measuring section, the CCD/CMOS sensor of the camera is rectangular, and one side of the rectangle is parallel to the connecting line of the pulleys at two sides.

5. A machine vision based deep level shift monitoring system as claimed in claim 3, wherein: in the measuring section, the distance between the CCD/CMOS sensor of the camera and the plane where the LED lamp is located is smaller than 10mm.

6. The machine vision based deep level shift monitoring system of claim 1, wherein: the data acquisition unit is positioned in the data acquisition unit DTU, and the data acquisition unit DTU is arranged at the wellhead position of the foundation pit;

The data acquisition unit is responsible for controlling the work of each unit section, receiving the image data of each measurement section and analyzing and calculating the received data; or the data acquisition unit is responsible for controlling each measuring section to work, receiving images, processing the images and uploading the images to the cloud platform, and the cloud platform analyzes, calculates and processes the data.

7. The machine vision based deep level shift monitoring system of claim 1, wherein: the measuring section comprises a metal shell, a camera and a plurality of LED lamps;

a camera and a plurality of LED lamps are respectively arranged in the lower end and the lower end of the metal shell; the LED lamps are positioned in the radial section of the metal shell, and the positions of the LED lamps are positioned in the radial section; the distance between the LED lamps needs to be accurately measured;

the metal shell is provided with a data transmission interface which is electrically connected with the camera; the data acquisition unit is in butt joint with the data transmission interface through a cable;

the left side and the right side of the metal shell are respectively provided with a roller wheel through torsion springs.

8. The machine vision based deep level shift monitoring system of claim 1, wherein: in the connecting string, the interval between the measuring joints is 500mm or 1000mm;

In the connection string, the hose is connected with the measuring joint in a sealing way.

9. The measurement method for the machine vision-based deep horizontal displacement monitoring system of any one of claims 1 to 8, characterized by: the method comprises the following steps: s1, sliding a traditional inclinometer down into an inclinometer pipe along a sliding chute to perform initial measurement, and obtaining initial spatial position data of the inclinometer pipe; if the geometrical space shape of the inclinometer pipe is not considered during installation, the step is omitted;

selecting one stable end of the inclinometer pipe as a starting position, namely selecting the bottom of the inclinometer pipe or selecting the pipe orifice at the top of the inclinometer pipe;

selecting one stable end of the inclinometer as a starting point, and respectively arranging a first measuring section, a second measuring section and an nth measuring section from one end of the starting point of the measuring tube to the other end of the starting point of the measuring tube; measuring initial space coordinates of the LED target lamps of the corresponding measuring sections by a traditional inclinometer; wherein the initial spatial coordinates of the first measurement node are (X ₁ ，Y ₁ ，H ₁ ），X ₁ 、Y ₁ LED target lamp plane coordinate representing 1 st measuring section, H ₁ The height of the LED target lamp is 1 st measuring section, and X is calculated because the first measuring section is the starting point ₁ =0，Y ₁ =0; wherein the initial spatial coordinates of the nth measurement node are (X _n ，Y _n ，H _n ），X _n 、Y _n Representing the plane coordinates of the LED target lamp of the nth measuring section, H _n The height of the LED target lamp is the nth measuring section;

s3, shooting;

after a period of time, starting each measuring section camera to shoot for the second time, shooting an LED target lamp of the last measuring section by the camera to obtain an image of the LED target lamp, wherein the image is an observation image of the time, and transmitting the shot image to a data acquisition unit;

10. The machine vision based deep level shift monitoring system measurement method of claim 9, wherein: in step S3, after the latest observation image is uploaded each time, the data acquisition unit analyzes and processes the current observation image and the last observation image change amount, calculates the current horizontal offset, and calculates the current accumulated horizontal displacement:

A. Calculating coefficients;

when observing the image, obtaining image coordinates corresponding to the corresponding LED target lamp, the second auxiliary LED lamp and the third auxiliary LED lamp in the first measurement section to the nth measurement section in the second shooting; wherein, the image coordinates corresponding to the LED target lamp, the second auxiliary LED lamp and the third auxiliary LED lamp in the nth measurement section are (x) _n1-2 ,y _n1-2 ）、（x _n2-2 ,y _n2-2 ）、（x _n3-2 ,y _n3-2 ) In the subscript of the image coordinates, the first number represents the corresponding measurement section, the second number represents the LED target lamp if it is 1, the second auxiliary lamp if it is 2, the third auxiliary lamp if it is 3, and the third number represents the number of times of shooting;