Visual detection method for shield segment assembly quality and segment assembly system

Abstract

The application discloses a kind of visual detection method and segment system for shield segment for segment quality, method includes: the image of the surface of shield segment including the image of the target to be measured is collected The row projection curve and column projection curve of image are calculated;Respectively, the row projection curve and column projection curve are pretreated, and the minimum point of each is determined;According to the minimum point of row projection curve and the minimum point of column projection curve, the region of interest is determined in image;The pixel value of external region outside the region of interest is differentiated and set, so that external region and region of interest produce boundary line;Image is processed by algorithm to determine the contour in the region of interest, and fit contour line to it;The center of the region surrounded by the contour line fitted to is determined as the location of the target to be measured in image.The visual detection method based on gray projection curve limits the target to be measured in the region of interest, improves detection accuracy and recognition efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of visual detection, and in particular to a visual detection method for shield segment assembly quality and a segment assembly system. BACKGROUND

[0002] The segment erector is an important subsystem of the shield system for tunnel lining segment assembly, and its working performance is directly related to construction safety and tunnel quality. At present, the segment assembly process relies on manual control of the segment erector through a remote controller. With the increasing demand for shield construction and the increasing diameter of tunnels, this method is increasingly unable to meet the construction needs in terms of construction safety, segment forming quality, and construction efficiency.

[0003] Chinese patent CN 108104834B discloses a shield segment automatic assembly method and system, which proposes a method for completing fine adjustment assembly of the current segment to be installed according to image information, identification position information, and assembly strategy. Chinese patent CN 109736836B discloses a shield segment automatic assembly method based on Tag code visual measurement, which provides a shield segment automatic assembly method based on Tag code visual measurement. The computer calculates the segment offset according to the collected Tag code pose image, and controls the shield machine to complete the initial adjustment positioning and fine adjustment positioning of the segment according to the position offset, thereby completing the automatic assembly of the segment. Chinese patent application CN 113107550A discloses a shield segment automatic assembly method based on machine vision. This method predicts the expected position point of the segment to be assembled by detecting the target ArUco code feature corner on the segment, and obtains the expected speed of each mechanism of the segment erector using an image-based visual servo strategy. Finally, the joint speed closed-loop control is used to obtain the control amount of each actuator, thereby realizing the automatic assembly of the segment. Chinese patent application CN 113863966A discloses a segment gripping pose detection device and method based on deep learning vision, which realizes the gripping pose detection of the segment to be assembled.

[0004] The above methods all use visual measurement methods for pose detection, although the to-be-detected targets are different, including Tag codes, triangular marks, ArUco codes and grooves, but all need to find the feature points of the to-be-detected target in the camera plane coordinates, and the basic method to determine the coordinates is to find the contour of the to-be-detected target, fit the contour, and take the position of the feature point of the fitted shape in the image as its coordinates in the camera plane. The technical solutions provided in the above patents / patent applications realize detection by two methods, traditional image method and deep learning method. The traditional image method has a contradiction between noise resistance and detection accuracy, and cannot be applied to complex shield construction environments (low environmental illumination, easy pollution of the surface of the segment, uneven illumination, etc.). The deep learning method needs a large amount of data and has very slow detection speed, and is difficult to be applied to real-time control equipment, and its noise resistance also cannot meet the shield environment.

[0005] The disclosure of the above background art content is only used to assist in understanding the inventive concept and technical solutions of the present application, and does not necessarily belong to the prior art of the present patent application, nor does it necessarily give technical teaching; in the absence of explicit evidence that the above content has been disclosed before the filing date of the present patent application, the above background art should not be used to evaluate the novelty and inventiveness of the present application. SUMMARY

[0006] The purpose of the present application is to provide a visual detection method for shield segment assembly quality with high recognition accuracy and high detection efficiency, and to realize automatic assembly of shield segments by using the detection method to improve the assembly quality of shield segments.

[0007] To achieve the above purpose, the technical scheme adopted by the present application is as follows:

[0008] A visual detection method for shield segment assembly quality, comprising:

[0009] Collecting an image containing a to-be-detected target on the surface of a shield segment, the size of the image being M*N;

[0010] The row projection curve and the column projection curve of the image are calculated respectively by using the following formulas:

[0011]

[0012] Wherein, C(i,j) is the gray value of the (i,j) coordinate point in the image, P row (i) is the gray projection value of the i-th row in the image, P col (j) is the gray projection value of the j-th column in the image;

[0013] The row projection curve and the column projection curve are preprocessed respectively, and the minimum value points are determined respectively;

[0014] determining a region of interest in the image according to the minimum value points of the row projection curve and the minimum value points of the column projection curve;

[0015] distinguishing pixel values of an external region outside the region of interest, and setting a boundary between the external region and the region of interest;

[0016] algorithmically processing the image to determine a contour in the region of interest and fitting a contour line to the contour;

[0017] determining a center of a region enclosed by the fitted contour line as a position of the target object in the image.

[0018] Further, any of the technical solutions or combinations of the technical solutions described above, the region of interest in the image is determined by the following method:

[0019] determining a coordinate point corresponding to the minimum value points as a center point of interest according to the values of the row and column corresponding to the minimum value points;

[0020] expanding outwardly to obtain the region of interest with the center point of interest or pixel points of n adjacent pixels thereof as an expansion center point, wherein n is an integer less than 2%*min(M, N).

[0021] Further, any of the technical solutions or combinations of the technical solutions described above, the visual detection method for the shield segment assembly quality further comprises measuring a size of a region occupied by the target object in the image in advance, comprising: image acquisition of the shield segment at different shooting angles to obtain multiple test images; determining an external rectangle of the target object in each test image by manual or image processing to obtain a maximum value of a long side length of each external rectangle;

[0022] expanding the region of interest according to the maximum value of the long side length of the external rectangle, so that the region of interest can cover a region of a square with the maximum value of the long side length as a side length value.

[0023] Further, any of the technical solutions or combinations of the technical solutions described above, the number of rows or columns occupied by the longest long side of each external rectangle is T, and the region of interest is obtained by expanding T columns to the left and right directions and T rows to the upward and downward directions from the expansion center point; or,

[0024] taking the expansion center point as a center and T rows of corresponding length as a radius to make a circle to obtain the region of interest.

[0025] Further, any one of the above technical solutions or a combination of the above technical solutions, the number of rows or columns occupied by the longest side in each circumscribed rectangle is T, and T / 2 columns are extended to the left and right directions from the expansion center point, and T / 2 rows are extended to the up and down directions, to obtain the region of interest.

[0026] If it is determined that the contour shape in the region of interest does not match the shape of each type of target to be detected stored in the database, secondary expansion is performed, and the amplitude of the secondary expansion is greater than the amplitude of the previous expansion.

[0027] Further, any one of the above technical solutions or a combination of the above technical solutions, the pre-processing of the row projection curve and the column projection curve and the determination of the respective minimum points include the following steps:

[0028] The row projection curve and the column projection curve are respectively smoothed by using a moving average method, and the first derivatives of the smoothed curves are respectively calculated to determine the respective minimum points of the two curves.

[0029] Further, any one of the above technical solutions or a combination of the above technical solutions, a Gaussian filter is used to filter the image, and a canny edge detection algorithm is used to determine the contour.

[0030] The contour fitting contour line is as follows: if the target to be detected is a non-special-shaped part, the contour is fitted with a minimum circumscribed rectangle, and the center of the circumscribed rectangle is taken as the coordinate value of the target to be detected in the image in the current state; if the target to be detected is a special-shaped part, the contour boundary is taken as a contour line, and the barycenter point of the internal region of the contour line is calculated as the coordinate value of the target to be detected in the image in the current state.

[0031] Further, any one of the above technical solutions or a combination of the above technical solutions, the operation of differentiating the pixel values of the external region outside the region of interest is to set the pixel values of the external region outside the region of interest to 255.

[0032] According to another aspect of the present application, the present application provides a shield segment automatic assembling method, comprising the following steps:

[0033] The end position of the target to be detected in the image in the assembled state is calibrated in advance;

[0034] The current position of the target to be detected in the image is determined by using the visual detection method described above;

[0035] The moving direction and distance of the target to be detected from the current position to the end position are determined;

[0036] The assembling machine loads the to-be-tested target at the starting position and moves to the ending position.

[0037] According to still another aspect of the present application, the present application provides a shield segment automatic assembling system, comprising an assembling machine, a camera and a processor, the camera is configured to collect an image of a shield segment surface containing a to-be-tested target, the imaging size of the camera is M*N; the assembling machine is configured to assemble the to-be-tested target according to the current position of the to-be-tested target in the image and the ending position of the to-be-tested target in the image in a pre-calibrated completed assembling state, wherein the current position of the to-be-tested target in the image is obtained by the processor through the following steps:

[0038] The row projection curve and the column projection curve of the image collected by the camera are calculated respectively by using the following formulas:

[0039]

[0040] Wherein, C(i,j) is the gray value of the (i,j) coordinate point in the image, P row (i) is the gray projection value of the i-th row in the image, P col (j) is the gray projection value of the j-th column in the image;

[0041] The row projection curve and the column projection curve are preprocessed respectively, and the minimum value points are determined respectively;

[0042] According to the minimum value points of the row projection curve and the minimum value points of the column projection curve, a region of interest is determined in the image;

[0043] The pixel values of an external region outside the region of interest are set to be different, so that a boundary line is generated between the external region and the region of interest;

[0044] The image is processed by an algorithm to determine the contour in the region of interest, and the contour line is fitted;

[0045] The center of the region surrounded by the fitted contour line is determined as the position of the to-be-tested target in the image.

[0046] The technical scheme provided by the present application has the following beneficial effects: the visual detection method based on gray projection curve, that is, the to-be-tested target is limited in a small region of interest in the image through the projection curve information of the collected image, and the coordinates of the feature points of the to-be-tested target in the camera plane are extracted, so as to effectively solve the problems of low detection precision and low recognition rate caused by low environmental illumination, easy pollution of the segment surface and uneven illumination, and finally improve the assembling quality of the shield segment. BRIEF DESCRIPTION OF DRAWINGS

[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description only represent some of the embodiments described in the present application, and for those skilled in the art, other drawings can be obtained without creative labor on the basis of these drawings.

[0048] Figure 1 The flowchart of the visual detection method for identifying the target position provided for an exemplary embodiment of the present application is shown in the figure.

[0049] Figure 2 The interface diagram of the minimum value point of the row projection curve provided for an exemplary embodiment of the present application is shown in the figure.

[0050] Figure 3 The interface diagram of the minimum value point of the column projection curve provided for an exemplary embodiment of the present application is shown in the figure.

[0051] Figure 4 The schematic diagram of the extended region of interest in the image provided for an exemplary embodiment of the present application is shown in the figure.

[0052] Figure 5 The working state schematic diagram of the shield segment erector provided for an exemplary embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0053] In order to make the person skilled in the art better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments only represent some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all the other embodiments obtained by those skilled in the art without creative labor should belong to the scope of protection of the present application.

[0054] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, device, product or equipment including a series of steps or units does not necessarily limit to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or equipment.

[0055] In one embodiment of the present application, a visual detection method for shield segment assembly quality is provided, as shown in Figure 1 The visual detection method comprises the following steps:

[0056] An image of a shield segment surface containing a target to be detected is collected, and the size of the image is M*N, wherein M is the number of pixel columns of the image, and N is the number of pixel rows of the image.

[0057] The row projection curve and the column projection curve of the image are calculated respectively by using the following formulas:

[0058]

[0059] Wherein C(i,j) is the gray value of the (i,j) coordinate point in the image, P row (i) is the gray projection value of the i-th row in the image, and P col (j) is the gray projection value of the j-th column in the image.

[0060] The row projection curve and the column projection curve are preprocessed respectively, and the minimum value points are determined, specifically, the sliding average method is used to smooth the row projection curve and the column projection curve respectively, and the first derivative of the smoothed curve is calculated to determine the minimum value points of the two curves, as shown in Figure 2 and Figure 3 .

[0061] According to the minimum value points of the row projection curve and the minimum value points of the column projection curve, a region of interest is determined in the image, and the specific manner is described in detail below.

[0062] The pixel values of the external region outside the region of interest are set to be different from the pixel values of the region of interest, so that a boundary line is formed between the external region and the region of interest. Specifically, the pixel values of the external region outside the region of interest can be set to 255.

[0063] The image is processed by an algorithm. Specifically, the image can be filtered by a Gaussian filter, and then the canny edge detection algorithm is used to determine the contour in the region of interest and fit the contour line. Specifically, if the target to be detected is a non-special-shaped part, the contour is fitted by a minimum circumscribed rectangle, and the center of the circumscribed rectangle is taken as the coordinate value of the target to be detected in the image in the current state. If the target to be detected is a special-shaped part, the contour boundary is taken as the contour line, and the center of gravity of the internal region of the contour line is calculated as the coordinate value of the target to be detected in the image in the current state, so as to avoid the center of the circumscribed rectangle of the special-shaped part deviating too much from the special-shaped part.

[0064] determining the center of the region enclosed by the contour line as the position of the target in the image.

[0065] In one embodiment, the region of interest is determined in the image by the following way:

[0066] According to the values of the row and column corresponding to the minimum point, the corresponding coordinate point is determined as the center point of interest;

[0067] The pixel point of the center point of interest or the pixel point adjacent to it is taken as the expansion center point, and the region of interest is obtained by outward expansion, wherein n is an integer less than 2%*min(M,N), and n is preferably 0. The expansion mode can be as shown in the following figure. Figure 4 The columns are expanded in the left and right directions, and the rows are expanded in the up and down directions.

[0068] The expanded region of interest needs to cover the target, so the size of the region occupied by the target in the image is also measured in advance, including: image acquisition of the shield segment at different shooting angles to obtain multiple test images; determining the circumscribed rectangle of the target in each test image by manual or image processing to obtain the maximum value of the long side length of each circumscribed rectangle.

[0069] According to the maximum value of the long side length of the circumscribed rectangle, the region of interest is expanded to make the region of interest cover the area of a square with the maximum value of the long side length as the side length value, so as to avoid that the region of interest is too small to completely accommodate the target object.

[0070] Specifically, there can be at least three ways:

[0071] Method one: taking the number of rows or columns occupied by the longest long side of each circumscribed rectangle as T, and expanding T columns to the left and right directions respectively and T rows to the up and down directions respectively from the expansion center point to obtain the region of interest, as shown in the following figure. Figure 4 Method two: taking the expansion center point as the center and T rows as the radius to make a circle to obtain the region of interest.

[0072] Method two: taking the expansion center point as the center and T rows as the radius to make a circle to obtain the region of interest.

[0073] The third mode is to record the number of rows or columns occupied by the longest side of each circumscribed rectangle as T, and to expand T / 2 columns to the left and right directions and T / 2 rows to the up and down directions respectively from the expansion center point to obtain the region of interest; if T is odd, the number of expanded rows and columns can be the integer plus 1, for example, if T is 21, 11 columns are expanded to the left and right directions and 11 rows are expanded to the up and down directions. If the contour shape in the region of interest is determined to be not matched with the shapes of various types of to-be-detected targets stored in the database, secondary expansion is performed, and the amplitude of the secondary expansion is greater than that of the previous expansion. For example, the number of columns expanded to the left and right directions after the secondary expansion is [3T / 4] ([] is the integer symbol), and the number of rows expanded to the up and down directions after the secondary expansion is [3T / 4] ([] is the integer symbol), and if a matched contour shape is still not obtained, T columns / rows are further expanded in each of the up, down, left and right directions.

[0074] The third mode has the advantage of high detection efficiency compared with the first and second modes, and the mechanism is that the image is projected in gray scale and the row / column gray scale projection minimum value is actually determined to be the feature point of the to-be-detected target, and the feature point of the to-be-detected target under the application of shield segment assembly is usually not an edge point, but close to the center or located at the center. In theory, 0.5T columns / rows are expanded to the left, right, up and down directions to cover the to-be-detected target. Of course, the present embodiment does not exclude special cases, for example, the to-be-detected object is a special-shaped part, and the side length of the circumscribed rectangle in the current orientation exceeds the longest side of the circumscribed rectangle of the to-be-detected target in the test image, so that the region of interest after the first expansion cannot completely accommodate the to-be-detected target, and therefore the present embodiment proposes the above-mentioned secondary expansion / multiple expansion solution.

[0075] In an embodiment of the present application, a shield segment automatic assembly method is provided, comprising the following steps:

[0076] The end position of the to-be-detected target in the image in the assembled state is calibrated in advance;

[0077] The current position of the to-be-detected target in the image is determined by using the visual detection method of the above-mentioned embodiment;

[0078] The moving direction and distance of the to-be-detected target from the current position to the end position are determined;

[0079] The to-be-detected target at the starting position is loaded by the assembly machine, and the assembly machine moves to the end position.

[0080] In an embodiment of the present application, a shield segment automatic assembly system is provided, as shown in Figure 5As shown, the assembling system comprises an assembling machine 1, a camera 2 and a processor, the camera is configured to collect an image of a surface of a shield segment 3 containing a target to be measured, and the imaging size of the camera is M*N; the assembling machine 1 is configured to assemble the target to be measured according to a current position of the target to be measured in the image and an end position of the target to be measured in the image in a pre-calibrated completed assembling state, wherein the current position of the target to be measured in the image is obtained by the processor through the following steps:

[0081] The row projection curve and the column projection curve of the image collected by the camera are calculated by using the following formulas respectively:

[0082]

[0083] Wherein, C(i,j) is the gray value of the (i,j) coordinate point in the image, P row (i) is the gray projection value of the i-th row in the image, P col (j) is the gray projection value of the j-th column in the image.

[0084] The row projection curve and the column projection curve are preprocessed respectively, and the minimum value points are determined respectively;

[0085] According to the minimum value points of the row projection curve and the minimum value points of the column projection curve, a region of interest is determined in the image;

[0086] The pixel values of an external region outside the region of interest are set to be differentiated, so that a boundary line is generated between the external region and the region of interest;

[0087] The image is processed by an algorithm to determine the contour in the region of interest, and the contour line is fitted;

[0088] The center of the region surrounded by the fitted contour line is determined as the position of the target to be measured in the image.

[0089] The concept of determining the current position of the target to be measured in the image in the embodiment of the automatic assembling system of the shield segment is the same as that of the above-mentioned visual detection method embodiment, and the whole content of the above-mentioned visual detection method embodiment is incorporated herein by reference.

[0090] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0091] The above description is only a specific embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for visual inspection of shield segment assembly quality, characterized in that, The method comprises the following steps: Collecting an image of a shield segment surface containing a target to be detected, the size of the image being M*N; Calculating the row projection curve and the column projection curve of the image respectively by using the following formulas: where C(i, j) is the gray value at the (i, j) coordinate point in the image, P row (i) is the gray projection value of the i-th row in the image, P col (j) is the gray projection value of the j-th column in the image; Preprocessing the row projection curve and the column projection curve respectively, and determining the minimum value points of the two curves respectively; Determining a region of interest in the image according to the minimum value points of the row projection curve and the column projection curve, which comprises the following steps: determining the coordinate point corresponding to the minimum value point as the center point of interest according to the row and column values corresponding to the minimum value point; taking the center point of interest or the pixel points adjacent to the center point of interest as the expansion center points, and expanding outward to obtain the region of interest, wherein n is an integer less than 2%*min(M,N); further comprising the following steps: measuring the size of the region occupied by the target to be detected in the image in advance, which comprises the following steps: collecting images of the shield segment from different shooting angles to obtain multiple test images; determining the circumscribed rectangle of the target to be detected in each test image by using manual or image processing, and obtaining the maximum value of the long side length of each circumscribed rectangle; expanding the region of interest according to the maximum value of the long side length of the circumscribed rectangle, so that the region of interest can cover the area of a square with the maximum value of the long side length as the side length value; Differentiating the pixel values of the external region outside the region of interest to set a demarcation line between the external region and the region of interest; Processing the image by using an algorithm to determine the contour in the region of interest and fit a contour line to the contour; Determining the center of the region surrounded by the fitted contour line as the position of the target to be detected in the image, which comprises the following steps: if the target to be detected is a non-irregular-shaped part, fitting a minimum circumscribed rectangle to the contour, and taking the center of the circumscribed rectangle as the coordinate value of the target to be detected in the image in the current state; if the target to be detected is an irregular-shaped part, taking the contour boundary as the contour line, calculating the barycentric point of the internal region of the contour line, and taking the barycentric point as the coordinate value of the target to be detected in the image in the current state. 2.The method for visual inspection of shield segment assembly quality according to claim 1, characterized in that, Taking the number of rows or columns occupied by the longest side of each circumscribed rectangle as T, expanding T columns to the left and right directions respectively from the expansion center point, and expanding T rows to the up and down directions respectively to obtain the region of interest; or Taking the expansion center point as the center of a circle and T rows as the radius to draw the circle to obtain the region of interest. 3.The method for visual inspection of shield segment assembly quality according to claim 1, characterized in that, Taking the number of rows or columns occupied by the longest side of each circumscribed rectangle as T, expanding T / 2 columns to the left and right directions respectively from the expansion center point, and expanding T / 2 rows to the up and down directions respectively to obtain the region of interest. If the contour shape in the region of interest is determined to be unmatched with the shapes of various targets to be detected stored in the database, secondary expansion is performed, and the amplitude of the secondary expansion is greater than that of the previous expansion.

4. The method for visual inspection of shield segment assembly quality according to claim 1, wherein, The preprocessing of the row projection curve and the column projection curve and the determination of the minimum value points of the two curves respectively comprise the following steps: Smoothing the row projection curve and the column projection curve respectively by using the moving average method, and calculating the first derivative of the smoothed curves to determine the minimum value points of the two curves respectively.

5. The method for visual inspection of shield segment assembly quality according to claim 1, wherein, The image is filtered by a Gaussian filter, and a canny edge detection algorithm is used to determine the contour.

6. The method for visual inspection of shield segment assembly quality according to claim 1, wherein, The operation of setting the pixel value of the external region outside the region of interest is to set the pixel value of the external region outside the region of interest to 255.

7. A method of automatically assembling a shield segment, characterized by, The method comprises the following steps: The end position of the target to be measured in the image in the completed assembly state is calibrated in advance; The current position of the target to be measured in the image is determined by using the visual detection method according to any one of claims 1 to 6; The moving direction and distance of the target to be measured from the current position to the end position are determined; The assembly machine loads the target to be measured at the starting position and moves to the end position.

8. A shield segment automatic assembly system, characterized in that, The method comprises the following steps: The row projection curve and the column projection curve of the image collected by the camera are calculated by using the following formulas respectively: where C(i, j) is the gray value at the (i, j) coordinate point in the image, P row (i) is the gray projection value of the i-th row in the image, P col (j) is the gray projection value of the j-th column in the image; The row projection curve and the column projection curve are preprocessed respectively, and the minimum value points are determined respectively; The region of interest is determined in the image according to the minimum value points of the row projection curve and the minimum value points of the column projection curve, including: determining the coordinate point corresponding to the minimum value points as the center point of interest according to the numerical values of the row and the column corresponding to the minimum value points; taking the center point of interest or the pixel points adjacent to the center point of interest as the expansion center points, and expanding outward to obtain the region of interest, wherein n is an integer less than 2%*min(M, N); further comprising measuring the size of the region occupied by the target to be measured in the image in advance, including: collecting images of the shield segment at different shooting angles to obtain multiple test images; determining the circumscribed rectangle of the target to be measured in each test image by using artificial or image processing to obtain the maximum value of the long side length of each circumscribed rectangle; expanding the region of interest according to the maximum value of the long side length of the circumscribed rectangle, so that the region of interest can cover the area of a square with the maximum value of the long side length as the side length value; The pixel value of the external region outside the region of interest is set to 255 to form a boundary between the external region and the region of interest; The image is processed by an algorithm to determine the contour in the region of interest and fit a contour line, including: if the target to be measured is a non-special-shaped part, fitting a minimum circumscribed rectangle to the contour, and taking the center of the circumscribed rectangle as the coordinate value of the target to be measured in the image in the current state; if the target to be measured is a special-shaped part, taking the contour boundary as a contour line, calculating the barycenter of the internal region of the contour line as the coordinate value of the target to be measured in the image in the current state; determining the center of the region enclosed by the contour line to which the fitting is performed as the position of the target to be detected in the image.

Images