Non-contact strain gauge identification and coordinate accurate positioning method

Abstract

The invention discloses a non-contact strain gauge identification and coordinate accurate positioning method. The method comprises the following steps: acquiring strain gauge image data; carrying out graying processing on the strain gauge image data, and obtaining screened strain gauge contour image data; determining contour data of a target strain gauge; determining an essential matrix of the image acquisition equipment according to the image rotation matrix and the translation vector, identifying the same strain gauge in the contour data of the target strain gauge based on the essential matrix, performing coordinate positioning on the same strain gauge based on binocular vision, and converting the coordinate into a coordinate under the image acquisition equipment. Coordinate information of the strain gauge under the image acquisition equipment is obtained; establishing a conversion relation between the coordinate information and the target coordinate system, and converting the coordinate information into the target coordinate system to obtain a strain gauge identification result under the target coordinate system; and repeating the steps S1-S5 to obtain strain gauge identification results of all structure tests. The accuracy of the position coordinates of the strain gauges can be improved.

Description

Technical Field

[0001] The invention relates to the technical field of structural testing, and in particular to a non-contact strain gauge identification and coordinate precise positioning method. Background Art

[0002] Strain gauges are the main means of measuring strain. In structural tests, a large number of strain gauges are often required to monitor the state of the structure. However, in traditional structural tests, strain gauges are usually manually pasted, and there are deviations in the pasting position. In addition, the coordinates of the strain gauges are usually not checked, resulting in the inability to guarantee measurement accuracy. In high strain gradient areas of the structure, the deviation in the strain gauge pasting position will lead to a significant decrease in measurement accuracy. Therefore, the development of strain gauge coordinate positioning methods is particularly important for improving test accuracy.

[0003] At present, the coordinate measurement of strain gauges is mainly assisted by tools such as rulers, and mainly adopts contact measurement. Due to the complexity and large size of the structure, there are inevitably problems such as difficult operation and inaccurate positioning. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the present invention provides a non-contact strain gauge identification and coordinate precise positioning method, which solves the problem of low accuracy of the identification result of the strain gauge position coordinates in the prior art.

[0005] In order to achieve the above-mentioned purpose of the invention, the technical solution adopted by the present invention is: a non-contact strain gauge identification and coordinate precise positioning method, comprising the following steps:

[0006] S1. Collect the strain gauge image data through the automated controllable device equipped with the image acquisition device, and perform rotation transformation during the acquisition process to obtain the strain gauge image data in different postures.

[0007] S2, graying the strain gauge image data to obtain strain gauge graying image data, extracting the strain gauge graying image data to obtain strain gauge contour image data, and filtering the strain gauge contour image data to obtain filtered strain gauge contour image data;

[0008] S3, determining target strain gauge contour data according to the screened strain gauge contour image data and the strain gauge image data corresponding to the screened strain gauge contour image data;

[0009] S4. Determine the intrinsic matrix of the image acquisition device according to the image rotation matrix and the translation vector, identify the same strain gauge in the target strain gauge contour data based on the intrinsic matrix, and locate the coordinates of the same strain gauge based on binocular vision, convert it into the coordinates under the image acquisition device, and obtain the coordinate information of the strain gauge under the image acquisition device;

[0010] S5, establishing a conversion relationship between the coordinate information and the target coordinate system, and converting the coordinate information to the target coordinate system to obtain a strain gauge identification result in the target coordinate system;

[0011] S6. Repeat steps S1-S5 to obtain strain gauge identification results for all structural tests.

[0012] The beneficial effect of the above scheme is as follows: the present invention utilizes the geometric and color features of the strain gauge for identification, obtains the information of the strain gauge contour and color respectively by processing the collected image in grayscale and color, performs strain gauge identification, and obtains the matching relationship between the same strain gauge in different images based on the posture relationship between the images, locates the coordinates based on the binocular principle, and finally converts the coordinates to the corresponding coordinate system to complete the non-contact positioning of the coordinates, thereby improving the accuracy of the strain gauge position coordinates.

[0013] Furthermore, step S2, screening the strain gauge contour image data, includes:

[0014] S21, determining a polygon measurement threshold according to the aspect ratio of the minimum circumscribed rectangle of the outline, the ratio of the area of ​​the minimum circumscribed rectangle of the outline to the area of ​​the outline, and the number of sides of the outline fitting polygon;

[0015] S22. Filter the strain gauge contour image data according to the polygon measurement threshold.

[0016] The beneficial effect of the above further scheme is that through the polygon measurement threshold, noise or non-target objects that do not conform to physical characteristics can be effectively eliminated, and only possible strain gauge contours can be retained, which helps to improve the accuracy of subsequent processing and reduce false detection and missed detection. Setting a reasonable polygon measurement threshold can filter out a large number of irrelevant contours at an early stage, reduce unnecessary further processing, thereby saving computing resources and time, and improving the processing efficiency of the overall system.

[0017] Further, step S3 specifically includes:

[0018] S31, overlaying the filtered strain gauge contour image data on the strain gauge image data corresponding to the filtered strain gauge contour image data, segmenting the strain gauge image data corresponding to the filtered strain gauge contour image data, and obtaining segmented strain gauge image data;

[0019] S32, obtaining the average value of the three channels of RGB color in the segmented strain gauge image data as the main color information of the contour area;

[0020] S33. Screen the segmented strain gauge image data according to the main color information of the contour area to determine the target strain gauge contour data.

[0021] The beneficial effect of the above further scheme is: by accurately covering and segmenting the original strain gauge image data with the screened contour, the strain gauge area is finely isolated, the background interference can be effectively eliminated, and the subsequent analysis is focused on the real strain gauge area, which improves the accuracy and pertinence of the analysis. Extracting the average value of the RGB color three channels of the segmented strain gauge area as the main color information of the contour area helps to identify and quantify the color characteristics of the strain gauge, judge the strain state based on color changes, improve the dimension and depth of data analysis, help to exclude areas that do not meet the requirements due to color deviation, and ensure the reliability and pertinence of the analysis results.

[0022] Furthermore, in step S22, the strain gauge contour image data is screened using the formula:

[0023]

[0024] Wherein, retain means retaining the strain gauge contour image data that meets the polygon measurement threshold, and delete means deleting the strain gauge contour image data that does not meet the polygon measurement threshold.

[0025] Furthermore, in step S33, the segmented strain gauge image data is screened according to the main color information of the contour area, and the formula used is:

[0026]

[0027] Among them, (c h ,c s ,c v ) represents the color information of the area within the contour in the segmented strain gauge image data, (r h ,r s ,r v ) represents the main color information of the contour area.

[0028] Furthermore, in step S4, the coordinates of the same strain gauge are positioned based on binocular vision, and the essential matrix E is:

[0029] E=T×R=[T] × R

[0030] Where T represents the translation vector, which is represented by t x ,t y ,t z R represents the image rotation matrix, whose size is 3*3, T×R represents the vector product, [T] × represents the antisymmetric matrix of the translation vector T, and t x ,t y ,t zThey represent the magnitude of translation in the x, y, and z directions respectively.

[0031] Furthermore, in step S4, based on the essential matrix, the same strain gauge in the target strain gauge contour data is identified using the formula:

[0032]

[0033] Among them, M nm represents the corresponding matching matrix of all strain gauges in the two images, represents the transposed matrix of the coordinates of the key points in the first target strain gauge contour data, N represents the number of key points in the first target strain gauge contour data, n represents the nth key point in the first target strain gauge contour data, p r =(u r ,v r ), p lm (m=1,2...,M) represents the coordinates of the key points in the second target strain gauge contour data, M represents the number of key points in the second target strain gauge contour data, m represents the mth key point in the second target strain gauge contour data, p l =(u l ,v l ), and the target strain gauge profile data includes first target strain gauge profile data and second target strain gauge profile data, Where E represents the essential matrix.

[0034] Furthermore, in step S4, the coordinates are converted to the image acquisition device using the following formula:

[0035]

[0036] (z ci z cj )=(C T C) -1 C T T

[0037]

[0038] Where A represents the camera intrinsic parameter matrix, f x 、f y , s represents the parameters in the camera intrinsic parameter matrix, (u0, v0) represents the pixel value of the key point in the key point, C represents the process matrix composed of pixel coordinates, camera intrinsic parameters, and rotation matrix, R represents the rotation matrix between different camera positions, (u i ,v i ) represents the pixel coordinates of the key points of the strain gauge in the first target strain gauge contour data, (u j ,v j) represents the pixel coordinates of the key points of the strain gauge in the second target strain gauge contour data, (z ci z cj ) represents the depth value of the strain gauge in different camera coordinate systems, T represents the translation vector between cameras, y cj Indicates the horizontal distance from the origin of the camera coordinate system, x cj Indicates the vertical distance from the origin of the camera coordinate system.

[0039] The beneficial effect of the above further scheme is that by calculating the essential matrix, the geometric relationship between the two camera coordinate systems can be accurately established, and combined with the principle of binocular stereo vision, the strain gauge can be accurately positioned in three-dimensional space, which significantly improves the accuracy and stability of spatial coordinate measurement. It supports automatic positioning and tracking of strain gauges, reduces the need for manual intervention, improves monitoring efficiency, and is suitable for large-scale, long-term structural health monitoring projects.

[0040] Furthermore, the automated controllable equipment includes collaborative robots and drones; the image acquisition equipment includes a single camera and two cameras with known posture relationships.

[0041] The beneficial effects of the above further scheme are: drones equipped with cameras can conduct remote, efficient and large-scale monitoring, significantly improving safety and work efficiency. Collaborative robots can perform detailed and accurate inspections indoors or in specific places. Binocular cameras provide depth information through parallax calculation, realize three-dimensional reconstruction of target objects, and provide richer visual information. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 The figure is a flow chart of a non-contact strain gauge identification and coordinate precise positioning method.

[0043] Figure 2 The figure is a schematic diagram of a device for non-contact strain gauge identification and precise coordinate positioning method.

[0044] Figure 3 A schematic diagram of another device for a non-contact strain gauge identification and precise coordinate positioning method.

[0045] Figure 4 Schematic diagram of the threshold value for polygon measurement.

[0046] Figure 5 Schematic diagram of the process of non-contact strain gauge identification. DETAILED DESCRIPTION

[0047] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0048] like Figure 1As shown, a non-contact strain gauge identification and coordinate precise positioning method includes the following steps:

[0049] S1. The strain gauge image data is collected by an automated controllable device equipped with an image acquisition device, and a rotation transformation is performed during the collection process to obtain the strain gauge image data in different postures.

[0050] In this embodiment, the automated controllable devices include collaborative robots and drones; the image acquisition devices include a single camera and two cameras with known posture relationships.

[0051] For example, this embodiment is carried out in a structural test. In addition, the strain gauge can be a unidirectional strain gauge, a strain rosette or other strain gauge with polygonal features.

[0052] For example, Figure 2 As shown, Figure 2 The figure is a schematic diagram of a non-contact strain gauge identification and coordinate precision positioning method. Figure 2 (a) Schematic diagram of a small square tube structure. Figure 2 (b) is a schematic diagram of an automated controllable device equipped with an image acquisition device corresponding to a small square tube structure. Figure 2 In (a), the small square tube structure can be fixed with a unidirectional strain gauge (e.g. Figure 2 (a) as shown in the small square block), and around the small square tube structure can be arranged as follows Figure 2 (b) and using Figure 2 The camera and lens in (b) collect the strain gauge image data of the unidirectional strain gauge.

[0053] For example, Figure 3 Another schematic diagram of a non-contact strain gauge identification and coordinate precision positioning method. Figure 3 In the embodiment, the automated controllable device is a six-degree-of-freedom collaborative robot. In the six-degree-of-freedom collaborative robot, a corresponding image acquisition device can be set to collect image data of strain gauges in different positions in the thin-walled cylindrical shell structure. Figure 3 The right side in the middle is a schematic diagram of the device facing the cylindrical shell structure. Figure 3 The left side of the middle is a schematic diagram of the six-degree-of-freedom collaborative robot corresponding to the cylindrical shell structure. Figure 3 In the embodiment, a unidirectional strain gauge can be fixed on the surface facing the cylindrical shell structure, and a six-degree-of-freedom collaborative robot can be arranged around the cylindrical shell structure. The camera and lens in the six-degree-of-freedom collaborative robot are used to collect the strain gauge image data of the unidirectional strain gauge. Figure 3The collaborative robot with six degrees of freedom is included in two postures, namely, the collaborative robot in the initial position and the collaborative robot after the posture is changed. The posture change can be completed by the mechanical arm of the collaborative robot with six degrees of freedom.

[0054] S2. Gray-scale the strain gauge image data to obtain strain gauge gray-scale image data, extract the strain gauge gray-scale image data to obtain strain gauge contour image data, and filter the strain gauge contour image data to obtain filtered strain gauge contour image data.

[0055] In step S2 of this embodiment, the strain gauge contour image data is screened, including:

[0056] S21, determining a polygon measurement threshold according to the aspect ratio of the minimum circumscribed rectangle of the outline, the ratio of the area of ​​the minimum circumscribed rectangle of the outline to the area of ​​the outline, and the number of sides of the outline fitting polygon;

[0057] S22. Filter the strain gauge contour image data according to the polygon measurement threshold.

[0058] In step S22 of this embodiment, the strain gauge contour image data is screened using the formula:

[0059]

[0060] Wherein, retain means retaining the strain gauge contour image data that meets the polygon measurement threshold, and delete means deleting the strain gauge contour image data that does not meet the polygon measurement threshold.

[0061] This embodiment can filter out interference contours and reduce the amount of calculation by screening the strain gauge contour image data.

[0062] Figure 4 This is a schematic diagram of the polygon measurement threshold. Figure 4 (a) is a schematic diagram of the threshold value measured by the aspect ratio of the minimum circumscribed rectangle of the contour. Figure 4 (b) is a schematic diagram of the polygon measurement threshold based on the ratio of the minimum enclosing rectangle area to the contour area.

[0063] Optionally, the screening of the strain gauge contour image data in step S2 may not be performed, that is, the strain gauge image data may be grayscaled in step S2 to obtain strain gauge grayscale image data, the strain gauge grayscale image data may be extracted to obtain strain gauge contour image data, and step S3 may be performed.

[0064] S3. Determine target strain gauge contour data according to the filtered strain gauge contour image data and the strain gauge image data corresponding to the filtered strain gauge contour image data.

[0065] In this embodiment, step S3 specifically includes:

[0066] S31, overlaying the filtered strain gauge contour image data on the strain gauge image data corresponding to the filtered strain gauge contour image data, segmenting the strain gauge image data corresponding to the filtered strain gauge contour image data, and obtaining segmented strain gauge image data;

[0067] S32, obtaining the average value of the three channels of RGB color in the segmented strain gauge image data as the main color information of the contour area;

[0068] S33. Screen the segmented strain gauge image data according to the main color information of the contour area to determine the target strain gauge contour data.

[0069] In step S33 of this embodiment, the segmented strain gauge image data is screened according to the main color information of the contour area, and the formula used is:

[0070]

[0071] Among them, (c h ,c s ,c v ) represents the color information of the area within the contour in the segmented strain gauge image data, (r h ,r s ,r v ) represents the main color information of the contour area.

[0072] Optionally, in step S32, a k-means clustering method may be used to cluster the segmented strain gauge image data to obtain main color information of the contour area, which is not specifically limited herein.

[0073] Optionally, in step S32, the image format of the segmented strain gauge image data may be an RGB image, an HSV image, or an HSL image, which is not specifically limited herein.

[0074] S4. Determine the intrinsic matrix of the image acquisition device according to the image rotation matrix and the translation vector. Based on the intrinsic matrix, identify the same strain gauge in the target strain gauge contour data, and locate the coordinates of the same strain gauge based on binocular vision, convert it into the coordinates under the image acquisition device, and obtain the coordinate information of the strain gauge under the image acquisition device.

[0075] In step S4 of this embodiment, the coordinates of the same strain gauge are positioned based on binocular vision, and the essential matrix E is:

[0076] E=T×R=[T] × R

[0077] Where T represents the translation vector, which is represented by t x ,t y ,t z R represents the image rotation matrix, whose size is 3*3, T×R represents the vector product, [T] × represents the antisymmetric matrix of the translation vector T, and t x ,t y ,t z They represent the magnitude of translation in the x, y, and z directions respectively.

[0078] Based on the essential matrix, the same strain gauge in the target strain gauge contour data is identified using the formula:

[0079]

[0080] Among them, M nm represents the corresponding matching matrix of all strain gauges in the two images, represents the transposed matrix of the coordinates of the key points in the first target strain gauge contour data, N represents the number of key points in the first target strain gauge contour data, n represents the nth key point in the first target strain gauge contour data, p r =(u r ,v r ), p lm (m=1,2...,M) represents the coordinates of the key points in the second target strain gauge contour data, M represents the number of key points in the second target strain gauge contour data, m represents the mth key point in the second target strain gauge contour data, p l =(u l ,v l ), and the target strain gauge profile data includes first target strain gauge profile data and second target strain gauge profile data, E represents the essential matrix.

[0081] The formula used to convert to the coordinates of the image acquisition device is:

[0082]

[0083] (z ci z cj )=(C T C) -1 C T T

[0084]

[0085] Where A represents the camera intrinsic parameter matrix, f x 、f y, s represents the parameters in the camera intrinsic parameter matrix, (u0, v0) represents the pixel value of the key point in the key point, C represents the process matrix composed of pixel coordinates, camera intrinsic parameters, and rotation matrix, R represents the rotation matrix between different camera positions, (u i ,v i ) represents the pixel coordinates of the key points of the strain gauge in the first target strain gauge contour data, (u j ,v j ) represents the pixel coordinates of the key points of the strain gauge in the second target strain gauge contour data, (z ci z cj ) represents the depth value of the strain gauge in different camera coordinate systems, T represents the translation vector between cameras, y cj Indicates the horizontal distance from the origin of the camera coordinate system, x cj Indicates the vertical distance from the origin of the camera coordinate system.

[0086] Figure 5 This is a schematic diagram of the process of non-contact strain gauge identification. Figure 5 (a) is a schematic diagram of using HSV color image as strain gauge image data. Figure 5 (b) is a schematic diagram of the process of screening the strain gauge contour image data. Figure 5 (c) Schematic diagram of the process of determining the target strain gauge profile data.

[0087] Optionally, the first target strain gauge contour data may be any target strain gauge contour data of the target strain gauge contour data, and the second target strain gauge contour data may also be any target strain gauge contour data of the target strain gauge contour data. Furthermore, based on the essential matrix, the same strain gauge in the target strain gauge contour data may be identified by using two images in the target strain gauge contour data, or multiple (greater than 2) images in the target strain gauge contour data may be used to complete the identification of the same strain gauge, or two images in the target strain gauge contour data may be used to complete the identification of multiple identical strain gauges, or multiple (greater than 2) images in the target strain gauge contour data may be used to complete the identification of multiple identical strain gauges.

[0088] Optionally, when multiple (greater than 2) images in the target strain gauge contour data are used to complete the identification of the strain gauge, the images may be combined in pairs first, and finally the least square method is calculated for the multiple positioning results to obtain the final identification result.

[0089] S5. Establish a conversion relationship between the coordinate information and the target coordinate system, and convert the coordinate information into the target coordinate system to obtain a strain gauge identification result in the target coordinate system.

[0090] Optionally, the target coordinate system may refer to the test structure coordinate system, and the conversion relationship between the coordinate information and the target coordinate system may be obtained by measurement or calculated from marker points with known spatial position relationships in the image.

[0091] S6. Repeat steps S1-S5 to obtain strain gauge identification results for all structural tests.

[0092] In this embodiment, non-contact positioning of strain gauge coordinates is achieved based on the principle of image processing and visual positioning, which effectively improves the positioning accuracy to 1mm. Compared with the traditional contact positioning method, this embodiment has a high degree of automation and is easy to operate. It can quickly position strain gauges for large and complex structures, shorten the test preparation cycle while ensuring accuracy, and improve test accuracy.

[0093] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific variations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed by the present invention, and these variations and combinations are still within the protection scope of the invention.

Claims

1. A non-contact strain gauge identification and coordinate precise positioning method, It is characterized in that The method comprises: S1. Collect the strain gauge image data through the automated controllable device equipped with the image acquisition device, and perform rotation and translation transformation during the acquisition process to obtain the strain gauge image data in different postures. S2, graying the strain gauge image data to obtain strain gauge graying image data, extracting the strain gauge graying image data to obtain strain gauge contour image data, and filtering the strain gauge contour image data to obtain filtered strain gauge contour image data; S3, determining target strain gauge contour data according to the screened strain gauge contour image data and the strain gauge image data corresponding to the screened strain gauge contour image data; S4, determining the intrinsic matrix of the image acquisition device according to the image rotation matrix and the translation vector, identifying the same strain gauge in the target strain gauge contour data based on the intrinsic matrix, and locating the same strain gauge based on binocular vision, converting it into the coordinates under the image acquisition device, and obtaining the coordinate information of the strain gauge under the image acquisition device; S5, establishing a conversion relationship between the coordinate information and a target coordinate system, and converting the coordinate information into the target coordinate system to obtain a strain gauge identification result in the target coordinate system; S6. Repeat steps S1-S5 to obtain strain gauge identification results for all structural tests.

2. The method according to claim 1, It is characterized in that In the step S2, the screening of the strain gauge contour image data includes: S21, determining a polygon measurement threshold according to the aspect ratio of the minimum circumscribed rectangle of the outline, the ratio of the area of ​​the minimum circumscribed rectangle of the outline to the area of ​​the outline, and the number of sides of the outline fitting polygon; S22. Filter the strain gauge contour image data according to the polygon measurement threshold.

3. The method according to claim 1, It is characterized in that The step S3 specifically includes: S31, overlaying the filtered strain gauge contour image data on the strain gauge image data corresponding to the filtered strain gauge contour image data, segmenting the strain gauge image data corresponding to the filtered strain gauge contour image data, to obtain segmented strain gauge image data; S32, obtaining the average value of the three channels of RGB color in the segmented strain gauge image data as the main color information of the contour area; S33: Screening the segmented strain gauge image data according to the main color information of the contour area to determine the target strain gauge contour data.

4. The method according to claim 2, It is characterized in that In the step S22, the strain gauge contour image data is screened using the formula: Wherein, retain means retaining the strain gauge contour image data that meets the polygon measurement threshold, and delete means deleting the strain gauge contour image data that does not meet the polygon measurement threshold.

5. The method according to claim 2, It is characterized in that In the step S33, the segmented strain gauge image data is screened according to the main color information of the contour area, and the formula used is: Among them, (c h ,c s ,c v ) represents the color information of the area within the contour in the segmented strain gauge image data, (r h ,r s ,r v ) represents the main color information of the contour area.

6. The method according to claim 1, It is characterized in that In step S4, the same strain gauge is positioned based on binocular vision, and the essential matrix E is: E=T×R=[T] × R Where T represents the translation vector, which is represented by t x ,t y ,t z R represents the image rotation matrix, whose size is 3*3, T×R represents the vector product, [T] × represents the antisymmetric matrix of the translation vector T, and t x ,t y ,t z They represent the magnitude of translation in the x, y, and z directions respectively.

7. The method according to claim 1, It is characterized in that In the step S4, based on the essential matrix, the same strain gauge in the target strain gauge contour data is identified using the formula: Among them, M nm represents the corresponding matching matrix of all strain gauges in the two images, represents the transposed matrix of the coordinates of the key points in the first target strain gauge contour data, N represents the number of key points in the first target strain gauge contour data, n represents the nth key point in the first target strain gauge contour data, p r =(u r ,v r ), p lm (m=1,2...,M) represents the coordinates of the key points in the second target strain gauge contour data, M represents the number of key points in the second target strain gauge contour data, m represents the mth key point in the second target strain gauge contour data, p l =(u l ,v l ), and the target strain gauge profile data includes first target strain gauge profile data and second target strain gauge profile data, Where E represents the essential matrix.

8. The method according to claim 1, It is characterized in that In step S4, the conversion to the coordinates of the image acquisition device uses the following formula: (z ci z cj )=(C T C) -1 C T T Where A represents the camera intrinsic parameter matrix, f x 、f y , s represents the parameters in the camera internal parameter matrix, (u 0 ,v 0 ) represents the pixel value of the key point in the key point, C represents the process matrix composed of pixel coordinates, camera intrinsic parameters, and rotation matrix, R represents the rotation matrix between different camera positions, (u i ,v i ) represents the pixel coordinates of the key points of the strain gauge in the first target strain gauge contour data, (u j ,v j ) represents the pixel coordinates of the key points of the strain gauge in the second target strain gauge contour data, (z ci z cj ) represents the depth value of the strain gauge in different camera coordinate systems, T represents the translation vector between cameras, y cj Indicates the horizontal distance from the origin of the camera coordinate system, x cj Indicates the vertical distance from the origin of the camera coordinate system.

9. The method according to claim 1, It is characterized in that The automated controllable equipment includes a collaborative robot and a drone; the image acquisition equipment includes a single camera and two cameras with known posture relationships.

Images