Method for evaluating positioning precision of underground equipment

Abstract

The invention discloses a method for evaluating the positioning precision of underground equipment. The method comprises the following steps: S1, mounting a camera and designing and arranging a verification mark; s2, acquiring image information; s3, checking mark detection; s4, carrying out binarization processing; s5, extracting a verification point; s6, performing pose matching; s7, performing precision evaluation; according to the real-time positioning accuracy evaluation scheme of the underground equipment, the problem of accurate evaluation of the positioning accuracy of the underground environment is solved, and safe and stable operation of the underground equipment is guaranteed.

Description

Technical Field

[0001] This invention relates to the field of evaluating downhole positioning systems, and more particularly to a method for evaluating the positioning accuracy of downhole equipment. Background Technology

[0002] Currently, intelligent and unmanned operations in underground mines are being gradually promoted nationwide. The key to achieving unmanned equipment operation lies in real-time underground positioning technology; therefore, accurately assessing the real-time positioning accuracy of underground equipment is crucial for ensuring its safe and stable operation. Currently, commonly used methods for assessing positioning accuracy on the surface mainly rely on high-precision RTK real-time dynamic differential positioning and motion capture equipment. However, these traditional methods are difficult to adapt to the underground environment: RTK cannot receive satellite signals underground, thus failing; and motion capture systems have extremely high requirements for the installation environment, making them unsuitable for installation in the narrow tunnels of underground mines. Furthermore, both methods are very expensive, typically requiring an investment of hundreds of thousands or even more. Therefore, there is an urgent need for a lower-cost, more adaptable solution specifically designed for assessing the accuracy of underground positioning systems. Summary of the Invention

[0003] The present invention aims to address the shortcomings of the prior art by providing a method for evaluating the positioning accuracy of downhole equipment.

[0004] To achieve the above objectives, the present invention adopts the following technical solution:

[0005] A method for evaluating the positioning accuracy of downhole equipment includes the following steps:

[0006] S1. Camera installation and verification mark design and layout: The camera is installed in the underground equipment and verification marks are posted in the mine roadway at preset intervals or randomly. The verification mark pattern consists of a QR code in the center and a black and white checkerboard pattern around it.

[0007] S2. Acquire image information: Capture the pattern of the verification mark using a camera installed on the downhole equipment;

[0008] S3. Verification Mark Detection: Perform QR code detection on the verification mark to obtain the ID information and region information of the verification mark;

[0009] S4. Binarization Processing: The detected check flag region is binarized to accelerate the matching speed;

[0010] S5. Extract checkpoints: Use a preset convolution kernel to extract checkpoints, obtaining a set of checkpoints corresponding to the check flag. ;

[0011] S6. Pose Matching: For each verification marker, a set of continuous detection values ​​is obtained. Each detection value consists of QR code information and verification point information, denoted as... For the same verification mark area, a set of consecutive detection results are obtained. , The number of images with a check mark at this location; during the mapping and localization processes, a series of results for check mark detection are obtained, and these results are matched one by one to obtain the pose corresponding to the mapping and the pose corresponding to the real-time localization at the same location;

[0012] S7. Accuracy Evaluation: For each checkpoint, a set of consecutive matching results is obtained. For a certain checkpoint, there is a set of consecutive images. Each frame of the image corresponds to a set of mutually matching mappings, which is the corresponding pose. Real-time positioning corresponds to pose. The position error for this matching set is calculated as follows: This is the absolute trajectory error between the mapped positioning result and the real-time positioning result.

[0013] ,

[0014] in The confidence coefficient is equal to the percentage of successful verification points in that frame of image, used to evaluate the positioning accuracy of downhole equipment in real time.

[0015] The steps for detecting the verification mark are as follows: First, QR code detection is performed to obtain the pattern ID information and the outline position of the QR code in the image; then, based on the fixed proportional relationship between the QR code and the entire verification mark, the region information and size information of the verification mark in the image are calculated. ,when or or or If the verification flag area exceeds the entire image range, the result should be discarded. Here, id represents the verification flag ID. The x-coordinate of the center point of the QR code. The vertical coordinate of the center point of the QR code. The width of the logo area, The height of the logo area, The width of the image. The height of the image.

[0016] The binarization process involves iterating through each pixel in the pattern area. When the pixel value is greater than 100 in all three channels (R, G, and B), the corresponding pixel value is set to 1; otherwise, it is set to 0, generating a binarization matrix composed of 0s and 1s.

[0017] The steps for extracting the verification point are as follows: A 5×5 convolution kernel is used to traverse the entire detected verification flag. When the convolution kernel completely matches the pixel value of the region, the center point of that region is marked as the verification point. Where n is the number of check points. Represents the coordinates of a pixel.

[0018] When the accuracy assessment is performed, if there are n corresponding images at the pattern marker, the average positioning error at the pattern marker is: The average positioning error corresponding to the remaining pattern marks is obtained based on the accuracy evaluation steps.

[0019] The specific steps of pose matching are as follows:

[0020] P1. Extract a detection result from the poses corresponding to all real-time localizations. Iterate through all the poses corresponding to the mapping to find a result. ,satisfy and The corresponding IDs are equal;

[0021] P2. For Each checkpoint in ,exist Search for a matching value in the middle, if There exists a condition where the following relationship is satisfied for a match to be considered successful:

[0022] - 2 <= +2;

[0023] - 2 <= +2, of which Indicates the allowable error range;

[0024] P3. Mark the successfully matched checkpoints and remove them from the matching process. Then, iterate through all points according to step P2, counting the number of successful matches. When 80% of the points are successfully matched, the matching is considered complete. and It is the same location;

[0025] P4. Extract separately and The timestamp of the image is used to locate the corresponding device location information.

[0026] The beneficial effects of this invention are: the real-time positioning accuracy evaluation scheme for downhole equipment proposed in this invention solves the problem of accurate evaluation of positioning accuracy in the downhole environment and ensures the safe and stable operation of downhole equipment. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of the camera installation and verification mark arrangement according to the present invention;

[0028] Figure 2 This is a design diagram of the verification mark for the present invention;

[0029] Figure 3 This is a schematic diagram of the pose matching process of the present invention;

[0030] Figure 4 This is the binarization process of the present invention;

[0031] Figure 5 This is a schematic diagram of the convolution kernel of the present invention;

[0032] Figure 6 This is the convolution calculation process of the present invention;

[0033] Figure 7 This is the convolution traversal process of the present invention;

[0034] Figure 8 This is a schematic diagram of the verification points of the present invention;

[0035] In the diagram: 1-Downhole equipment; 2-Camera; 3-Verification mark;

[0036] The following will describe in detail, with reference to the accompanying drawings, embodiments of the invention. Detailed Implementation

[0037] The present invention will be further described below with reference to the accompanying drawings and embodiments:

[0038] A method for evaluating the positioning accuracy of downhole equipment includes the following steps:

[0039] S1. Camera Installation and Verification Mark Design and Layout: Camera 2 is installed on underground equipment 1. Verification marks 3 are posted in the mine roadway at preset intervals or randomly. The pattern of verification mark 3 consists of a QR code in the center and a black and white checkerboard pattern around it. Camera 2 uses an RGB camera to provide stable color images. Underground equipment 1 needs to be equipped with a positioning algorithm to effectively output positioning information. The QR code in the center of verification mark 3 is used to provide ID information and pattern mark position information. The remaining black and white checkerboard pattern provides stable verification points to improve matching accuracy.

[0040] S2. Acquire image information: Capture the pattern of the verification mark 3 using the camera 2 installed on the downhole equipment 1;

[0041] S3. Verification mark detection: Perform QR code detection on verification mark 3 to obtain the ID information and region information of verification mark 3;

[0042] S4. Binarization Processing: The detected check mark 3 region is binarized to accelerate the matching speed;

[0043] S5. Extract checkpoints: Use a preset convolution kernel to extract checkpoints, obtaining a set of checkpoints corresponding to checkpoint flag 3. ;

[0044] S6. Pose Matching: At each verification marker 3, a set of continuous detection values ​​is obtained. Each detection value consists of QR code information and verification point information, denoted as... For the same area marked by check mark 3, a set of consecutive detection results were obtained. , The number of images with verification mark 3 appearing at this location; during the mapping and localization processes, a series of results for the detection of verification mark 3 are obtained, and these results are matched one by one to obtain the pose corresponding to the mapping and the pose corresponding to the real-time localization at the same location;

[0045] S7. Accuracy Evaluation: For each set of consecutive matching results obtained at checkpoint 3, for a certain checkpoint 3, there is a set of consecutive images. Each frame of the image corresponds to a set of mutually matching mappings, which is the corresponding pose. Real-time positioning corresponds to pose. The position error for this matching set is calculated as follows: This is the absolute trajectory error between the mapped positioning result and the real-time positioning result.

[0046] ,

[0047] in The confidence coefficient is equal to the percentage of successful verification points in the image frame, used to evaluate the positioning accuracy of downhole equipment 1 in real time.

[0048] The steps for detecting the verification mark 3 are as follows: First, QR code detection is performed to obtain the pattern ID information and the outline position of the QR code in the image; then, based on the fixed proportional relationship between the QR code and the entire verification mark 3, the region information and size information of the verification mark 3 in the image are calculated. ,when or or or If the area containing check flag 3 exceeds the entire image range, the result should be discarded. Here, id represents the check flag ID. The x-coordinate of the center point of the QR code. The vertical coordinate of the center point of the QR code. The width of the logo area, The height of the logo area, The width of the image. The height of the image.

[0049] The primary purpose of QR code detection is to obtain pattern ID information and the QR code's outline position within the image. General detection algorithms can be used for this purpose without affecting the essence of this invention. The QR code detection return value data is defined as follows: The physical meanings within the parentheses are respectively the verification flag id, the x-coordinate of the QR code center point, the y-coordinate of the QR code center point, the pixel width of the QR code area, and the pixel height of the QR code area.

[0050] Since the size of verification mark 3 is fixed, although the pixel size occupied by verification mark 3 in the image will vary depending on the distance between the camera and verification mark 3, there is a fixed proportional relationship between the size of the entire verification mark 3 and the size of the QR code in the image. When the QR code is detected, the width and height of the entire verification mark 3 can be obtained from the width and height of the QR code area. Let's assume the size ratio between the QR code and verification mark 3 is:

[0051] ;

[0052] ;

[0053] Considering the extraction of verification points, the entire pattern mark area is expanded, and the width and height of the pattern mark area can be obtained in the end:

[0054] ;

[0055] ;

[0056] For each frame of the image, we can obtain the corresponding ID and region information. ;

[0057] when or or or If the area marked by check mark 3 exceeds the entire image range, the result should be discarded. and This represents the width and height of the image.

[0058] Because the image frame rate is relatively high, typically 30 frames can be output per second, a continuous set of detection values ​​can be obtained each time the check flag 3 is detected. n represents the number of images that can detect the same check flag 3.

[0059] The binarization process involves iterating through each pixel in the pattern area. When the pixel value is greater than 100 in all three channels (R, G, and B), the corresponding pixel value is set to 1; otherwise, it is set to 0, generating a binarization matrix composed of 0s and 1s.

[0060] To accelerate the matching process, the detected check flag 3 region is binarized. In a color image, for black pixel values... For white pixel values However, due to the changes in lighting in the mine, the photos obtained by the camera are generally not exactly 0 or 255, but fluctuate around 0 or 255, so binarization is required.

[0061] The binarization process iterates through each pixel in the pattern area. If a pixel's value is greater than 100 in all three channels (R, G, B), the corresponding pixel value is set to 1; otherwise, it is set to 0. Assuming the pattern area is a 4×4 rectangular region, the binarization process is illustrated in the attached diagram. Figure 4 As shown. (Attached) Figure 4 The data to the left of the arrow represents the original image data, and the data to the right represents the processed binarized data. Let m be the binarized data of each check flag region, where m is a matrix consisting of 0s and 1s.

[0062] The steps for extracting the verification points are as follows: Traverse the entire detected verification flag 3 using a 5×5 convolution kernel, with the convolution kernel designed as follows: Figure 5 As shown, both the convolution kernel and the binarized checksum region 3 are currently matrices composed of 0s and 1s. During the traversal, the pixel values ​​at corresponding positions of the convolution kernel and the checksum region 3 are compared. When the pixel values ​​of the convolution kernel and the checksum region 3 are completely equal, it indicates that the center pixel of this 5×5 region is a checksum point. The convolution calculation process is shown in Appendix 6. If the convolution result of the upper half is 0, then the gray pixel is a feature point; if the convolution result of the lower half is not 0, then the gray pixel is not a feature point. Figure 7 This represents the convolution traversal process. Based on the above process, all check points can be obtained from all check flag regions, as shown in the appendix. Figure 8 As shown, finally, by extracting the verification points, the center point of the area is obtained as the verification point. Where n is the number of check points. Represents the coordinates of a pixel.

[0063] When the accuracy assessment is performed, if there are n corresponding images at the pattern marker, the average positioning error at the pattern marker is: The average positioning error corresponding to the remaining pattern marks is obtained based on the accuracy evaluation steps.

[0064] The specific steps of pose matching are as follows: For each verification marker 3, a set of continuous detection values ​​can be obtained. Each detection value consists of QR code information and verification point information, denoted as... Therefore, a set of consecutive detection results can be obtained for the same area of ​​verification mark 3. , This represents the number of images where the check flag 3 appears at this location.

[0065] During the mapping and real-time positioning processes, a series of results regarding the detection of check mark 3 can be obtained. These results can then be matched one by one to find the mapping and real-time positioning results at a specific location.

[0066] P1. Extract a detection result from the poses corresponding to all real-time localizations. Iterate through all the poses corresponding to the mapping to find a result. ,satisfy and The corresponding IDs are equal;

[0067] P2. For Each checkpoint in ,exist Search for a matching value in the middle, if There exists a condition where the following relationship is satisfied for a match to be considered successful:

[0068] - 2 <= +2;

[0069] - 2 <= +2, of which Indicates the allowable error range;

[0070] P3. Mark the successfully matched checkpoints and remove them from the matching process. Then, iterate through all points according to step P2, counting the number of successful matches. When 80% of the points are successfully matched, the matching is considered complete. and It is the same location;

[0071] P4. Extract separately and The timestamp of the image is used to locate the corresponding device location information.

[0072] In the description of the invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," "counterclockwise," "axial," "radial," and "circumferential" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention.

[0073] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of the invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0074] In this invention, unless otherwise expressly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection, an electrical connection, or a connection that allows communication between them; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components, unless otherwise expressly limited. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0075] The invention has been described above with reference to the accompanying drawings. Obviously, the specific implementation of the invention is not limited to the above-described manner. Any improvements made using the inventive concept and technical solution, or direct application to other situations without modification, are all within the scope of protection of the invention.

Claims

1. A method of evaluating positioning accuracy of downhole equipment, characterized by, Comprise the following steps: S1. Camera installation and calibration mark design and arrangement: the camera (2) is installed in the underground equipment (1), and the calibration mark (3) is pasted in the mine roadway at a predetermined interval or in a random manner. The calibration mark (3) pattern is composed of a two-dimensional code located in the middle and a black and white checkerboard pattern located on the periphery; S2. Obtain image information: capture the pattern of the calibration mark (3) through the camera (2) installed on the underground equipment (1); S3. Calibration mark detection: two-dimensional code detection is performed on the calibration mark (3) to obtain the ID information and area information of the calibration mark (3); S4. Binary processing: the detected calibration mark (3) area is subjected to binary processing to accelerate the matching speed; S5. Extracting checkpoint: using preset convolution kernel to extract checkpoint, obtaining a group of checkpoints corresponding to the checkpoint mark (3) ; S6. Pose matching: each check mark (3) gets a set of continuous detection values, each detection value consisting of two-dimensional code information and check point information, denoted as , for the same check mark (3) region, a set of continuous detection results is obtained , is the number of images in which the check mark (3) appears at this position; in the mapping positioning and real-time positioning processes, a series of check mark (3) detection results are obtained respectively, the results are matched one by one, and the corresponding poses of the mapping and the corresponding poses of the real-time positioning at the same position are obtained. S7. Precision evaluation: for each set of consecutive matching results obtained at the check mark (3), for a certain check mark (3), corresponding to a set of consecutive images, each frame of image corresponds to a set of mutually matched mapping is the corresponding pose and real-time positioning is the corresponding pose , for the set of matches, the position error is calculated as, that is, the absolute trajectory error between the mapping positioning result and the real-time positioning result is calculated: ， wherein is a confidence coefficient, whose value is equal to the percentage of the number of successful checks of the image points of the frame, which allows real-time evaluation of the positioning accuracy of the downhole device (1).

2. The method of claim 1, wherein, The step of the check mark (3) detection is: firstly, two-dimensional code detection is performed to obtain pattern ID information and contour position of the two-dimensional code in the image; then, according to the fixed proportional relationship between the two-dimensional code and the whole check mark (3), the area information and the size information of the check mark (3) in the image are calculated, the area information When or or or , it is indicated that the area of the check mark (3) exceeds the whole image range, and the result of this time is abandoned, wherein id is the check mark id, is the horizontal coordinate of the two-dimensional code center point, is the vertical coordinate of the two-dimensional code center point, is the width of the pattern mark area, is the height of the pattern mark area, is the width of the image, is the height of the image.

3. The method of claim 1, wherein, The binary processing step is to traverse each pixel of the pattern area. When the pixel value in R, G, B three channels is greater than 100, the corresponding pixel value is set to 1, otherwise it is set to 0, generating a binary matrix composed of 0 and 1.

4. The method of claim 1, wherein, The step of extracting the check points is: using a 5*5 convolution kernel to traverse the whole check mark (3) detected, when the convolution kernel is completely matched with the pixel value of the region, marking the center point of the region as a check point wherein n is the number of check points, represents the pixel point coordinates.

5. The method of claim 4, wherein, The average positioning error at the pattern mark when n frames of corresponding images exist at the pattern mark during the precision evaluation is: ; and the average positioning error at the corresponding pattern mark is obtained based on the precision evaluation step.

6. The method of claim 5, wherein, The specific steps of the pose matching are: P1. Take out a detection result in all real-time positioning corresponding poses , traverse all mapping corresponding poses to find a result , meet and corresponding id are equal; P2. For each checkpoint in the set of checkpoints , find a matching value in , if one exists, such that the following relationship holds: the match is considered successful. - 2 <= + 2； - 2 <= + 2, wherein denotes the error allowance range; P3. Mark the successfully matched check points and remove them from the matching process. According to step P2, traverse all the points and calculate the number of successfully matched points. When 80% of the points are successfully matched, it is considered that and are the same position; P4. Extracting and the timestamp of the image in which it is located and finding the corresponding device location information from the timestamp.

Images