A high-precision large-range pose measurement method and system based on an absolute space coding calibration board

Abstract

The application discloses a high-precision large-range pose measurement method and system based on an absolute space coding calibration board, and relates to the technical field of visual measurement. The application discloses a high-precision large-range pose measurement method and system based on an absolute space coding calibration board, and relates to the technical field of visual measurement. The application discloses a high-precision large-range pose measurement method and system based on an absolute space coding calibration board, and relates to the technical field of visual measurement. The application discloses a high-precision large-range pose measurement method and system based on an absolute space coding calibration board, and relates to the technical field of visual measurement. The application discloses a high-precision large-range pose measurement method and system based on an absolute space coding calibration board, and relates to the technical field of visual measurement. The application discloses a high-precision large-range pose measurement method and system based on an absolute space coding calibration board, and relates to the technical field of visual measurement.

Description

Technical Field

[0001] This invention relates to the field of visual measurement technology, and in particular to a high-precision, large-range pose measurement method and system based on an absolute spatial coding calibration board. Background Technology

[0002] Visual measurement technology, with its advantages of low cost, non-contact operation, and flexibility, has gradually become a research hotspot for improving the accuracy of pose error compensation in multi-axis equipment and robot end effectors. It is widely used in industrial fields such as object motion posture tracking, robotic arm end effector guidance, and multi-axis platform kinematic calibration. Visual measurement technology provides image information of feature points on the calibration board surface to measure pose, thereby identifying the deviation between the actual motion model and the theoretical model of the mechanical system and completing the motion error compensation. Clearly, visual calibration is a crucial foundation for subsequent high-precision operations.

[0003] Traditional visual calibration methods rely heavily on measurements using checkerboard or dot matrix calibration boards. While they offer high accuracy in small-scale calibration, they have significant limitations when it comes to large-scale, high-precision calibration. For example, checkerboard calibration boards are sensitive to shooting angles and lighting conditions, and the acquired images must completely cover the entire checkerboard to provide sufficient geometric constraints for solving pose, thus limiting their application in precise attitude calibration under large-scale motion.

[0004] When visual measurement technology is used for motion platform calibration, high precision and large measurement range are often mutually exclusive. Large calibration boards offer wide coverage but reduce the density of feature points per unit area, leading to low local accuracy. High-resolution, small-field-of-view measurements require reducing the size of the calibration board, which in turn reduces the measurement range. Furthermore, planar calibration boards lack depth information in the depth direction, and limited geometric constraints can easily lead to low accuracy in extrinsic parameter calculations, especially when the calibration board's viewing angle is limited or its tilt is significant. When the minimum coding unit of the calibration board is too large, tilting can cause the coding unit to exceed the depth of field. Expanding the field of view can increase the number of image pixels but increases the image processing burden, while reducing the image pixel precision leads to an increase in the final kinematic calibration error.

[0005] Therefore, current visual measurement methods struggle to meet both the requirements of large-scale and high-precision calibration, and there is an urgent need to provide a method that can improve the geometric and range limitations of planar calibration plates to achieve high-precision, large-scale pose measurement. Summary of the Invention

[0006] This invention provides a high-precision, large-range pose measurement method and system based on an absolute spatial coding calibration board, which solves the technical problem that existing visual measurement methods cannot simultaneously meet the calibration requirements of large range and high precision.

[0007] The application provides a high-precision large-range pose measurement method based on an absolute space coding calibration board.

[0008] The target minimum identification unit is photographed by using the target calibration board to determine target mark circle center pixel coordinates, a target grating center point pixel coordinate and a target code center point pixel coordinate.

[0009] According to the absolute phase distribution of the two-dimensional sinusoidal grating pattern to which the target minimum identification unit belongs, the target grating center point pixel coordinate and the target code center point pixel coordinate, the actual physical distance between the corresponding grating center point and the same line-of-sight point of the code center point is determined.

[0010] The code recognition matrix of the target minimum identification unit is matched with mark circle center world coordinates, combined with the target mark circle center pixel coordinates and the camera intrinsic parameter matrix, to perform pose solving, so that the initial pose is determined.

[0011] Based on the target code center point pixel coordinate, the actual physical distance and the grating center point world coordinate of the target minimum identification unit, the initial pose is locally corrected in depth information, and the target pose is output.

[0012] Optionally, the target minimum identification unit is photographed by using the target calibration board to determine target mark circle center pixel coordinates, a target grating center point pixel coordinate and a target code center point pixel coordinate, comprising:

[0013] After the target minimum identification unit is extracted from the target calibration board image, corner point recognition is performed to determine actual two-dimensional code corner point coordinates.

[0014] Based on the actual two-dimensional code corner point coordinates, a minimum mark identification area containing a circular mark point group is circled in the target minimum identification unit.

[0015] According to the code recognition matrix of the target minimum identification unit, searching is performed in a preset calibration board code table to determine designed two-dimensional code corner point coordinates.

[0016] Based on the transformation relationship between the actual two-dimensional code corner point coordinates and the associated designed two-dimensional code corner point coordinates, perspective transformation is performed on the target calibration board image to determine a transformed image.

[0017] extracting initial mark circle center pixel coordinates of a circular mark point group in each of the minimum mark recognition regions and initial raster center point pixel coordinates from the transformed image;

[0018] performing geometric center fitting using the initial mark circle center pixel coordinates to determine initial encoding center point pixel coordinates of an encoding center point of the target minimum recognition unit;

[0019] inverse transforming the initial mark circle center pixel coordinates, the initial raster center point pixel coordinates and the initial encoding center point pixel coordinates to the target calibration board image to determine target mark circle center pixel coordinates, target raster center point pixel coordinates and target encoding center point pixel coordinates.

[0020] Optionally, the determining of the actual physical distance between the same line-of-sight point of the corresponding raster center point and the encoding center point according to the absolute phase distribution of the two-dimensional sinusoidal grating pattern to which the target minimum recognition unit belongs, the target raster center point pixel coordinates and the target encoding center point pixel coordinates comprises:

[0021] cropping a raster image of a region where the two-dimensional sinusoidal grating pattern of the target minimum recognition unit is located from the target calibration board image;

[0022] sequentially performing two-dimensional Fourier transform, band-pass filtering, two-dimensional inverse Fourier transform and phase unwrapping on the raster image to determine the absolute phase distribution;

[0023] determining a first absolute phase of the same line-of-sight point associated with the corresponding encoding center point from the absolute phase distribution according to the encoding center point pixel coordinates;

[0024] determining a second absolute phase of the corresponding raster center point from the absolute phase distribution according to the target raster center point pixel coordinates;

[0025] determining a pixel distance using a phase difference between the first absolute phase and the second absolute phase;

[0026] determining the actual physical distance between the same line-of-sight point and the raster center point according to the pixel distance and a physical distance corresponding to a unit pixel distance.

[0027] Optionally, the matching of the mark circle world coordinates through the encoding recognition matrix of the target minimum recognition unit, the pose solving combining the target mark circle pixel coordinates and the camera intrinsic parameter matrix to determine the initial pose comprises:

[0028] searching in a preset calibration board encoding table based on the encoding recognition matrix of the target minimum recognition unit to match the mark circle world coordinates;

[0029] Adopt the mark circle center world coordinate, the target mark circle center pixel coordinate and the camera internal parameter matrix, it is based on the linear transformation of the pinhole imaging, determine the homography matrix;

[0030] The initial rotation matrix and the initial translation matrix are decomposed from the homography matrix as the initial pose.

[0031] Optionally, the target pose is output by performing depth information local correction on the initial pose based on the target encoding center point pixel coordinate, the actual physical distance and the raster center point world coordinate of the target minimum identification unit, comprising:

[0032] The encoding identification matrix of the target minimum identification unit is searched in a preset calibration board encoding table based on the raster center point world coordinate, and the raster center point world coordinate is matched.

[0033] The raster center point world coordinate and the actual physical distance are summed to determine the isosight point world coordinate of the isosight point.

[0034] The depth information is determined by performing nonlinear least squares solution on the projection error target function according to the target encoding center point pixel coordinate, the isosight point world coordinate and the initial pose.

[0035] The initial pose is updated based on the depth information to determine the target pose.

[0036] Optionally, the determination process of the target calibration board image comprises:

[0037] The calibration board image of the double-layer transparent absolute space encoding calibration board is preprocessed to output the target calibration board image.

[0038] The second aspect of the application provides a high-precision large-range pose measurement system based on an absolute space encoding calibration board, which relates to a double-layer transparent absolute space encoding calibration board, the upper surface of the double-layer transparent absolute space encoding calibration board is arrayed with Aruco two-dimensional codes, and the lower surface is arrayed with two-dimensional sinusoidal grating patterns, circular marker points are equidistantly arranged between adjacent Aruco two-dimensional codes, the Aruco two-dimensional codes of the minimum identification unit form a unique encoding identification matrix, and the encoding center point of the minimum identification unit corresponds to the raster center point of the two-dimensional sinusoidal grating pattern one by one; the system comprises:

[0039] A pixel coordinate extraction module is configured to determine target mark circle center pixel coordinates, target raster center point pixel coordinates and target encoding center point pixel coordinates of a target minimum identification unit by using a target calibration board image.

[0040] a physical distance determination module configured to determine an actual physical distance between a same line-of-sight point of a grating center point and an encoding center point according to an absolute phase distribution of a two-dimensional sinusoidal grating pattern to which the target minimum identification unit belongs, the target grating center point pixel coordinate and the target encoding center point pixel coordinate;

[0041] a pose solving module configured to match a mark circle center world coordinate through an encoding identification matrix of the target minimum identification unit, and to determine an initial pose by combining the target mark circle center pixel coordinate and an intrinsic parameter matrix of a camera;

[0042] a pose correction module configured to perform a depth information local correction on the initial pose based on the target encoding center point pixel coordinate, the actual physical distance and the grating center point world coordinate of the target minimum identification unit, and to output a target pose.

[0043] A computer device provided in the third aspect of the present application comprises a memory and a processor, the memory stores a computer program, and the computer program is executed by the processor to make the processor execute the steps of the high-precision large-range pose measurement method based on the absolute space encoding calibration board according to any one of the above aspects.

[0044] A computer readable storage medium provided in the fourth aspect of the present application stores a computer program, and the computer program is executed to implement the high-precision large-range pose measurement method based on the absolute space encoding calibration board according to any one of the above aspects.

[0045] A computer program product provided in the fifth aspect of the present application comprises computer programs / instructions, and the computer programs / instructions are executed by a processor to implement the high-precision large-range pose measurement method based on the absolute space encoding calibration board according to any one of the above aspects.

[0046] It can be seen from the above technical solutions that the present application has the following advantages:

[0047] The above-described solution of the present invention provides a high-precision, large-range pose measurement method based on an absolute spatial coding calibration board, involving a double-layer transparent absolute spatial coding calibration board, comprising: using the target calibration board to capture images to determine the target marker center pixel coordinates, target grating center point pixel coordinates, and target coding center point pixel coordinates of the target minimum recognition unit; determining the actual physical distance between the corresponding grating center point and the coding center point co-line point based on the absolute phase distribution of the two-dimensional sinusoidal grating pattern to which the target minimum recognition unit belongs, the target grating center point pixel coordinates, and the target coding center point pixel coordinates; matching the world coordinates of the marker center point through the coding recognition matrix of the target minimum recognition unit, and performing pose calculation by combining the target marker center point pixel coordinates and the camera intrinsic parameter matrix to determine the initial pose; and performing local depth information correction on the initial pose based on the target coding center point pixel coordinates, the actual physical distance, and the grating center point world coordinates of the target minimum recognition unit to output the target pose. The above scheme is based on a designed double-layer transparent absolute spatial coding calibration board for pose calculation. On the one hand, it reduces the smallest identifiable unit of the calibration board, solving the problem that traditional calibration methods require capturing the complete calibration board coding, and enabling XY measurement over a wider range and measurement and calibration under large angle changes. On the other hand, it increases the spatial coding dimension of the coding unit, and provides more accurate depth measurement data for calibration through the spatial distribution of calibration points, which helps to achieve large-scale, high-precision spatial calibration. Attached Figure Description

[0048] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 This is a flowchart illustrating the steps of a high-precision, large-range pose measurement method based on an absolute spatial coding calibration board, as provided in Embodiment 1 of the present invention.

[0050] Figure 2 A partial view of the absolute spatial coding calibration board provided in Embodiment 1 of the present invention;

[0051] Figure 3 A camera image provided in Embodiment 1 of the present invention;

[0052] Figure 4 A schematic diagram of the smallest identification unit provided in Embodiment 1 of the present invention;

[0053] Figure 5 The two-dimensional sinusoidal grating pattern provided in Embodiment 1 of the present invention;

[0054] Figure 6 A schematic diagram of absolute spatial coding pose measurement provided in Embodiment 1 of the present invention;

[0055] Figure 7 A flowchart of a high-precision, large-range pose measurement method based on an absolute spatial coding calibration board provided in Embodiment 1 of the present invention;

[0056] Figure 8 This is a structural block diagram of a high-precision, large-range pose measurement system based on an absolute spatial coding calibration board, provided in Embodiment 2 of the present invention. Detailed Implementation

[0057] This invention provides a high-precision, large-range pose measurement method and system based on an absolute spatial coding calibration board, which solves the technical problem that existing visual measurement methods cannot simultaneously meet the calibration requirements of large range and high precision.

[0058] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0059] Please see Figure 1 , Figure 1 The flowchart illustrates the steps of a high-precision, large-range pose measurement method based on an absolute spatial coding calibration board, as provided in Embodiment 1 of the present invention.

[0060] This embodiment provides a high-precision, large-range pose measurement method based on an absolute spatial coding calibration board, involving a double-layer transparent absolute spatial coding calibration board, and the method includes steps 101 to 104.

[0061] It should be noted that this embodiment studies and designs a double-layer transparent absolute spatial coding calibration board, such as... Figures 2 to 4 As shown, the upper surface of the double-layer transparent absolute spatial coding calibration board is arrayed with Aruco QR codes, and the lower surface is arrayed with a two-dimensional sinusoidal grating pattern. Circular markers are equidistantly spaced between adjacent Aruco QR codes, forming a unique coding and recognition matrix with the Aruco QR codes forming the smallest recognition unit. The center point of the smallest recognition unit corresponds one-to-one with the center point of the grating in the two-dimensional sinusoidal grating pattern. From the design concept of the calibration board:

[0062] 1) Regarding the design of the surface coding pattern and feature points on the calibration plate:

[0063] The core of calibrating extrinsic parameters in this embodiment lies in the design of an absolute calibration board, whose smallest identification unit carries uniquely coded information that can be decoded. By taking local photos of each smallest identification area, the absolute coordinates of these points on the calibration board can be determined.

[0064] This embodiment uses Aruco encoding to achieve rapid positioning and identification, which has the advantages of rapid detection, ease of use, strong anti-interference ability, and low cost; such as Figure 4 As shown, every four Aruco QR codes constitute a minimum identification unit (ID0 / ID1 / ID2). If the dictionary for generating Aruco QR codes is DICT_4X4_250, then 250 uniquely identified QR codes participate in the calibration board pattern encoding. Every four QR code ID values ​​form a 2*2 encoding and recognition matrix. For example... There are a total of 250 4 There are several combination methods, and more codes can be generated from the dictionary according to actual needs; in short, among the absolute coding patterns on the entire calibration board surface, the smallest identification unit that conforms to the absolute coding template is unique. Therefore, as long as any part of the coding pattern on the calibration board that is larger than the smallest identification unit is photographed, the absolute position of that part of the coding pattern in the original image can be determined. Thus, the calibration board can use the arrangement of QR codes on the upper surface to achieve absolute coding to support a wide range of calibration applications.

[0065] After the identification coding design and fast absolute position acquisition, camera extrinsic parameter calibration is required. In this embodiment, circular patterns are used to construct circular marker points as calibration feature points. The center of the circular pattern is less affected by lens distortion (such as radial distortion and tangential distortion) in the image. Even if the circle is stretched into an ellipse due to distortion, its center point can still be accurately calculated by geometric fitting (such as the least squares method). Moreover, the sub-pixel level detection accuracy of the circular marker points is higher. In contrast, the corner points of the traditional checkerboard are prone to significant increases in positioning deviation due to distortion.

[0066] In practical implementation, each minimum recognition unit can carry multiple circular marker point groups. A set of circular marker point groups is set between any two adjacent Aruco QR codes. Each circular marker point group includes one or more evenly distributed circular marker points. For example, the number of circular marker point groups is four, and each set of circular marker point groups includes four circular marker points. That is, each minimum recognition unit carries a total of 16 circular marker points, which are used as feature points, i.e. target points, to solve the camera extrinsic parameters.

[0067] 2) Regarding the design of the two-dimensional sinusoidal grating pattern on the lower surface of the calibration board:

[0068] This implementation example Figure 5 The two-dimensional sinusoidal grating pattern shown can be used to generate a grayscale image using the following formula:

[0069]

[0070] wherein, is a gray scale image, is a luminance direct current component, is an x-direction fringe frequency, is an x-direction pixel coordinate, is a phase, is a y-direction fringe frequency, is a y-direction pixel coordinate;

[0071] The two-dimensional sinusoidal grating pattern can simultaneously obtain phase information in x and y directions, since the information in two directions has been integrated in the same pattern, the phase data in x and y directions can be separated without rotating or moving the grating, thereby improving the measurement efficiency; by analyzing the phase distribution of the fringes in the image, a mapping relationship between the phase difference and the physical distance is established, the phase difference can be converted into the physical distance and then the actual world coordinates of the target point are obtained, combined with the circular mark points on the upper surface, the three-dimensional distribution of the calibration points is realized, which is helpful to construct more accurate spatial geometric constraints, thereby improving the calibration accuracy.

[0072] 3) Overall design of the calibration board:

[0073] In specific implementation, in order to realize the transparent effect, the calibration board can be prepared by using a glass plate, the thickness of the calibration board is known when it is processed, on the upper surface of the calibration board, the Aruco two-dimensional code is used for the absolute positioning of the calibration board, and the circular mark points are used for the rapid solving of the pose of the calibration board, on the lower surface of the calibration board, the two-dimensional sinusoidal grating pattern generated based on a mathematical model is used to provide high-precision target point information, the world coordinate of the center of the feature point circular pattern on the upper surface has been determined during processing, therefore, the points on the upper surface can form three-dimensional observation data together with the target points on the lower surface to participate in the solving of the pose of the calibration board, and the points on the upper surface can also be used alone to solve the pose of the calibration board as plane target points, to preliminarily verify whether the solving result of the pose of the calibration board is correct, the center point of each group of two-dimensional sinusoidal grating patterns on the lower surface (i.e. the center point of the grating) corresponds to the center point of the line connecting the center of each group of circular codes on the upper surface (i.e. the center point of the code), and the world coordinates of the center points are also determined during processing, under the camera view, the two planes on the upper surface and the lower surface are displayed on the same image, and the ideal case is as shown in Figure 3 .

[0074] Step 101, using the target calibration board to shoot an image to determine the target mark center pixel coordinates, the target grating center point pixel coordinates and the target code center point pixel coordinates of the target minimum recognition unit.

[0075] The calibration board image refers to the image obtained by shooting the double-layer transparent absolute space coding calibration board.

[0076] Minimum recognition unit refers to the minimum recognition area containing unique coding information conforming to the absolute coding template.

[0077] Marked circle center pixel coordinates refer to the coordinates of the center of the circular mark in the circular mark group in the pixel coordinate system.

[0078] Grating center point pixel coordinates refer to the coordinates of the center point of the two-dimensional sinusoidal grating pattern in the pixel coordinate system.

[0079] Encoding center point pixel coordinates refer to the coordinates of the center point of the upper surface of the two-dimensional code encoding in the pixel coordinate system.

[0080] The determination process of the target calibration board image includes:

[0081] The calibration board image of the double-layer transparent absolute space coding calibration board is preprocessed, and the target calibration board image is output.

[0082] It should be noted that after the calibration board image is obtained by shooting the double-layer transparent absolute space coding calibration board through the camera, preprocessing can be performed first, which can include steps such as binarization, denoising, grayscale correction, and geometric correction (eliminating lens distortion), and then the final target calibration board image for calibration is output. In order to solve the problem of the need for complete shooting of the traditional calibration board, the calibration board image can correspond to the image obtained under the condition of shooting only part of the calibration board, as long as at least one minimum recognition unit is contained in the calibration board image, the calibration can be performed. Therefore, after obtaining the target calibration board image, the minimum recognition unit is identified and extracted from the target calibration board image, one of the extracted minimum recognition units is selected as the minimum recognition unit for calibration, that is, the target minimum recognition unit, and then the pixel coordinates of the mark circle center, grating center point and encoding center point of the target minimum recognition unit are identified, so as to determine the corresponding target mark circle center pixel coordinates, target grating center point pixel coordinates and target encoding center point pixel coordinates.

[0083] In one specific embodiment of the present embodiment, step 101 includes the following sub-steps:

[0084] After the target minimum recognition unit is extracted from the target calibration board image, the corner point recognition is performed, and the actual two-dimensional code corner point coordinates are determined;

[0085] Based on the actual two-dimensional code corner point coordinates, the minimum mark recognition area containing the circular mark group is circled in the target minimum recognition unit;

[0086] According to the encoding recognition matrix of the target minimum recognition unit, the design two-dimensional code corner point coordinates are determined by searching in the preset calibration board encoding table.

[0087] Based on the transformation relationship between the actual two-dimensional code corner point coordinates and the associated design two-dimensional code corner point coordinates, the perspective transformation is performed on the target calibration board image to determine a transformed image;

[0088] The initial marker center pixel coordinates of the circular marker point group in each minimum marker recognition region and the initial grating center point pixel coordinates are extracted from the transformed image;

[0089] The initial encoding center point pixel coordinates of the encoding center point of the target minimum recognition unit are determined by performing geometric center fitting on the initial marker center pixel coordinates;

[0090] The target marker center pixel coordinates, the target grating center point pixel coordinates, and the target encoding center point pixel coordinates are determined by inversely transforming the initial marker center pixel coordinates, the initial grating center point pixel coordinates, and the initial encoding center point pixel coordinates to the target calibration board image.

[0091] Actual two-dimensional code corner point coordinates refer to the coordinates of the corner points of the Aruco two-dimensional code in the actual captured calibration board image.

[0092] Minimum marker recognition region refers to the smallest region capable of recognizing the coordinates of the circular marker point group.

[0093] Encoding recognition matrix refers to the matrix composed of the two-dimensional code ID values of the multiple minimum recognition units that constitute the minimum recognition unit.

[0094] Calibration board encoding table refers to the table containing the distribution of Aruco two-dimensional code ID values of the entire double-layer transparent absolute space encoding calibration board.

[0095] Design two-dimensional code corner point coordinates refer to the coordinates of the corner points of the Aruco two-dimensional code that have been determined according to design drawings or other methods during the processing of the double-layer transparent absolute space encoding calibration board.

[0096] Transformed image refers to the image obtained after perspective transformation.

[0097] It should be noted that after the target minimum recognition unit is determined from the target calibration board image, corner point recognition is performed on the target minimum recognition unit. Since the minimum recognition unit is surrounded by multiple Aruco two-dimensional codes, the minimum recognition unit constitutes the corner point with the Aruco two-dimensional code, thereby determining the actual two-dimensional code corner point coordinates, and defining the minimum recognition unit region (ROI) for solving the pose, and then extracting the marker center pixel coordinates of the entire minimum recognition unit in the target calibration board image with each minimum recognition unit region as the range, for example, Figure 3As shown, the minimum recognition unit region can be formed by two Aruco two-dimensional codes, and the minimum recognition unit region can only need to meet the requirement of facilitating extraction of the pixel coordinates of the center of the marker, and preferably, the region where the two-dimensional sinusoidal grating pattern is located is not included in the minimum recognition unit region to avoid affecting the extraction effect of the circular marker point.

[0098] In order to eliminate the distortion caused by the shooting angle, the target calibration plate image is projected onto the design plane of the calibration plate by perspective transformation to regularize the image, so as to simplify the feature extraction process. In the perspective transformation process, the coding recognition matrix composed of the ID values of the Aruco two-dimensional codes on the upper surface of the target minimum recognition unit is identified, and the row and column positions of the coding recognition matrix in the calibration plate coding table of the absolute calibration plate are searched. The corresponding design two-dimensional code corner point coordinates can be determined in combination with the design of the calibration plate. The one-to-one correspondence between the actual two-dimensional code corner point coordinates and the design two-dimensional code corner point coordinates can be used to establish a perspective transformation matrix, so that the perspective transformation matrix is used to perform perspective transformation on the target calibration plate image to obtain a transformed image.

[0099] In the transformed image, the pixel coordinates of the center of the marker of each group of circular marker points and the pixel coordinates of the center of the grating of the two-dimensional sinusoidal grating pattern are extracted, so as to obtain the initial pixel coordinates of the center of the marker and the initial pixel coordinates of the center of the grating. Then, the pixel coordinates of the center of the marker surrounded by each circular marker point are calculated by connecting the initial pixel coordinates of the center of the marker of the circular marker point, so as to determine the initial pixel coordinates of the center of the coding. The feature points extracted are restored to the pixel coordinates of the original image by inverse transformation, and the target pixel coordinates obtained can provide high-precision target point information.

[0100] Step 102, determining the actual physical distance between the corresponding grating center point and the same line-of-sight point between the coding center point according to the absolute phase distribution of the two-dimensional sinusoidal grating pattern to which the target minimum recognition unit belongs, the target grating center point pixel coordinates and the target coding center point pixel coordinates.

[0101] The absolute phase distribution refers to continuous and unique phase data reflecting the spatial distribution of the grating fringe, which can be understood with reference to the prior art.

[0102] The same line-of-sight point refers to, in the pinhole imaging camera model, if two three-dimensional points are located on the same camera light line, then their projections on the image will fall on the same pixel point, as shown in Figure 6 C point is the "same line-of-sight point" of A point (coding center point) on the camera line of sight.

[0103] The actual physical distance refers to the actual distance between two points in the double-layer transparent absolute space coding calibration plate.

[0104] It should be noted that the embodiment extracts the fringe phase information of the two-dimensional sinusoidal grating pattern to which the target minimum recognition unit in the target calibration plate image belongs to determine the absolute phase distribution, and then combines the absolute phase distribution with the calibration plate, and converts the actual physical distance between the grating center point of the target minimum recognition unit and the same line-of-sight point of the code center point according to the two-dimensional phase data by the analytic geometry method, so as to facilitate the subsequent determination of the world coordinate.

[0105] In one specific embodiment of the present embodiment, step 102 comprises the following sub-steps:

[0106] cropping the grating image of the region where the two-dimensional sinusoidal grating pattern of the target minimum recognition unit is located from the target calibration plate image;

[0107] performing two-dimensional Fourier transform, band-pass filtering, two-dimensional inverse Fourier transform and phase unwrapping on the grating image in sequence to determine the absolute phase distribution;

[0108] determining the first absolute phase of the same line-of-sight point associated with the code center point from the absolute phase distribution according to the code center point pixel coordinates;

[0109] determining the second absolute phase of the grating center point from the absolute phase distribution according to the target grating center point pixel coordinates;

[0110] determining the corresponding pixel distance by using the phase difference between the first absolute phase and the second absolute phase;

[0111] determining the actual physical distance between the same line-of-sight point and the grating center point according to the pixel distance and the physical distance corresponding to the unit pixel distance.

[0112] The grating image refers to the image containing the region where the two-dimensional sinusoidal grating pattern is located.

[0113] The pixel distance refers to the distance between two pixel points in the image.

[0114] It should be noted that first, the grating image of the region where the two-dimensional sinusoidal grating pattern of the target minimum recognition unit is located is cropped from the target calibration plate image. Since the actual image is obtained by shooting through an industrial camera, the background and the non-uniformity of the modulation amplitude may be introduced, and the following model is established:

[0115]

[0116] In the formula, is the pixel coordinate, is the background light (low frequency component), and are the local modulation amplitudes of the x and y direction fringes respectively, and x, y direction phase components (containing actual measurement information) to be extracted;

[0117] The fringe term can be written as:

[0118]

[0119]

[0120] wherein, is the phase angle, is the imaginary unit;

[0121] Then, a two-dimensional Fourier transform is performed, defined as:

[0122]

[0123] wherein, is the frequency variable, is the frequency spectrum; the inverse transform is:

[0124]

[0125] According to the foregoing model, after Fourier transform, the frequency spectrum will have:

[0126] 1) DC component: from , mainly concentrated near ;

[0127] 2) x direction carrier component: one at positive frequency corresponding to ; one at negative frequency corresponding to its conjugate component;

[0128] 3) y direction carrier component: one at corresponding to ; one at corresponding to its conjugate component;

[0129] Actual images will have scaling and distortion during printing and shooting, resulting in the ideal set and the actual frequency appearing in the image may not be completely consistent; the two-dimensional Fourier transform can be performed on the captured image to complete the calculation of the frequency spectrum extraction; using the frequency spectrum amplitude , local maximum values (peaks) are searched in the plane; in the ideal case, the position of the peak should be close to:

[0130]

[0131] Then for a certain direction (e.g. x direction), define an x direction band-pass filter which satisfies in frequency domain:

[0132]

[0133] wherein, is the filter radius, selected to enclose the whole main peak but exclude other interfering components; similarly, for y direction, define a y direction band-pass filter:

[0134]

[0135] After using the filter, get the filtered frequency spectrum, including x direction filtered frequency spectrum and y direction filtered frequency spectrum :

[0136]

[0137] In this way, other frequency components are filtered out, and only the carrier part of the corresponding direction is reserved. After shifting the filtered positive first order spectrum in x and y directions from , to the origin, perform two-dimensional inverse Fourier transform to get:

[0138]

[0139] wherein, is the x direction complex fringe signal, is the two-dimensional inverse Fourier transform, is the y direction complex fringe signal;

[0140] From equation 11, get the phase:

[0141]

[0142] wherein, is the phase of x direction fringe, is the phase of y direction fringe, is the imaginary part, is the real part;

[0143] Since the obtained phase value is within , there is a jump problem, and phase unwrapping is needed to get the continuous phase function, i.e. absolute phase distribution.

[0144] According to the actual size of the printed or photographed pattern, calibrate the scale factor and The phase difference can be converted into a physical distance; for the same sight line point, the pixel coordinates of the C point are the same as the pixel coordinates of the A point in the pinhole imaging model, therefore, first, the first absolute phase of the corresponding sight line point (C point) is determined from the absolute phase distribution according to the encoding center point pixel coordinates of the A point, and then the second absolute phase of the corresponding grating center point (B point) is determined from the absolute phase distribution according to the target grating center point pixel coordinates The phase difference is constructed according to the first absolute phase and the second absolute phase, the offset is calculated through the phase difference, and the actual physical distance from the C point to the grating center point (B point) is calculated, the process of which can be referred to as the following formula:

[0145]

[0146]

[0147] In the formula, That is, the x-direction phase difference relative to the grating center point; That is, the y-direction phase difference relative to the grating center point, is the x-direction pixel distance relative to the grating center point, is the y-direction pixel distance relative to the grating center point; and is the physical distance corresponding to the unit pixel distance in the x-direction and the y-direction, and That is, the actual physical distance from the x-direction and y-direction pixel point to the grating center point.

[0148] Step 103, match the mark circle center world coordinates through the encoding recognition matrix of the target minimum recognition unit, combine the target mark circle center pixel coordinates and the camera intrinsic parameter matrix to solve the pose, and determine the initial pose.

[0149] The mark circle center world coordinates refer to the coordinates of the circular mark point in the calibration board coordinate system (world coordinate system).

[0150] It should be noted that in the present embodiment, the world coordinates of the corresponding feature points on the double-layer transparent absolute space encoding calibration board are determined during processing, therefore, the world coordinates of the circular mark point center of the target minimum recognition unit in the calibration board coordinate system, that is, the mark circle center world coordinates, can be determined from the calibration board through the encoding recognition matrix of the target minimum recognition unit, and the mark circle center world coordinates, the target mark circle center pixel coordinates and the camera intrinsic parameter matrix are used to solve the initial pose according to the pinhole imaging principle, thereby determining the initial pose.

[0151] In one specific embodiment of the present embodiment, step 103 includes the following sub-steps:

[0152] The coding recognition matrix based on the smallest target recognition unit is searched in the preset calibration plate coding table to match the world coordinates of the center of the marked circle;

[0153] Using the world coordinates of the marker center, the pixel coordinates of the target marker center, and the camera intrinsic parameter matrix, a linear transformation is performed based on pinhole imaging to determine the homography matrix;

[0154] The initial rotation matrix and initial translation matrix are decomposed from the homography matrix to form the initial pose.

[0155] The homography matrix is ​​a matrix that represents the mapping relationship between the calibration plate coordinate system and the pixel coordinate system.

[0156] It should be noted that after searching the encoding recognition matrix of the target's smallest recognition unit in the encoding table of the absolute calibration board to determine the row and column positions of the encoding recognition matrix in the encoding table, the world coordinates of the mark center of the target's smallest recognition unit can be calculated and determined based on the actual physical spacing between each row and each column.

[0157] The known pinhole imaging model is as follows:

[0158]

[0159] In the formula, The camera intrinsic parameter matrix is ​​known. For the three-dimensional points in the calibration plate coordinate system, For transpose, For pixel coordinates, As a scale factor, The rotation matrix in the camera coordinate system. The translation vector in the camera coordinate system; the rotation matrix. column vector , The first rotation matrix Column vectors Matrix column index;

[0160] If the calibration plate coordinate system is established on the upper surface, then the corresponding circular marker points on the upper surface satisfy the following:

[0161]

[0162] In the formula, Let x be the x-coordinate of the calibration plate coordinate system on the upper surface of the calibration plate. Let be the y-coordinate of the calibration plate coordinate system on the upper surface of the calibration plate; at this point, the projection equation degenerates into:

[0163]

[0164] In the formula, The upper surface scale factor;

[0165] Introducing homogeneous coordinates of plane points Then it can be written as:

[0166]

[0167] Define homography matrix Then the points on the upper surface satisfy:

[0168]

[0169] Next, the coordinate correspondence of multiple circular marker points of the smallest target recognition unit on the surface of the calibration board is used. The direct linear transformation (DLT) method can be used to obtain it. ,set up (each) for The vector represents the first element of the homography matrix. (column vector), then we have:

[0170]

[0171]

[0172] In the formula, This is an intermediate variable representing the first column vector of the unnormalized rotation matrix in the camera coordinate system. This is an intermediate variable of the second column vector of the unnormalized rotation matrix in the camera coordinate system. For the intermediate variable of the unnormalized translation vector in the camera coordinate system; there exists a global scale factor. Make:

[0173]

[0174] make From the orthogonality requirement, we can obtain:

[0175]

[0176] Simultaneously, the following requirements are also required:

[0177]

[0178] Thus, the first two columns of the rotation matrix are obtained from the upper surface. With translation vector (in , Translation vector The translation vectors in the x, y, and z directions are used to obtain the initial pose, but at this point, since all points are located in... Up, There is uncertainty in the scale (and the scale related to Therefore, it is necessary to make a local correction of the depth information.

[0179] Step 104, based on the target encoding center point pixel coordinates, the actual physical distance and the raster center point world coordinates of the target minimum recognition unit, the initial pose is locally corrected by the depth information, and the target pose is output.

[0180] Step 104 includes the following sub-steps:

[0181] Based on the encoding recognition matrix of the target minimum recognition unit, search in the preset calibration board encoding table, match the raster center point world coordinates;

[0182] The sum operation of the raster center point world coordinates and the actual physical distance is used to determine the same line-of-sight point world coordinates of the same line-of-sight point;

[0183] According to the target encoding center point pixel coordinates, the same line-of-sight point world coordinates and the initial pose, the projection error target function is solved by nonlinear least squares to determine the depth information;

[0184] Based on the depth information, the initial pose is updated to determine the target pose.

[0185] The projection error target function refers to a function that constructs the projection error depending on the depth information.

[0186] The depth information refers to the information that can correct the pose from the depth angle.

[0187] It should be noted that the same line-of-sight point (C point) of the encoding center point of the target minimum recognition unit corresponds to the lower surface of the calibration board, so the projection of the C point on the lower surface of the image satisfies:

[0188]

[0189] Wherein, is the pixel coordinates of the same line-of-sight point, is the coordinates of the same line-of-sight point on the lower surface of the calibration board coordinate system, that is, the same line-of-sight point world coordinates, To calibrate the plate thickness; similar to the way the world coordinates of the center of the mark circle are determined, the world coordinates of the grating center point (point B) are searched for in the preset calibration plate encoding table based on the encoding recognition matrix of the target minimum recognition unit, and the actual physical distance of point C relative to point B has been extracted through the phase information in the foregoing step. The sum value operation of the world coordinates of the grating center point and the actual physical distance corresponding to the same line-of-sight point can determine the world coordinates of the same line-of-sight point; since point C is the same line-of-sight point of point A, according to the image projection relationship, the pixel coordinates of the two points are the same, so the pixel coordinates of the target encoding center point are the pixel coordinates of the same line-of-sight point;

[0190] Generally, due to the scale selection when decomposing , and , it can be considered that the upper surface has been determined and , but since the points on the plane cannot provide depth information, and , there is ambiguity in the "absolute scale", for this purpose, two unknown parameters can be introduced: : used to correct the scale of , that is , in theory, if the decomposition is correct, there should be , : the offset of the translation vector in the z direction (from the upper surface to the camera origin); therefore, the projection equation of the lower surface can be written as:

[0191]

[0192] In the formula, is the scale factor of the lower surface; let:

[0193]

[0194] then:

[0195]

[0196] wherein, is the depth correction term, and the plane transformation term only is unknown, and is also a to-be-determined parameter, then the constraint is given by using the known image pixel coordinates and the intrinsic matrix to solve and ; let the form of

[0197]

[0198]

[0199] wherein, is the equivalent focal length in x direction, is the equivalent focal length in y direction, is the x direction pixel coordinate of the principal point, is the y direction pixel coordinate of the principal point, , and is the decomposition of the planar transformation term along x, y and z directions, , and is the decomposition of the depth correction term along x, y and z directions; wherein is related to the known quantity , and contains the unknown quantity ;

[0200]

[0201] The homogeneous scale is obtained from equation 31 as:

[0202]

[0203] From equation 28, equation 31 and equation 32, we have:

[0204]

[0205] Equation 33 is a nonlinear constraint on and ( ), which defines the projection error (for and components) and the constructed projection error objective function is obtained by nonlinear least squares (Levenberg-Marquardt method) to find the parameters and , , is the weight coefficient to balance the projection error and the unit orthogonality constraint:

[0206]

[0207] After the above steps, the depth information and are determined, and finally the target rotation matrix and the target translation vector in the target pose of the absolute space coded calibration board can be obtained as:

[0208]

[0209] wherein the third column vector of the target rotation matrix .

[0210] For better illustration, refer to Figure 7 , the overall framework diagram of the embodiment one of the present application is shown:

[0211] 1) Information encoding: generate unique encoding identification matrix by using multiple aruco codes;

[0212] 2) Calibration board design: design double-layer transparent absolute space encoding calibration board according to multiple encoding identification matrix and two-dimensional sinusoidal grating pattern;

[0213] 3) Image acquisition: acquire calibration board shooting image by shooting double-layer calibration board through camera;

[0214] 4) Preprocessing: perform threshold processing such as binarization on calibration board shooting image;

[0215] 5) Feature extraction: perform coordinate calculation of minimum identification unit such as marker center pixel coordinate, grating center point pixel coordinate and encoding center point pixel coordinate;

[0216] 6) Information decoding: complete two-dimensional pattern phase information extraction of minimum identification unit and determine actual physical distance;

[0217] 7) Pose calculation: find absolute position of minimum identification unit on calibration board and solve pose according to encoding identification matrix of minimum identification unit.

[0218] It should be pointed out that only the general flow of high-precision large-range pose measurement method based on absolute space encoding calibration board is briefly described here, and the specific implementation process of each step can be understood by referring to the related contents in the foregoing embodiments, which will not be described here. It can be understood that the present application does not limit this.

[0219] In the embodiment of the present application, in the scheme of solving pose based on designed double-layer transparent absolute space encoding calibration board: 1) the minimum identifiable unit of the calibration board is reduced, only local unit encoding information (minimum identification unit) can determine the world coordinate system of the calibration board, solving the problem that the traditional calibration method needs to shoot the complete calibration board code, realizing larger range XY measurement and measurement and calibration under large angle change; 2) the spatial encoding dimension of the encoding unit is increased, and more accurate Z-axis measurement data is provided for calibration by the spatial distribution of the calibration points; 3) a high-precision pose solving method based on absolute calibration board is given; overall, based on the designed calibration board, the limitation of the traditional calibration board is broken through, realizing large-range, high-precision space calibration, and the structure is simple and easy to realize, providing an effective tool for precise kinematic modeling and error compensation of multi-axis platform.

[0220] Please refer to Figure 8 , Figure 8 A structural block diagram of a high-precision large-range pose measurement system based on an absolute space coding calibration board is provided for Embodiment Two of the present application.

[0221] The high-precision large-range pose measurement system based on an absolute space coding calibration board provided in this embodiment relates to a double-layer transparent absolute space coding calibration board, the upper surface of the double-layer transparent absolute space coding calibration board is arrayed with Aruco two-dimensional codes and the lower surface is arrayed with a two-dimensional sinusoidal grating pattern, a circular marker point group is equidistantly arranged between adjacent Aruco two-dimensional codes, the Aruco two-dimensional codes surrounding the minimum recognition unit form a unique coding recognition matrix, and the coding center point of the minimum recognition unit corresponds to the grating center point of the two-dimensional sinusoidal grating pattern one by one; the system comprises:

[0222] A pixel coordinate extraction module 801 is configured to determine target marker circle center pixel coordinates, a target grating center point pixel coordinate and a target coding center point pixel coordinate of a target minimum recognition unit by using a target calibration board to capture an image;

[0223] A physical distance determination module 802 is configured to determine the actual physical distance between the same line-of-sight points of the grating center point and the coding center point according to the absolute phase distribution of the two-dimensional sinusoidal grating pattern to which the target minimum recognition unit belongs, the target grating center point pixel coordinate and the target coding center point pixel coordinate;

[0224] A pose solution module 803 is configured to match marker circle center world coordinates through the coding recognition matrix of the target minimum recognition unit, and to determine an initial pose by combining the target marker circle center pixel coordinates and a camera intrinsic parameter matrix for pose solution;

[0225] A pose correction module 804 is configured to perform deep information local correction on the initial pose based on the target coding center point pixel coordinate, the actual physical distance and the grating center point world coordinates of the target minimum recognition unit, and to output a target pose.

[0226] Further, the pixel coordinate extraction module 801 is specifically configured to:

[0227] extract the target minimum recognition unit from the target calibration board captured image and perform corner point recognition to determine actual two-dimensional code corner point coordinates;

[0228] based on the actual two-dimensional code corner point coordinates, the minimum marker recognition area containing the circular marker point group is circled in the target minimum recognition unit;

[0229] the design two-dimensional code corner point coordinates are determined by searching in a preset calibration board coding table according to the coding recognition matrix of the target minimum recognition unit;

[0230] The perspective transformation is performed on the target calibration plate image based on the transformation relationship between the actual two-dimensional code corner point coordinates and the associated design two-dimensional code corner point coordinates;

[0231] The initial marker center pixel coordinates of the circular marker point group in each minimum marker recognition region and the initial grating center point pixel coordinates are extracted from the transformed image;

[0232] The initial encoding center point pixel coordinates of the encoding center point of the target minimum recognition unit are determined by geometric center fitting using the initial marker center pixel coordinates;

[0233] The target marker center pixel coordinates, the target grating center point pixel coordinates and the target encoding center point pixel coordinates are determined by inversely transforming the initial marker center pixel coordinates, the initial grating center point pixel coordinates and the initial encoding center point pixel coordinates to the target calibration plate image.

[0234] Further, the physical distance determination module 802 is specifically configured to:

[0235] The grating image of the region where the two-dimensional sinusoidal grating pattern of the target minimum recognition unit is located is cropped from the target calibration plate image;

[0236] The grating image is sequentially subjected to two-dimensional Fourier transform, band-pass filtering, two-dimensional inverse Fourier transform and phase unwrapping to determine the absolute phase distribution;

[0237] The first absolute phase of the corresponding boresight point associated with the encoding center point pixel coordinates is determined from the absolute phase distribution;

[0238] The second absolute phase of the corresponding grating center point is determined from the absolute phase distribution according to the target grating center point pixel coordinates;

[0239] The phase difference between the first absolute phase and the second absolute phase is used to determine the corresponding pixel distance;

[0240] The actual physical distance between the boresight point and the grating center point is determined according to the pixel distance and the physical distance corresponding to the unit pixel distance.

[0241] Further, the pose solution module 803 is specifically configured to:

[0242] The marker center world coordinates are matched by searching the encoding recognition matrix of the target minimum recognition unit in the preset calibration plate encoding table;

[0243] The homography matrix is determined by performing linear transformation based on pinhole imaging using the marker center world coordinates, the target marker center pixel coordinates and the camera intrinsic parameter matrix;

[0244] Decompose the initial rotation matrix and the initial translation matrix from the homography matrix as the initial pose.

[0245] Further, the pose solving module 803 is specifically configured to:

[0246] Search the encoding recognition matrix of the target minimum recognition unit in the preset calibration plate encoding table, and match the raster center point world coordinates;

[0247] Perform sum operation on the raster center point world coordinates and the actual physical distance to determine the same line-of-sight point world coordinates of the same line-of-sight point;

[0248] According to the target encoding center point pixel coordinates, the same line-of-sight point world coordinates and the initial pose, perform nonlinear least square solving on the projection error target function to determine the depth information;

[0249] Update the initial pose based on the depth information to determine the target pose.

[0250] Further, the pre-processing module is further included, and is configured to:

[0251] Pre-process the calibration plate shooting image of the double-layer transparent absolute space encoding calibration plate to output the target calibration plate shooting image.

[0252] The embodiment of the present application further provides a computer device, including a memory and a processor, and the memory stores a computer program; the computer program is executed by the processor, so that the processor executes the steps of the high-precision large-range pose measurement method based on the absolute space encoding calibration plate of any one of the above-mentioned embodiments.

[0253] The embodiment of the present application further provides a computer readable storage medium, which stores a computer program / instruction, and the computer program / instruction is executed by the processor to realize the steps of the high-precision large-range pose measurement method based on the absolute space encoding calibration plate of any one of the above-mentioned embodiments.

[0254] The embodiment of the present application further provides a computer program product, including a computer program / instruction, and the computer program / instruction is executed by the processor to realize the steps of the high-precision large-range pose measurement method based on the absolute space encoding calibration plate of any one of the above-mentioned embodiments.

[0255] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the system and the module described above can refer to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0256] In several embodiments provided in the present application, it should be understood that the disclosed system and method can be implemented in other manners. For example, the division of the system embodiments described above is merely illustrative, and the units can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interface, device or unit, and can be electrical, mechanical or in other forms.

[0257] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they can be located in one place or distributed on multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0258] In addition, the functional units in each embodiment of the present application can be integrated into a processing unit, or each unit can be physically present separately, or two or more units can be integrated into one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0259] The integrated unit, if realized in the form of a software functional unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application or the essential part or all or part of the technical solutions that contribute to the prior art can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0260] The above embodiments are merely used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features can be replaced by equivalent ones; and these modifications or replacements do not make the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A high-precision large-range pose measurement method based on an absolute space coding calibration board, characterized in that, The application relates to a double-layer transparent absolute space coding calibration plate, wherein the upper surface of the double-layer transparent absolute space coding calibration plate is arrayed with Aruco two-dimensional codes, the lower surface is arrayed with a two-dimensional sinusoidal grating pattern, circular mark points are equidistantly arranged between adjacent Aruco two-dimensional codes, the Aruco two-dimensional codes of a minimum recognition unit form a unique coding recognition matrix, and the coding center point of the minimum recognition unit corresponds to the grating center point of the two-dimensional sinusoidal grating pattern; and the method comprises the following steps: target mark circle center pixel coordinates, target grating center point pixel coordinates and target coding center point pixel coordinates of a target minimum recognition unit are determined by using a target calibration plate to shoot an image; an actual physical distance between a same line-of-sight point of a corresponding grating center point and a coding center point is determined according to an absolute phase distribution of a two-dimensional sinusoidal grating pattern to which the target minimum recognition unit belongs, the target grating center point pixel coordinates and the target coding center point pixel coordinates; an initial pose is determined by matching mark circle world coordinates of a coding recognition matrix of the target minimum recognition unit, combining the target mark circle pixel coordinates and a camera internal parameter matrix and performing pose solving; a target pose is output by performing local correction of depth information on the initial pose based on the target coding center point pixel coordinates, the actual physical distance and grating center point world coordinates of the target minimum recognition unit.

2. The high-precision large-range pose measurement method based on the absolute space coding calibration board according to claim 1, characterized in that, the target mark circle center pixel coordinates, the target grating center point pixel coordinates and the target coding center point pixel coordinates of the target minimum recognition unit are determined by using the target calibration plate to shoot the image, and the method comprises the following steps: actual two-dimensional code corner point coordinates are determined after a target minimum recognition unit is extracted from a target calibration plate image and corner point recognition is performed; a minimum mark recognition area containing a circular mark point group is circled based on the actual two-dimensional code corner point coordinates in the target minimum recognition unit; design two-dimensional code corner point coordinates are determined by searching a preset calibration plate coding table according to a coding recognition matrix of the target minimum recognition unit; a transformed image is determined by performing perspective transformation on the target calibration plate shooting image based on a transformation relationship between the actual two-dimensional code corner point coordinates and the associated design two-dimensional code corner point coordinates; initial mark circle center pixel coordinates and initial grating center point pixel coordinates of the circular mark point group in each minimum mark recognition area are extracted from the transformed image; initial coding center point pixel coordinates of the coding center point of the target minimum recognition unit are determined by performing geometric center fitting on each initial mark circle center pixel coordinate; the initial mark circle center pixel coordinates, the initial grating center point pixel coordinates and the initial coding center point pixel coordinates are inversely transformed to the target calibration plate shooting image to determine the target mark circle center pixel coordinates, the target grating center point pixel coordinates and the target coding center point pixel coordinates.

3. The high-precision large-range pose measurement method based on the absolute space coding calibration board according to claim 1, characterized in that, the actual physical distance between the same line-of-sight point of the corresponding grating center point and the coding center point is determined according to the absolute phase distribution of the two-dimensional sinusoidal grating pattern to which the target minimum recognition unit belongs, the target grating center point pixel coordinates and the target coding center point pixel coordinates, and the method comprises the following steps: cropping a grating image of a region where a two-dimensional sinusoidal grating pattern of the target minimum recognition unit is located from the image captured from the target calibration board; performing two-dimensional Fourier transform, band-pass filtering, two-dimensional inverse Fourier transform and phase unwrapping on the grating image in sequence to determine an absolute phase distribution; determining a first absolute phase of a boresight point associated with the encoding center point from the absolute phase distribution according to the encoding center point pixel coordinates; determining a second absolute phase of the grating center point from the absolute phase distribution according to the target grating center point pixel coordinates; determining a corresponding pixel distance by using a phase difference between the first absolute phase and the second absolute phase; determining an actual physical distance between the boresight point and the grating center point according to the pixel distance and a physical distance corresponding to a unit pixel distance.

4. The high-precision large-range pose measurement method based on the absolute space coding calibration board according to claim 1, characterized in that, The matching of the mark circle center world coordinates through the encoding recognition matrix of the target minimum recognition unit is combined with the target mark circle center pixel coordinates and the camera intrinsic parameter matrix to perform pose calculation to determine an initial pose, including: searching for the encoding recognition matrix of the target minimum recognition unit in a preset calibration board encoding table to match mark circle center world coordinates; performing linear transformation based on pinhole imaging by using the mark circle center world coordinates, the target mark circle center pixel coordinates and the camera intrinsic parameter matrix to determine a homography matrix; decomposing an initial rotation matrix and an initial translation matrix from the homography matrix as the initial pose.

5. The high-precision large-range pose measurement method based on the absolute space coding calibration board according to claim 1, characterized in that, The local correction of the initial pose based on the target encoding center point pixel coordinates, the actual physical distance and the grating center point world coordinates of the target minimum recognition unit is used to output a target pose, including: searching for the encoding recognition matrix of the target minimum recognition unit in a preset calibration board encoding table to match grating center point world coordinates; performing sum operation on the grating center point world coordinates and the actual physical distance to determine boresight point world coordinates of the boresight point; performing nonlinear least square solution on a projection error target function according to the target encoding center point pixel coordinates, the boresight point world coordinates and the initial pose to determine depth information; updating the initial pose based on the depth information to determine a target pose.

6. The high-precision large-range pose measurement method based on the absolute space coding calibration board according to claim 1, characterized in that, The determination process of the target calibration board captured image includes: performing preprocessing on the calibration board captured image of the double-layer transparent absolute space encoding calibration board to output a target calibration board captured image.

7. A high-precision large-range pose measurement system based on an absolute space-encoding calibration board, characterized in that, The double-layer transparent absolute space encoding calibration board has Aruco two-dimensional codes arranged on an upper surface and two-dimensional sinusoidal grating patterns arranged on a lower surface, and circular mark point groups are equidistantly arranged between adjacent Aruco two-dimensional codes; the Aruco two-dimensional codes of the minimum recognition unit form a unique encoding recognition matrix, and the encoding center point of the minimum recognition unit corresponds to the grating center point of the two-dimensional sinusoidal grating pattern one by one; the system includes: a pixel coordinate extraction module configured to determine target mark circle center pixel coordinates, target grating center point pixel coordinates and target encoding center point pixel coordinates of a target minimum recognition unit by using a target calibration board captured image; a physical distance determination module configured to determine an actual physical distance between a same line-of-sight point of a corresponding grating center point and a code center point according to an absolute phase distribution of a two-dimensional sinusoidal grating pattern to which the target minimum identification unit belongs, the target grating center point pixel coordinate, and the target code center point pixel coordinate; a pose solving module configured to match a code identification matrix of the target minimum identification unit with a marker circle center world coordinate, and to perform pose solving by combining the target marker circle center pixel coordinate and an intrinsic parameter matrix of a camera to determine an initial pose; a pose correction module configured to perform deep information local correction on the initial pose based on the target code center point pixel coordinate, the actual physical distance, and the grating center point world coordinate of the target minimum identification unit, and to output a target pose.

8. A computer device, comprising: A computer program product including a memory and a processor, the memory storing a computer program, the computer program being executed by the processor to cause the processor to perform the steps of the high-precision large-range pose measurement method based on the absolute spatial encoding calibration board according to any one of claims 1-6.

9. A computer readable storage medium having stored thereon computer programs / instructions, characterized in that, The computer program / instructions are executed by the processor to implement the steps of the high-precision large-range pose measurement method based on the absolute spatial encoding calibration board according to any one of claims 1-6.

10. A computer program product comprising computer programs / instructions, characterized in that, The computer program / instructions are executed by the processor to implement the steps of the high-precision large-range pose measurement method based on the absolute spatial encoding calibration board according to any one of claims 1-6.

Images