WO2024011764A1 - 标定参数确定方法、混合标定板、装置、设备和介质 - Google Patents

标定参数确定方法、混合标定板、装置、设备和介质 Download PDF

Info

Publication number
WO2024011764A1
WO2024011764A1 PCT/CN2022/123363 CN2022123363W WO2024011764A1 WO 2024011764 A1 WO2024011764 A1 WO 2024011764A1 CN 2022123363 W CN2022123363 W CN 2022123363W WO 2024011764 A1 WO2024011764 A1 WO 2024011764A1
Authority
WO
WIPO (PCT)
Prior art keywords
calibration
image
reference object
hybrid
plate
Prior art date
Application number
PCT/CN2022/123363
Other languages
English (en)
French (fr)
Inventor
朱烙盛
蒋念娟
沈小勇
吕江波
Original Assignee
深圳思谋信息科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳思谋信息科技有限公司 filed Critical 深圳思谋信息科技有限公司
Publication of WO2024011764A1 publication Critical patent/WO2024011764A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
    • G06T3/04—
    • G06T5/80—
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/13—Edge detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10004—Still image; Photographic image

Definitions

  • the present application relates to the technical field of camera calibration, and in particular to a method for determining calibration parameters, a hybrid calibration plate, a device, equipment and a medium.
  • a two-dimensional plane calibration plate is usually used to calibrate equipment parameters.
  • a circular array calibration plate can be used for camera calibration.
  • the calibration reference graphics in some two-dimensional plane calibration plates are easily deformed after perspective transformation, which leads to inaccurate determination of the calibration parameters of the image acquisition equipment.
  • the circles in the circular array calibration plate are very difficult to deform after perspective transformation. It may become an ellipse in the calibration image, so there may be errors in the position of the calibration point determined through the calibration image, which will affect the accuracy of the calibration parameters.
  • this application provides a calibration parameter determination method, hybrid calibration plate, device, equipment and medium that can improve the accuracy of calibration parameters.
  • this application provides a method for determining calibration parameters.
  • the method includes the following steps.
  • the hybrid calibration plate includes a first calibration reference object and a second calibration reference object; the position of the corner point of the first calibration reference object does not undergo perspective deformation.
  • the reference object image corresponding to the second calibration reference object is determined from the calibration image.
  • the initial calibration point position of the second calibration reference object is determined based on the reference object image, and the initial calibration point position is corrected.
  • the application also provides a hybrid calibration plate, including: a plate body; the plate body has a planar structure; at least one first calibration reference object and at least one second calibration reference object are provided on the surface of the plate body;
  • a calibration reference object is an object whose corner point position does not undergo perspective deformation.
  • the first calibration reference object and the second calibration reference object are arranged at intervals on the surface of the plate body.
  • first calibration reference objects there are multiple first calibration reference objects and multiple second calibration reference objects; multiple first calibration reference objects are centrally symmetrically arranged on the surface of the board; and multiple second calibration reference objects are arranged on the surface of the board.
  • the surface of the body is arranged symmetrically around the center.
  • the first calibration reference object is a rectangular pattern
  • the second calibration reference object is a concentric circle pattern.
  • this application also provides a calibration parameter determination device.
  • the device includes an image acquisition module, an image determination module, a position correction module and a parameter determination module.
  • the image acquisition module is used to acquire the calibration image obtained by image acquisition of the hybrid calibration plate;
  • the hybrid calibration plate includes a first calibration reference object and a second calibration reference object; the position of the corner point of the first calibration reference object does not undergo perspective deformation. .
  • the image determination module is configured to determine the reference object image corresponding to the second calibration reference object from the calibration image based on the position of the corner point of the first calibration reference object in the calibration image.
  • the position correction module is used to determine the initial calibration point position of the second calibration reference object based on the reference object image, and correct the initial calibration point position.
  • the parameter determination module is used to determine the image acquisition calibration parameters corresponding to the hybrid calibration plate based on the corrected calibration point position.
  • the corner point position is the corner image coordinate of the corner point of the first calibration reference object in the image coordinate system where the calibration image is located.
  • the image determination module includes a coordinate acquisition unit, a relationship determination unit and an image positioning unit.
  • the coordinate acquisition unit is used to obtain the corner point calibration coordinates of the corner point of the first calibration reference object in the calibration coordinate system where the hybrid calibration plate is located;
  • the relationship determination unit is used to obtain the positional relationship between the corner point image coordinates and the corner point calibration coordinates.
  • mapping conversion unit determines the mapping conversion relationship; the mapping conversion unit is used to realize the mapping conversion between the plane where the calibration image is located and the plane where the calibration plate template map is located under the calibration coordinate system; the image positioning unit is used to implement the mapping conversion relationship and the calibration plate template map based on the mapping conversion relationship. Locate the reference object image corresponding to the second calibration reference object from the calibration image.
  • the image positioning unit is also used to map the calibration image to the plane where the calibration plate template map is located in the calibration coordinate system according to the mapping conversion relationship to obtain the reference image; and combine the second calibration reference object in the reference image with the calibration
  • the position of the second calibration reference object in the template image is compared to determine the encoding identifier of the second calibration reference object in the reference image; the encoding identifier has unique corresponding position information; based on the mapping conversion relationship, the encoding identifier is The corresponding position information is reversely mapped to the calibration image to locate the reference object image corresponding to the second calibration reference object from the calibration image.
  • the second calibration reference object is a concentric circle pattern
  • the position correction module includes a contour extraction unit and a position determination unit.
  • the contour extraction unit is used to extract multiple reference circle outlines of the concentric circle pattern from the reference object image; the position determination unit is used to obtain the corresponding center coordinates of the reference circle outline in the reference object image as the initial calibration point position.
  • the contour extraction unit is also used to perform binarization processing on the reference object image to obtain a binarized image; perform edge detection on the binarized image to extract a preliminary edge contour of the binarized image; The edge contours are fitted to obtain multiple reference circle contours.
  • the position correction module is also used to determine the calibration circle profile that matches the reference circle profile from the hybrid calibration plate; perform eccentricity error correction at the physical radius corresponding to the hybrid calibration plate based on the initial calibration point position and the calibration circle profile. , get the corrected calibration point position.
  • the calibration parameter determination device further includes a parameter optimization module, which is used to calculate the reprojection error of the calibration image according to the image acquisition calibration parameters; construct a cost function based on the reprojection error and the image acquisition calibration parameters; to minimize The cost function is the optimization target, and the image acquisition and calibration parameters are optimized to obtain the optimized image acquisition and calibration parameters.
  • a parameter optimization module which is used to calculate the reprojection error of the calibration image according to the image acquisition calibration parameters; construct a cost function based on the reprojection error and the image acquisition calibration parameters; to minimize The cost function is the optimization target, and the image acquisition and calibration parameters are optimized to obtain the optimized image acquisition and calibration parameters.
  • this application also provides a computer device.
  • the computer device includes a memory and a processor.
  • the memory stores a computer program.
  • the processor executes the computer program, it implements the steps in the above calibration parameter determination method.
  • this application also provides a computer-readable storage medium.
  • a computer-readable storage medium has a computer program stored thereon. When the computer program is executed by a processor, the steps in the above calibration parameter determination method are implemented.
  • this application also provides a computer program product.
  • a computer program product includes a computer program that implements the steps in the above calibration parameter determination method when executed by a processor.
  • the above-mentioned calibration parameter determination method, hybrid calibration plate, calibration parameter determination device, computer equipment, storage medium and computer program product are obtained by acquiring a calibration image obtained by image acquisition of the hybrid calibration plate; the hybrid calibration plate includes a first calibration reference object and The second calibration reference object; the position of the corner point of the first calibration reference object does not undergo perspective deformation; according to the position of the corner point of the first calibration reference object in the calibration image, the reference corresponding to the second calibration reference object is determined from the calibration image Object image; determine the initial calibration point coordinates of the second calibration reference object based on the reference object image, and correct the initial calibration point coordinates; determine the image acquisition calibration parameters corresponding to the hybrid calibration plate based on the corrected calibration point coordinates.
  • This application designs a hybrid calibration plate that includes different calibration reference objects, and proposes a new calibration parameter determination method based on the hybrid calibration plate, that is, the first calibration reference object that does not undergo perspective deformation through the corner position is accurately viewed from The second calibration reference object image is determined in the calibration image, thereby ensuring the accuracy of the initial calibration point position obtained based on the second calibration reference object; correcting the initial calibration point position can effectively improve the accuracy of the corrected calibration point position. , thereby improving the accuracy of calibration parameters.
  • Figure 1 is a schematic flowchart of a method for determining calibration parameters in an embodiment of the present application.
  • Figure 2 is a schematic diagram of a hybrid calibration plate in an embodiment of the present application.
  • Figure 3 is a schematic diagram of the reference circle outline in the reference object image in one embodiment of the present application.
  • Figure 4 is a schematic flowchart of a method for determining calibration parameters in another embodiment of the present application.
  • Figure 5 is a structural block diagram of a calibration parameter determination device in an embodiment of the present application.
  • Figure 6 is an internal structure diagram of a computer device in an embodiment of the present application.
  • a calibration parameter determination method is provided.
  • the computer device may be a server or a terminal. It can be understood that this method can also be applied to a system including a server and a terminal, and is implemented through the interaction between the server and the terminal, where the terminal includes at least one of a mobile phone, a tablet computer, a notebook computer, or a desktop computer.
  • the method includes the following steps 102 to 108.
  • Step 102 Obtain the calibration image obtained by image acquisition of the hybrid calibration plate.
  • the calibration plate refers to a geometric model with a fixed spacing pattern array
  • the hybrid calibration plate refers to a geometric model with a fixed spacing and containing two or more different pattern arrays.
  • the hybrid calibration plate includes a first calibration reference object and a second calibration reference object.
  • the first calibration reference object refers to one of the patterns in the hybrid calibration plate, which is used for subsequent calibration.
  • the second calibration reference object refers to another pattern in the hybrid calibration plate that is different from the first calibration reference object, and is also used for subsequent calibration.
  • a corner point is an extreme point, that is, a point with particularly outstanding properties in some aspect, such as the end point of a line segment, the vertex of a geometric figure, or the point on a curve where the local curvature is maximum.
  • a corner point in this application may refer to a vertex in a pattern.
  • the position of the corner point of the first calibration reference object will not undergo perspective deformation during the calibration process.
  • perspective distortion refers to the fact that an object and its surrounding area are seen differently in reality than through the standard lens of the image capture device. That is, the object and its surrounding area in the standard lens of the image acquisition device may be bent or deformed due to changes in the relative proportions of far and near features.
  • the corner points of the rectangle will basically not undergo perspective deformation during the calibration process
  • the first calibration reference object may be a rectangle, and the corner points of the first calibration reference object may be the vertices of the rectangle.
  • the first calibration reference object includes at least one of a checkerboard pattern, a two-dimensional code pattern, a binary coding pattern, and the like.
  • the binary encoding pattern refers to a pattern composed of a binary matrix composed of a black border. The black border helps improve the accuracy of its positioning detection in the image, and the binary matrix is used to represent the uniqueness of the mark.
  • the graph complexity of the second calibration reference object is lower than the graph complexity of the first calibration reference object.
  • This application introduces a second calibration reference object whose graphic complexity is lower than that of the first calibration reference object in the hybrid calibration board. On the premise of ensuring that there is no offset error using the first calibration reference object, it can also reduce the error through the second calibration reference object. Process requirements for making hybrid calibration plates.
  • graph complexity refers to the complexity of the structure and elements of the graph. If the graph has many elements and a complicated structure, the complexity of the graph will be considered to be high; otherwise, the complexity of the graph will be considered to be relatively low. .
  • the second calibration reference object includes at least one of a circular pattern or a concentric circle pattern.
  • the ArUco pattern since the ArUco pattern has good robustness and the concentric circle pattern has good wear resistance, the ArUco pattern can be selected as the first calibration reference object, and the concentric circle pattern can be selected as the second Calibrate the reference object to combine the strengths of both patterns.
  • a blank area for placing objects can also be provided in the middle of the hybrid calibration plate to ensure that the object can be placed on the hybrid calibration plate without blocking at least one of the first calibration reference object or the second calibration reference object. kind.
  • the design diagram of the hybrid calibration plate including the ArUco pattern 202 and the concentric circle pattern 204 can be referred to FIG. 2 .
  • the overall shape of the hybrid calibration plate is a symmetrical figure, and the ArUco pattern 202 and the concentric circle pattern 204 are arranged at intervals.
  • the overall shape of the hybrid calibration plate is a circle, and the object is placed within the range of a regular polygon inscribed in the circle.
  • Each vertex position of the regular polygon inscribed in the circle is correspondingly placed with an ArUco pattern 202, and several concentric circle patterns 204 are placed between each adjacent ArUco pattern 202 connected by a straight line.
  • the coordinates of several other concentric circles forming an equilateral triangle formed by these several concentric circle patterns 204 can also be calculated, thereby forming a group of concentric circle distribution.
  • the image acquisition device takes photos of the hybrid calibration plate at multiple angles to obtain calibration images. Then, the computer device obtains the corresponding calibration image from the image acquisition device.
  • the image acquisition device refers to a device with a camera function, which can be but is not limited to various cameras and mobile devices.
  • Step 104 Determine the reference object image corresponding to the second calibration reference object from the calibration image based on the corner point position of the first calibration reference object in the calibration image.
  • the reference object image is an image composed of an image area corresponding to the second calibration reference object in the calibration image.
  • the computer device can determine the mapping conversion relationship between the plane where the corner point of the first calibration reference object is located and the plane where the calibration image is located based on the corner point position of the corner point of the first calibration reference object in the calibration image.
  • the computer device performs a rough positioning of the second calibration reference object in the calibration image based on the mapping conversion relationship, and determines the corresponding reference object image from the calibration image based on the position of the second calibration reference object obtained through the rough positioning in the calibration image. .
  • Step 106 Determine the initial calibration point position of the second calibration reference object based on the reference object image, and correct the initial calibration point position.
  • the initial calibration point position refers to the position of the calibration point of the second calibration reference object in the reference object image.
  • the calibration point of the second calibration reference object refers to one or more points in the second calibration reference object used to position the second calibration reference object. For example, it may be a center point or an edge point in the second calibration reference object. At least one of the others.
  • the calibration point of the second calibration reference object that is, the position of the initial calibration point may deviate during the perspective transformation process.
  • the initial calibration point position needs to be corrected.
  • the computer device determines the position of the calibration point of the second calibration reference object in the reference object image from the reference object, that is, the initial calibration point position. Then, the computer device corrects the coordinate position of the initial calibration point to obtain the corrected calibration point coordinate to eliminate the offset error caused by perspective transformation, thereby effectively improving the accuracy of the corrected calibration point position.
  • Step 108 Based on the corrected calibration point position, determine the image acquisition calibration parameters corresponding to the hybrid calibration plate.
  • the image acquisition calibration parameters include at least one of device internal parameters of the image acquisition device or device extrinsic parameters from the coordinate system where the calibration image is located to the coordinate system where the hybrid calibration plate is located.
  • the computer device can establish a mathematical model from the coordinate system where the calibration image is located to the coordinate system where the hybrid calibration plate is located based on the corrected calibration point position. Based on the calibration in the established mathematical model The function calculates the image acquisition calibration parameters.
  • the calibration function in OpenCV can be used to calculate the image acquisition calibration parameters.
  • OpenCV is a cross-platform computer vision and machine learning software library.
  • the calibration image is obtained by image acquisition of the hybrid calibration plate; the hybrid calibration plate includes a first calibration reference object and a second calibration reference object; the position of the corner point of the first calibration reference object does not occur.
  • Perspective deformation determine the reference object image corresponding to the second calibration reference object from the calibration image based on the position of the corner point of the first calibration reference object in the calibration image; determine the initial calibration point coordinates of the second calibration reference object based on the reference object image , and correct the initial calibration point coordinates; based on the corrected calibration point coordinates, determine the image acquisition calibration parameters corresponding to the hybrid calibration plate.
  • This application designs a hybrid calibration plate that includes different calibration reference objects, and proposes a new calibration parameter determination method based on the hybrid calibration plate, that is, the first calibration reference object that does not undergo perspective deformation through the corner position is accurately viewed from The second calibration reference object image is determined in the calibration image, thereby ensuring the accuracy of the initial calibration point position obtained based on the second calibration reference object; correcting the initial calibration point position can also effectively improve the accuracy of the corrected calibration point position. accuracy, thereby improving the accuracy of calibration parameters.
  • the corner position is the corner image coordinate of the corner point of the first calibration reference object in the image coordinate system where the calibration image is located.
  • Step 104 specifically includes but is not limited to: obtaining the corner point of the first calibration reference object.
  • the mapping conversion relationship is used to realize the mapping conversion between the plane where the calibration image is located and the plane where the calibration plate template map is located under the calibration coordinate system. That is to say, the calibration image can be converted to the plane where the calibration plate template image is located in the calibration coordinate system according to the mapping relationship, and the calibration template image can also be converted to the plane where the calibration image is located based on the mapping relationship.
  • the calibration template image refers to the top view taken only for the hybrid calibration plate at a plane parallel to the plane where the hybrid calibration plate is located.
  • the computer device obtains the corner point calibration coordinates of the corner point of the first calibration reference object in the calibration coordinate system where the hybrid calibration plate is located. Then, the computer device determines the mapping conversion relationship between the plane where the calibration image is located and the plane where the calibration plate template map under the calibration coordinates is located, based on the positional relationship between the corner point image coordinates and the corner point calibration coordinates. Finally, according to the above mapping conversion relationship, the computer device can locate the position of the second calibration reference image in the calibration image based on the position of the second calibration reference object in the calibration template image, and based on the position of the second calibration reference image in the calibration image. position, and extract the corresponding image area from the calibration image to obtain the reference object image.
  • This application can obtain an accurate mapping conversion relationship through the second calibration reference object that does not undergo perspective deformation at the corner position, thereby ensuring that the corresponding second calibration reference object is located from the calibration image based on the mapping conversion relationship and the calibration plate template diagram. the accuracy of the reference object image.
  • the step "locating the reference object image corresponding to the second calibration reference object from the calibration image according to the mapping conversion relationship and the calibration plate template” specifically includes but is not limited to: mapping the calibration image according to the mapping conversion relationship Go to the plane where the calibration plate template map is located in the calibration coordinate system to obtain the reference image; compare the position of the second calibration reference object in the reference image with the second calibration reference object in the calibration plate template map to determine the second calibration reference object in the reference image. Second, the coding identifier of the calibration reference object; based on the mapping conversion relationship, the position information corresponding to the coding identifier is reversely mapped to the calibration image, so as to locate the reference object image corresponding to the second calibration reference object from the calibration image.
  • the coding identifier is used to distinguish different second positioning reference objects, and the coding identifier has unique corresponding position information, that is, the coding identifier corresponding to each second calibration reference object has a unique corresponding position information corresponding to the second calibration reference object.
  • Position information such as position coordinates under calibration coordinates. It can be understood that the encoding identifier can be uniquely determined in advance according to the position of each second calibration reference object in the hybrid calibration plate or the calibration plate template map.
  • the computer device maps the calibration image to the plane where the calibration plate template map is located under the calibration coordinate system according to the mapping conversion relationship between the plane where the calibration image is located and the plane where the calibration plate template map is located under the calibration coordinate system, to obtain the reference image. . Since the reference image and the calibration template are located on the same plane, the computer device can compare the positions of the second calibration reference object in the reference image and the second calibration reference object in the calibration plate template. If the computer device recognizes that the position of a certain second calibration reference object in the reference image is the same as the position of a certain second calibration reference object in the calibration plate template, then the coding identifier corresponding to the two second calibration reference objects is determined.
  • the codes are the same, thereby determining the coding identifier of the second calibration reference object in the reference image. Then, based on the mapping conversion relationship between the plane where the calibration image is located and the plane where the calibration plate template map under the calibration coordinates is located, the computer device reversely maps the position information corresponding to the encoded identifier to the calibration image to locate the position from the calibration image
  • the second calibration reference object corresponds to the reference object image.
  • the mapping transformation relationship can be reflected by a single mapping transformation matrix.
  • the computer device can perform the second calibration based on the positional relationship between the corner point image coordinates and the corner point calibration coordinates. A single mapping transformation matrix for coarse positioning with reference to the object. Then, the computer device transforms the calibration image to the plane where the calibration plate template image is located under the calibration coordinates through the single mapping transformation matrix to obtain the reference image. After obtaining the reference image, compare the positions of the second calibration reference object in the reference image and the second calibration reference object in the calibration plate template to determine the coding identifier of the second calibration reference object in the reference image.
  • the computer device reversely maps the position information corresponding to the encoding identifier to the calibration image according to the inverse matrix of the single mapping transformation matrix, so as to locate the reference object image corresponding to the second calibration reference object from the calibration image.
  • the reference object image obtained by reverse mapping the position information corresponding to the encoding identifier is shown in Figure 3. As can be seen from Figure 3, the reference object image is the result of deformation of concentric circles.
  • the second calibration reference object is a concentric circle pattern.
  • Step 106 specifically includes but is not limited to: extracting multiple reference circle outlines of the concentric circle pattern from the reference object image; obtaining the reference circle outline in the reference object image. The corresponding circle center coordinate is used as the initial calibration point position.
  • concentric circles refer to circles with the same center and different radii.
  • the reference circle outline refers to the outline corresponding to circles in concentric circles with the same center and different radii. It should be noted that the outline of the concentric circles in the calibration image obtained after perspective transformation of the hybrid calibration plate is likely to become an ellipse. Correspondingly, the reference circle outline may be an elliptical outline.
  • the computer device extracts multiple reference circle outlines of the concentric circle pattern from the reference object image, and obtains the corresponding center coordinates of each reference circle outline in the reference object image as the initial calibration point coordinates.
  • the multiple reference circle outlines of the concentric circle pattern please refer to Figure 3 .
  • the step "extracting multiple reference circle outlines of the concentric circle pattern from the reference object image” specifically includes but is not limited to: performing binarization processing on the reference object image to obtain a binarized image; Perform edge detection on the binary image to extract the preliminary edge contour of the binary image; perform fitting processing on the preliminary edge contour to obtain multiple reference circle contours.
  • binarization processing is to set the gray value of the pixels on the image to 0 or 255, that is, to present an obvious visual effect of only black and white on the entire image.
  • Edge detection is a basic problem in image processing and computer vision. The purpose of edge detection is to identify points in digital images where brightness changes significantly.
  • Fitting is to connect a series of points on the plane with a smooth curve.
  • the computer device performs binarization processing on the reference object to obtain a binary image with only black and white visual effects.
  • the computer device performs edge detection on the binary image to obtain a preliminary edge contour formed by multiple edge points.
  • the computer equipment further performs fitting processing on multiple edge points obtained through edge detection to obtain clearer multiple reference circle outlines.
  • the computer device after the computer device performs edge detection on the binary image, it can also extract only sub-pixel-level edges, perform least-squares fitting on the extracted sub-pixel-level edges, and eliminate fitting points with errors greater than the threshold. , to obtain the fitting results, and cluster the fitting results into multiple reference circle contours.
  • sub-pixels are units smaller than pixels obtained by subdividing the basic unit of pixels, which can improve image resolution.
  • sub-pixel edges exist in areas of an image that experience progressive, excessive changes.
  • step 106 specifically includes but is not limited to: determining a calibration circle profile that matches the reference circle profile from the hybrid calibration plate; and determining the corresponding physical radius of the hybrid calibration plate based on the initial calibration point position and the calibration circle profile. Perform eccentric error correction to obtain the corrected calibration point position.
  • the calibration circle outline refers to the outline corresponding to circles with the same center and different radii as the concentric circles in each concentric circle pattern of the hybrid calibration plate.
  • the physical radius corresponding to the calibration circle outline on the hybrid calibration plate refers to the radius value of the calibration circle outline actually measured on the hybrid calibration plate.
  • the eccentricity error refers to the fact that the center projection of the circle in space is not equal to the center of the projected ellipse, and there is an error between the two center centers.
  • the computer device determines from the calibration mixing plate the calibration circle profile that matches the reference circle profile.
  • Matching means that the area of the reference circle profile is the same as the area of the calibration circle profile, and the radius of the reference circle profile is the same as the calibration circle profile. At least one of the same radii or the same diameter of the reference circle outline and the diameter of the calibration circle outline.
  • the computer equipment performs eccentric error correction at the physical radius corresponding to the hybrid calibration plate based on the initial calibration point position and the calibration circle outline to obtain the corrected calibration point position.
  • the calibration point position refers to the corresponding coordinates of the calibration point in the calibration image.
  • the concentric circle pattern in the reference object includes three ellipses with the same center as the concentric circles and different radii, they are recorded as ellipse 1, ellipse 2 and ellipse 3, and the center coordinates of ellipse 1, ellipse 2 and ellipse 3 are respectively: ( u B1, v B1 ), (u B2, v B2 ), (u B3, v B3 ), and the corresponding physical radii of ellipse 1, ellipse 2 and ellipse 3 in the hybrid calibration plate are r1, r2 and r3 respectively.
  • the eccentric error principle of the circle in perspective transformation determines the abscissa and ordinate of the calibration point in the calibration image, and determines the corresponding coordinates of the calibration point in the calibration image based on the abscissa and ordinate of the calibration point.
  • the computer device uses the abscissa coordinate of ellipse 1 in the calibration image, the abscissa coordinate of ellipse 2 in the calibration image, the abscissa coordinate of ellipse 3 in the calibration image, and the corresponding coordinates of ellipse 1 in the hybrid calibration plate.
  • the first intermediate variable, namely P 2 is calculated based on the physical radius, the corresponding physical radius of ellipse 2 in the hybrid calibration plate, and the corresponding physical radius of ellipse 3 in the hybrid calibration plate.
  • the computer equipment passes the abscissa coordinate of ellipse 1 in the calibration image, the abscissa coordinate of ellipse 2 in the calibration image, the corresponding physical radius of ellipse 1 in the hybrid calibration plate, the corresponding physical radius of ellipse 2 in the hybrid calibration plate and the third An intermediate variable is used to calculate the second intermediate variable, which is K.
  • the computer equipment determines the ordinate of ellipse 1 in the calibration image, the ordinate of ellipse 2 in the calibration image, the corresponding physical radius of ellipse 1 in the hybrid calibration plate, the corresponding physical radius of ellipse 2 in the hybrid calibration plate, and the third An intermediate variable is used to calculate the third intermediate variable, which is L. It should be noted that this application can calculate the first intermediate variable, the second intermediate variable and the third intermediate variable respectively through formula (1), formula (2) and formula (3).
  • the computer device calculates the first intermediate variable, the second intermediate variable and the third intermediate variable, according to the abscissa coordinate of the ellipse 1 in the calibration image, the corresponding physical radius of the ellipse 1 in the hybrid calibration plate, the first intermediate variable and the third intermediate variable.
  • Two intermediate variables are used to calculate the abscissa coordinate of the calibration point in the calibration image.
  • the computer device calculates the ordinate of the calibration point in the calibration image based on the ordinate of the ellipse 1 in the calibration image, the corresponding physical radius of the ellipse 1 in the hybrid calibration plate, the first intermediate variable and the third intermediate variable. It should be noted that this application can calculate the abscissa coordinate of the calibration point in the calibration image and the ordinate coordinate of the calibration point in the calibration image through formula (4) and formula (5) respectively.
  • P 2 , K and L are all intermediate variables used to calculate the coordinates of the corrected calibration point.
  • u c refers to the abscissa of the corrected calibration point
  • v c refers to the corrected calibration point.
  • the ordinate, (u c , v c ) refers to the coordinates of the corrected calibration point.
  • the calibration parameter determination method of the present application specifically includes but is not limited to: calculating the reprojection error of the calibration image based on the image acquisition calibration parameters; constructing a cost function based on the reprojection error and the image acquisition calibration parameters; to minimize The cost function is the optimization target, and the image acquisition and calibration parameters are optimized to obtain the optimized image acquisition and calibration parameters.
  • the reprojection error refers to the difference between the projection of the real three-dimensional space point on the plane of the calibration image and the reprojection.
  • the projection of the real three-dimensional space point on the plane of the calibration image refers to the pixel point on the calibration image.
  • Reprojection It refers to the virtual pixel points obtained based on the image acquisition calibration parameters.
  • the computer device extracts the pixel coordinates of the corrected calibration point position in the calibration image, and back-calculates to obtain the new pixel coordinates based on the coordinates of the calibration point in the calibration coordinate system where the hybrid calibration plate is located and the image acquisition calibration parameters. Then, the computer device calculates the sum of the two-norm of the corrected pixel coordinates of the calibration point position and the new pixel coordinates and averages them to obtain the reprojection error of the calibration image. After calculating the reprojection error, the computer device constructs a cost function based on the reprojection error and the image acquisition calibration parameters, and then minimizes the cost function to optimize the single mapping transformation matrix, thereby further optimizing the calculation based on the single mapping transformation matrix.
  • Image acquisition calibration parameters are obtained to obtain optimized image acquisition calibration parameters. This application optimizes the image acquisition calibration parameters through reprojection error, and can simultaneously consider the calculation error of the single mapping transformation matrix and the measurement error of the calibration point position, so its accuracy will be higher.
  • the coordinates can be obtained by using the calibration point position of the calibration parameters.
  • An optimization is performed as a constraint, and the specific cost function is shown in formula (6):
  • r rpj K(R kca X ia +t kca )-m ik (6)
  • K() represents the three-dimensional spatial coordinates in the device coordinate system where the image acquisition device is located, and is projected to the two-dimensional image coordinates of the calibration image through the internal parameters of the device.
  • X ia represents the three-dimensional coordinates of the i-th calibration point position to be optimized
  • m ik represents the two-dimensional image coordinate position of the i-th calibration point detected on the k-th calibration image.
  • the motion change of the hybrid calibration plate can also be used as a constraint to perform secondary optimization, that is, the three-dimensional point in the world coordinate system is X w , and the three-dimensional point in the device coordinate system where the image acquisition device is located is X c , then the relationship between the device coordinate system and the world coordinate system is as shown in formula (7):
  • R cw is the orientation of the image acquisition device
  • c cw is the coordinate of the center of the image acquisition device in the world coordinate system.
  • R kwa is the rotation of the calibration plate relative to the world coordinate system
  • c wa is the offset of the calibration plate coordinate system relative to the world coordinate system.
  • r rpjt K(R cw R kwa X ia -R cw (c cw +R kwa c wa ))-m ik (10)
  • the first calibration reference object in the hybrid calibration plate is an ArUco pattern
  • the second calibration reference object is a concentric circle pattern, as shown in Figure 4.
  • the calibration parameter determination method of the present application specifically includes but is not limited to: The following steps 402 to 412.
  • Step 402 Obtain the calibration image obtained by image acquisition of the hybrid calibration plate.
  • Step 404 Roughly locate the position of the concentric circle pattern in the calibration image to obtain a reference object image.
  • the computer device can obtain the corner point calibration coordinates of the corner points of the ArUco pattern in the calibration coordinate system where the hybrid calibration plate is located.
  • the mapping conversion relationship is determined.
  • the calibration image is mapped to the plane where the calibration plate template picture is located under the calibration coordinate system to obtain the reference image.
  • the position information corresponding to the encoding identifier is reversely mapped to the calibration image to locate the reference object image corresponding to the concentric circle pattern from the calibration image.
  • Step 406 Perform ellipse edge detection on the reference object image to obtain the initial calibration point position of the reference circle outline in the reference object image.
  • the reference object image is binarized to obtain a binarized image.
  • Perform edge detection on the binarized image to extract the preliminary edge contour of the binarized image.
  • the preliminary edge contours are fitted to obtain multiple reference circle contours. Obtain the center coordinates corresponding to the reference circle outline in the reference object image as the initial calibration point position.
  • Step 408 Perform eccentric error correction on the initial calibration point position to obtain the corrected calibration point position.
  • a calibration circle profile that matches a reference circle profile is determined from the hybrid calibration plate. According to the initial calibration point position and the calibration circle outline, the eccentricity error is corrected at the physical radius corresponding to the hybrid calibration plate, and the corrected calibration point position is obtained.
  • Step 410 Based on the corrected calibration point position, determine the image acquisition calibration parameters corresponding to the hybrid calibration plate.
  • Step 412 Optimize the image acquisition calibration parameters.
  • the reprojection error of the hybrid calibration plate is calculated based on the image acquisition calibration parameters.
  • the cost function is constructed based on the reprojection error and image acquisition calibration parameters. Taking the minimization of the cost function as the optimization goal, the image acquisition and calibration parameters are optimized to obtain the optimized image acquisition and calibration parameters.
  • embodiments of the present application also provide a calibration parameter determination device for implementing the above-mentioned calibration parameter determination method.
  • the solution to the problem provided by this device is similar to the solution recorded in the above method. Therefore, the specific limitations in the embodiments of one or more calibration parameter determination devices provided below can be found in the above description of the calibration parameter determination method. Limitations will not be repeated here.
  • the present application also provides a hybrid calibration plate, which includes a plate body.
  • the board body is a planar structure, and at least one first calibration reference object and at least one second calibration reference object are provided on the surface of the board body; the first calibration reference object is an object whose corner point position does not undergo perspective deformation, and the third calibration reference object A calibration reference object and a second calibration reference object are arranged at intervals on the surface of the plate body.
  • first calibration reference objects there are multiple first calibration reference objects and multiple second calibration reference objects; multiple first calibration reference objects are centrally symmetrically arranged on the surface of the board; and multiple second calibration reference objects are arranged on the surface of the board.
  • the surface of the body is arranged symmetrically around the center.
  • the first calibration reference object is a rectangular pattern
  • the second calibration reference object is a concentric circle pattern.
  • a calibration parameter determination device including: an image acquisition module 502 , an image determination module 504 , a position correction module 506 and a parameter determination module 508 .
  • the image acquisition module 502 is used to acquire the calibration image obtained by image acquisition of the hybrid calibration plate; the hybrid calibration plate includes a first calibration reference object and a second calibration reference object; the position of the corner point of the first calibration reference object does not cause perspective deformation.
  • the image determination module 504 is configured to determine the reference object image corresponding to the second calibration reference object from the calibration image based on the position of the corner point of the first calibration reference object in the calibration image.
  • the position correction module 506 is used to determine the initial calibration point position of the second calibration reference object based on the reference object image, and correct the initial calibration point position.
  • the parameter determination module 508 is used to determine the image acquisition calibration parameters corresponding to the hybrid calibration plate based on the corrected calibration point position.
  • the calibration image is obtained by image acquisition of the hybrid calibration plate;
  • the hybrid calibration plate includes a first calibration reference object and a second calibration reference object; the position of the corner point of the first calibration reference object does not occur.
  • Perspective deformation determine the reference object image corresponding to the second calibration reference object from the calibration image based on the position of the corner point of the first calibration reference object in the calibration image; determine the initial calibration point coordinates of the second calibration reference object based on the reference object image , and correct the initial calibration point coordinates; based on the corrected calibration point coordinates, determine the image acquisition calibration parameters corresponding to the hybrid calibration plate.
  • This application designs a hybrid calibration plate that includes different calibration reference objects, and proposes a new calibration parameter determination method based on the hybrid calibration plate, that is, the first calibration reference object that does not undergo perspective deformation through the corner position is accurately viewed from The second calibration reference object image is determined in the calibration image, thereby ensuring the accuracy of the initial calibration point position obtained based on the second calibration reference object; correcting the initial calibration point position can also effectively improve the accuracy of the corrected calibration point position. accuracy, thereby improving the accuracy of calibration parameters.
  • the corner point position is the corner image coordinate of the corner point of the first calibration reference object in the image coordinate system where the calibration image is located.
  • the image determination module 504 includes a coordinate acquisition unit, a relationship determination unit and an image positioning unit.
  • the coordinate acquisition unit is used to obtain the corner point calibration coordinates of the corner point of the first calibration reference object in the calibration coordinate system where the hybrid calibration plate is located;
  • the relationship determination unit is used to obtain the positional relationship between the corner point image coordinates and the corner point calibration coordinates.
  • mapping conversion unit determines the mapping conversion relationship; the mapping conversion unit is used to realize the mapping conversion between the plane where the calibration image is located and the plane where the calibration plate template map is located under the calibration coordinate system; the image positioning unit is used to implement the mapping conversion relationship and the calibration plate template map based on the mapping conversion relationship. Locate the reference object image corresponding to the second calibration reference object from the calibration image.
  • the image positioning unit is also used to map the calibration image to the plane where the calibration plate template map is located in the calibration coordinate system according to the mapping conversion relationship to obtain the reference image; and combine the second calibration reference object in the reference image with the calibration
  • the position of the second calibration reference object in the template image is compared to determine the encoding identifier of the second calibration reference object in the reference image; the encoding identifier has unique corresponding position information; based on the mapping conversion relationship, the encoding identifier is The corresponding position information is reversely mapped to the calibration image to locate the reference object image corresponding to the second calibration reference object from the calibration image.
  • the second calibration reference object is a concentric circle pattern
  • the position correction module includes a contour extraction unit and a position determination unit.
  • the contour extraction unit is used to extract multiple reference circle outlines of the concentric circle pattern from the reference object image; the position determination unit is used to obtain the corresponding center coordinates of the reference circle outline in the reference object image as the initial calibration point position.
  • the contour extraction unit is also used to perform binarization processing on the reference object image to obtain a binarized image; perform edge detection on the binarized image to extract a preliminary edge contour of the binarized image; The edge contours are fitted to obtain multiple reference circle contours.
  • the position correction module is also used to determine the calibration circle profile that matches the reference circle profile from the hybrid calibration plate; perform eccentricity error correction at the physical radius corresponding to the hybrid calibration plate based on the initial calibration point position and the calibration circle profile. , get the corrected calibration point position.
  • the calibration parameter determination device further includes a parameter optimization module, which is used to calculate the reprojection error of the calibration image according to the image acquisition calibration parameters; construct a cost function based on the reprojection error and the image acquisition calibration parameters; to minimize The cost function is the optimization target, and the image acquisition and calibration parameters are optimized to obtain the optimized image acquisition and calibration parameters.
  • a parameter optimization module which is used to calculate the reprojection error of the calibration image according to the image acquisition calibration parameters; construct a cost function based on the reprojection error and the image acquisition calibration parameters; to minimize The cost function is the optimization target, and the image acquisition and calibration parameters are optimized to obtain the optimized image acquisition and calibration parameters.
  • Each module in the above-mentioned calibration parameter determination device can be implemented in whole or in part by software, hardware and combinations thereof.
  • Each of the above modules may be embedded in or independent of the processor of the computer device in the form of hardware, or may be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.
  • a computer device is provided.
  • the computer device may be a server, and its internal structure diagram may be as shown in Figure 6 .
  • the computer device includes a processor, memory, and network interfaces connected through a system bus. Wherein, the processor of the computer device is used to provide computing and control capabilities.
  • the memory of the computer device includes non-volatile storage media and internal memory.
  • the non-volatile storage medium stores operating systems, computer programs and databases. This internal memory provides an environment for the execution of operating systems and computer programs in non-volatile storage media.
  • the database of the computer device is used to store data related to calibration parameters.
  • the network interface of the computer device is used to communicate with external terminals through a network connection.
  • the computer program implements a calibration parameter determination method when executed by the processor.
  • FIG. 6 is only a block diagram of a partial structure related to the solution of the present application, and does not constitute a limitation on the computer equipment to which the solution of the present application is applied.
  • Specific computer equipment can May include more or fewer parts than shown, or combine certain parts, or have a different arrangement of parts.
  • a computer device including a memory and a processor.
  • a computer program is stored in the memory.
  • the processor executes the computer program, it implements the steps in the above method embodiments.
  • a computer-readable storage medium is provided, with a computer program stored thereon.
  • the computer program is executed by a processor, the steps in the above method embodiments are implemented.
  • a computer program product including a computer program that implements the steps in each of the above method embodiments when executed by a processor.
  • the computer program can be stored in a non-volatile computer-readable storage.
  • the computer program when executed, may include the processes of the above method embodiments.
  • Any reference to memory, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory.
  • Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetic variable memory (Magnetoresistive Random Access Memory (MRAM), ferroelectric memory (Ferroelectric Random Access Memory, FRAM), phase change memory (Phase Change Memory, PCM), graphene memory, etc.
  • Volatile memory may include random access memory (Random Access Memory, RAM) or external cache memory, etc.
  • RAM Random Access Memory
  • RAM random access memory
  • RAM Random Access Memory
  • the databases involved in the various embodiments provided in this application may include at least one of a relational database and a non-relational database.
  • Non-relational databases may include blockchain-based distributed databases, etc., but are not limited thereto.
  • the processors involved in the various embodiments provided in this application may be general-purpose processors, central processing units, graphics processors, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to this.

Abstract

一种标定参数确定方法、混合标定板、装置、设备和介质，所述方法包括：获取对混合标定板进行图像采集得到的标定图像，所述混合标定板中包括第一标定参照对象和第二标定参照对象，第一标定参照对象的角点的位置不发生透视形变；根据所述第一标定参照对象的角点在标定图像中的位置，从所述标定图像中确定第二标定参照对象对应的参照对象图像；根据所述参照对象图像确定所述第二标定参照对象的初始标定点坐标，并对所述初始标定点坐标进行校正；基于校正后的标定点位置，确定所述混合标定板对应的图像采集标定参数。

Description

标定参数确定方法、混合标定板、装置、设备和介质
相关申请的交叉引用
本申请要求于2022年7月11日提交中国专利局，申请号为202210809183.6，申请名称为“标定参数确定方法、混合标定板、装置、设备和介质”的中国专利申请的优先权,在此将其全文引入作为参考。
技术领域
本申请涉及相机标定技术领域，特别是涉及一种标定参数确定方法、混合标定板、装置、设备和介质。
背景技术
在进行三维测量和三维重建的过程中，对图像采集设备的设备参数进行标定是非常重要一步。目前，通常采用二维平面标定板进行设备参数的标定处理，例如可以采用圆阵列标定板进行相机标定。其中，在利用二维平面标定板进行标定的过程中，需要对二维平面标定板中的标定参照图形进行透视变换。
然而，有些二维平面标定板中的标定参照图形在透视变换后容易发生形变，从而导致后续确定图像采集设备的标定参数不够准确，比如，圆阵列标定板中的圆形在经过透视变换后很有可能在标定图像中变成椭圆，因此通过标定图像所确定的标定点位置可能会存在误差，进而影响标定参数的准确性。
发明内容
基于此，本申请提供一种能够提高标定参数准确性的标定参数确定方法、混合标定板、装置、设备和介质。
第一方面，本申请提供了一种标定参数确定方法。所述方法包括如下步骤。
获取对混合标定板进行图像采集得到的标定图像；混合标定板中包括第一标定参照对象和第二标定参照对象；第一标定参照对象的角点的位置不发生透视形变。
根据第一标定参照对象的角点在标定图像中的角点位置，从标定图像中确定第二标定参照对象对应的参照对象图像。
根据参照对象图像确定第二标定参照对象的初始标定点位置，并对初始标定点位置进 行校正。
基于校正后的标定点位置，确定混合标定板对应的图像采集标定参数。
第二方面，本申请还提供了一种混合标定板，包括：板体；板体为平面结构；板体的表面上设置有至少一个第一标定参照对象和至少一个第二标定参照对象；第一标定参照对象是角点的位置不发生透视形变的对象。
第一标定参照对象和第二标定参照对象在板体的表面上呈间隔排布。
在一些实施例中，第一标定参照对象和第二标定参照对象皆为多个；多个第一标定参照对象在板体的表面上呈中心对称排布；多个第二标定参照对象在板体的表面上呈中心对称排布。
在一些实施例中，第一标定参照对象为矩形图案，第二标定参照对象为同心圆图案。
第三方面，本申请还提供了一种标定参数确定装置。装置包括图像获取模块、图像确定模块、位置校正模块和参数确定模块。
图像获取模块，用于获取对混合标定板进行图像采集得到的标定图像；混合标定板中包括第一标定参照对象和第二标定参照对象；第一标定参照对象的角点的位置不发生透视形变。
图像确定模块，用于根据第一标定参照对象的角点在标定图像中的角点位置，从标定图像中确定第二标定参照对象对应的参照对象图像。
位置校正模块，用于根据参照对象图像确定第二标定参照对象的初始标定点位置，并对初始标定点位置进行校正。
参数确定模块，用于基于校正后的标定点位置，确定混合标定板对应的图像采集标定参数。
在一些实施例中，角点位置是第一标定参照对象的角点在标定图像所在的图像坐标系下的角点图像坐标。图像确定模块包括坐标获取单元、关系确定单元和图像定位单元。坐标获取单元用于获取第一标定参照对象的角点在混合标定板所在的标定坐标系下的角点标定坐标；关系确定单元用于根据角点图像坐标与角点标定坐标之间的位置关系，确定映射转换关系；映射转换单元，用于实现标定图像所在平面与标定坐标系下的标定板模板图所在平面之间的映射转换；图像定位单元用于根据映射转换关系和标定板模板图，从标定图像中定位第二标定参照对象对应的参照对象图像。
在一些实施例中，图像定位单元还用于根据映射转换关系，将标定图像映射到标定坐标系下的标定板模板图所在平面，以得到参考图像；将参考图像中第二标定参照对象与标 定板模板图中的第二标定参照对象进行位置比对，以确定参考图像中第二标定参照对象的编码标识符；编码标识符具有唯一对应的位置信息；基于映射转换关系，将编码标识符所对应的位置信息反向映射至标定图像中，以从标定图像中定位第二标定参照对象对应的参照对象图像。
在一些实施例中，第二标定参照对象为同心圆图案，位置校正模块包括轮廓提取单元和位置确定单元。轮廓提取单元用于从参照对象图像中提取同心圆图案的多个参照圆轮廓；位置确定单元用于获取参照圆轮廓在参照对象图像中对应的圆心坐标，作为初始标定点位置。
在一些实施例中，轮廓提取单元还用于对参照对象图像进行二值化处理，得到二值化图像；对二值化图像进行边缘检测，以提取二值化图像的初步边缘轮廓；对初步边缘轮廓进行拟合处理，以得到多个参照圆轮廓。
在一些实施例中，位置校正模块还用于从混合标定板中确定与参照圆轮廓相匹配的标定圆轮廓；根据初始标定点位置和标定圆轮廓在混合标定板对应的物理半径进行偏心误差校正，得到校正后的标定点位置。
在一些实施例中，标定参数确定装置还包括参数优化模块，参数优化模块用于根据图像采集标定参数计算标定图像的重投影误差；基于重投影误差和图像采集标定参数构建代价函数；以最小化代价函数为优化目标，对图像采集标定参数进行优化，得到优化后的图像采集标定参数。
第四方面，本申请还提供了一种计算机设备。计算机设备包括存储器和处理器，存储器存储有计算机程序，处理器执行计算机程序时实现上述标定参数确定方法中的步骤。
第五方面，本申请还提供了一种计算机可读存储介质。计算机可读存储介质，其上存储有计算机程序，计算机程序被处理器执行时实现上述标定参数确定方法中的步骤。
第六方面，本申请还提供了一种计算机程序产品。计算机程序产品，包括计算机程序，该计算机程序被处理器执行时实现上述标定参数确定方法中的步骤。
上述标定参数确定方法、混合标定板、标定参数确定装置、计算机设备、存储介质和计算机程序产品，通过获取对混合标定板进行图像采集得到的标定图像；混合标定板中包括第一标定参照对象和第二标定参照对象；第一标定参照对象的角点的位置不发生透视形变；根据第一标定参照对象的角点在标定图像中的位置，从标定图像中确定第二标定参照对象对应的参照对象图像；根据参照对象图像确定第二标定参照对象的初始标定点坐标，并对初始标定点坐标进行校正；基于校正后的标定点坐标，确定混合标定板对应的图像采 集标定参数。本申请设计包括不同的标定参照对象的混合标定板，并基于该混合标定板提出了一种全新的标定参数确定方法，即通过角点位置不会发生透视形变的第一标定参照对象准确地从标定图像中确定第二标定参照对象图像，进而保证基于该第二标定参照对象中获取的初始标定点位置的准确性；对初始标定点位置进行校正，能够有效提高校正后的标定点位置的精度，进而提高标定参数准确性。
附图说明
图1为本申请一个实施例中标定参数确定方法的流程示意图。
图2为本申请一个实施例中混合标定板的示意图。
图3为本申请一个实施例中参照对象图像中参照圆轮廓的示意图。
图4为本申请另一个实施例中标定参数确定方法的流程示意图。
图5为本申请一个实施例中标定参数确定装置的结构框图。
图6为本申请一个实施例中计算机设备的内部结构图。
具体实施方式
为了使本申请的目的、技术方案及优点更加清楚明白，以下结合附图及实施例，对本申请进行进一步详细说明。应当理解，此处描述的具体实施例仅仅用以解释本申请，并不用于限定本申请。
在一些实施例中，如图1所示，提供了一种标定参数确定方法，本实施例以该方法应用于计算机设备进行举例说明，该计算机设备可以是服务器或终端。可以理解的是，该方法还可以应用于包括服务器和终端的系统，并通过服务器和终端的交互实现，其中，该终端包括手机、平板电脑、笔记本电脑或台式电脑中的至少一种。本实施例中，该方法包括以下步骤102至步骤108。
步骤102，获取对混合标定板进行图像采集得到的标定图像。
其中，标定板指的是一种带有固定间距图案阵列的几何模型，混合标定板指的是一种带有固定间距、且包含两个或两个以上不同图案阵列的几何模型。
在一些实施例中，混合标定板中包括第一标定参照对象和第二标定参照对象。其中，第一标定参照对象指的是混合标定板中的其中一种图案，用于进行后续的标定。第二标定参照对象指的是混合标定板中另一种区别于第一标定参照对象的图案，同样用于进行后续的标定。
角点，就是极值点，即在某方面属性特别突出的点，例如线段的终点、几何图形的顶点，或者是曲线上局部曲率最大的点。在本申请中的角点可以指图案中的顶点。
在一些实施例中，第一标定参照对象的角点的位置在标定过程中不会发生透视形变。其中，透视变形指的是一个物体及其周围区域在现实中看到的与在图像采集设备的标准镜头中看到的不同。也就是说，由于远近特征的相对比例变化，可能会导致图像采集设备的标准镜头中的物体及其周围区域发生了弯曲或变形。由于矩形的角点在标定过程中基本不会发生透视形变，因此第一标定参照对象可以为矩形，且第一标定参照对象的角点可以为矩形的顶点。
在一些实施例中，第一标定参照对象包括棋盘格图案、二维码图案或二进制编码图案等中的至少一种。其中，二进制编码图案即ArUco图案，是指一个由黑色边框组成的二进制矩阵所构成的图案，黑色边框有助于提高其在图像中定位检测的准确性，二进制矩阵用来表示标记的唯一性。
在一些实施例中，第二标定参照对象的图形复杂度低于第一标定参照对象的图形复杂度。本申请在混合标定板中引入图形复杂度低于第一标定参照对象的第二标定参照对象，在保证利用第一标定参照对象无偏移误差的前提下，还能够通过第二标定参照对象降低制作混合标定板的工艺要求。其中，图形复杂度指的是图形的结构和元素的复杂程度，如果图形的元素繁多且结构繁琐，则认为该图形的复杂度将对较高；反之，则认为该图形的复杂度相对较低。
在一些实施例中，第二标定参照对象包括圆形图案或同心圆图案中的至少一种。
在一些实施例中，由于ArUco图案具有很好的鲁棒性，且同心圆图案具有很好的耐磨性，则可以选择ArUco图案作为第一标定参照对象，且可以选择同心圆图案作为第二标定参照对象，以结合两种图案各自的优势。
在一些实施例中，还可以在混合标定板的中间设置用于放置物体的空白区域，以保证物体放置到混合标定板上能够不遮挡第一标定参照对象或第二标定参照对象中的至少一种。
在一些实施例中，包括ArUco图案202和同心圆图案204的混合标定板的设计图可参照图2。可以看出，该混合标定板的整体形状为对称图形，且ArUco图案202和同心圆图案204为间隔设置。其中，混合标定板的整体形状为圆形，物体放置在圆内接正多边形范围内。圆内接正多边形的各个顶点位置对应放置ArUco图案202，且在每相邻的ArUco图案202直线相连之间放置若干个同心圆图案204。若干个同心圆图案204设置以后，还可 以计算这若干个同心圆图案204组成等边三角形的另外多个同心圆坐标，从而组成一组同心圆的分布。
具体地，图像采集设备在多个角度下针对混合标定板进行拍摄，得到标定图像。接着，计算机设备则从图像采集设备中获取对应的标定图像。其中，图像采集设备指的是具备拍照功能的设备，可以但不限于是各种相机和移动设备。
步骤104，根据第一标定参照对象的角点在标定图像中的角点位置，从标定图像中确定第二标定参照对象对应的参照对象图像。
其中，参照对象图像为标定图像中对应于第二标定参照对象的图像区域所构成的图像。
具体地，计算机设备根据第一标定参照对象的角点在标定图像中的角点位置，可以确定第一标定参照对象的角点所在的平面与标定图像所在的平面之间的映射转换关系。计算机设备根据该映射转换关系在标定图像中对第二标定参照对象进行粗定位，并根据通过粗定位得到的第二标定参照对象在标定图像中的位置，从标定图像中确定对应的参照对象图像。
步骤106，根据参照对象图像确定第二标定参照对象的初始标定点位置，并对初始标定点位置进行校正。
其中，初始标定点位置指的是第二标定参照对象的标定点在参照对象图像中的位置。第二标定参照对象的标定点指的是第二标定参照对象中的一个或多个用于对第二标定参照对象进行定位的点，例如可以是第二标定参照对象中的中心点或边缘点等中的至少一种。
可以理解，第二标定参照对象的标定点，即初始标定点位置在经过透视变换的过程中可能存在偏差。为了保证标定参数计算的准确性，考虑在确定初始标定点位置之后，还需要对初始标定点位置进行校正。
具体地，计算机设备从参照对象中确定出第二标定参照对象的标定点在参照对象图像中的位置，即初始标定点位置。接着，计算机设备对初始标定点坐标位置进行校正，得到校正后的标定点坐标，以消除通过透视变换所导致的偏移误差，进而有效提高校正后的标定点位置的精度。
步骤108，基于校正后的标定点位置，确定混合标定板对应的图像采集标定参数。
其中，图像采集标定参数包括图像采集设备的设备内参或标定图像所在的坐标系到混合标定板所在的坐标系的设备外参中的至少一种。
具体地，计算机设备基于校正后的标定点位置，就可以根据校正后的标定点位置建立标定图像所在的坐标系到混合标定板所在的坐标系的数学模型，基于建立好的数学模型中 的标定函数计算得到图像采集标定参数。
在一些实施例中，可以利用OpenCV中的标定函数计算图像采集标定参数。其中，OpenCV是一个跨平台计算机视觉和机器学习软件库。
上述标定参数确定方法中，通过获取对混合标定板进行图像采集得到的标定图像；混合标定板中包括第一标定参照对象和第二标定参照对象；第一标定参照对象的角点的位置不发生透视形变；根据第一标定参照对象的角点在标定图像中的位置，从标定图像中确定第二标定参照对象对应的参照对象图像；根据参照对象图像确定第二标定参照对象的初始标定点坐标，并对初始标定点坐标进行校正；基于校正后的标定点坐标，确定混合标定板对应的图像采集标定参数。本申请设计包括不同的标定参照对象的混合标定板，并基于该混合标定板提出了一种全新的标定参数确定方法，即通过角点位置不会发生透视形变的第一标定参照对象准确地从标定图像中确定第二标定参照对象图像，进而保证基于该第二标定参照对象中获取的初始标定点位置的准确性；对初始标定点位置进行校正，还能够有效提高校正后的标定点位置的精度，进而提高标定参数准确性。
在一些实施例中，角点位置是第一标定参照对象的角点在标定图像所在的图像坐标系下的角点图像坐标，步骤104具体包括但不限于包括：获取第一标定参照对象的角点在混合标定板所在的标定坐标系下的角点标定坐标；根据角点图像坐标与角点标定坐标之间的位置关系，确定映射转换关系；根据映射转换关系和标定板模板图，从标定图像中定位第二标定参照对象对应的参照对象图像。
其中，映射转换关系，用于实现标定图像所在平面与标定坐标系下的标定板模板图所在平面之间的映射转换。也就是说，标定图像可以根据该映射关系转换至标定坐标系下的标定板模板图所在的平面上，标定模板图同样也可以根据该映射关系转换至标定图像所在的平面上。
标定模板图，指的是在平行于混合标定板所在的平面处，只针对混合标定板进行拍摄得到的俯视图。
具体地，计算机设备获取第一标定参照对象的角点在混合标定板所在的标定坐标系下的角点标定坐标。接着，计算机设备根据角点图像坐标与角点标定坐标之间的位置关系，确定标定图像所在平面与标定坐标下的标定板模板图所在的平面之间的映射转换关系。最后，计算机设备根据上述映射转换关系，就能够根据第二标定参照对象在标定模板图的位置来定位第二标定参照图像在标定图像中的位置，并基于第二标定参照图像在标定图像中的位置，从标定图像中提取对应图像区域，以得到参照对象图像。本申请通过角点位置不 会发生透视形变的第二标定参照对象，能够获取到准确的映射转换关系，进而保证根据该映射转换关系和标定板模板图从标定图像中定位第二标定参照对象对应的参照对象图像的准确性。
在一些实施例中，步骤“根据映射转换关系和标定板模板图，从标定图像中定位第二标定参照对象对应的参照对象图像”具体包括但不限于包括：根据映射转换关系，将标定图像映射到标定坐标系下的标定板模板图所在平面，以得到参考图像；将参考图像中第二标定参照对象与标定板模板图中的第二标定参照对象进行位置比对，以确定参考图像中第二标定参照对象的编码标识符；基于映射转换关系，将编码标识符所对应的位置信息反向映射至标定图像中，以从标定图像中定位第二标定参照对象对应的参照对象图像。
其中，编码标识符用于区分不同的第二定位参照对象，编码标识符具有唯一对应的位置信息，即每个第二标定参照对象所对应的编码标识符具有唯一对应于第二标定参照对象的位置信息，例如标定坐标下的位置坐标。可以理解，编码标识符可以预先根据各个第二标定参照对象在混合标定板或标定板模板图中的位置来唯一确定。
具体地，计算机设备根据标定图像所在平面与标定坐标下的标定板模板图所在的平面之间的映射转换关系，将标定图像映射到标定坐标系下的标定板模板图所在平面，以得到参考图像。由于参考图像和标定模板图都位于同一个平面，所以计算机设备可以将参考图像中第二标定参照对象与标定板模板图中的第二标定参照对象进行位置比对。若计算机设备识别出参考图像中的某个第二标定参照对象的位置和标定板模板图中的某个第二标定参照对象的位置相同，则确定这两个第二标定参照对象对应的编码标识符相同，从而确定参考图像中第二标定参照对象的编码标识符。接着，计算机设备基于标定图像所在平面与标定坐标下的标定板模板图所在的平面之间的映射转换关系，将编码标识符对应的位置信息反向映射至标定图像中，以从标定图像中定位第二标定参照对象对应的参照对象图像。本申请通过在参考图像中匹配各个第二标定参照对象的编码标识符，能够准确地确定对应第二标定参照对象在标定图像中的位置，从而保证基于该位置所提取的第二标定参照对象对应的参照对象图像的准确性。
在一些实施例中，映射转换关系可以通过单映射变换矩阵来体现，计算机设备在获取到角点标定坐标之后，可以根据角点图像坐标与角点标定坐标之间的位置关系，对第二标定参照对象进行粗定位的单映射变换矩阵。接着，计算机设备通过单映射变换矩阵，将标定图像变换到标定坐标下的标定板模板图所在的平面，以得到参考图像。得到参考图像之后，将参考图像中第二标定参照对象与标定板模板图中的第二标定参照对象进行位置比对， 以确定参考图像中第二标定参照对象的编码标识符。接着，计算机设备根据单映射变换矩阵的逆矩阵，将编码标识符对应的位置信息反向映射至标定图像中，以从标定图像中定位第二标定参照对象对应的参照对象图像。将编码标识符对应的位置信息反向映射所得到的参照对象图像如图3所示。从图3可知，参照对象图像是同心圆发生形变后的结果。
在一些实施例中，第二标定参照对象为同心圆图案，步骤106具体包括但不限于包括：从参照对象图像中提取同心圆图案的多个参照圆轮廓；获取参照圆轮廓在参照对象图像中对应的圆心坐标，作为初始标定点位置。
其中，同心圆指的是圆心相同半径不同的圆。参照圆轮廓指的是同心圆中各个圆心相同且半径不同的圆所对应的轮廓。需要说明的是，混合标定板在经过透视变换之后得到的标定图像中的同心圆的轮廓很有可能变成椭圆。对应的，参照圆轮廓可能为椭圆轮廓。
具体地，计算机设备从参照对象图像中提取同心圆图案的多个参照圆轮廓，并获取各个参照圆轮廓在参照对象图像中对应的圆心坐标，作为初始标定点坐标。其中，同心圆图案的多个参照圆轮廓可参照图3。使用同心圆图案的圆坐标作为初始标定点坐标，由于可以利用同心圆的圆外围的所有像素，因此可以减少图像噪声的影响，进而保证初始标定点坐标的准确性。
在一些实施例中，步骤“从参照对象图像中提取同心圆图案的多个参照圆轮廓”具体包括但不限于包括：对参照对象图像进行二值化处理，得到二值化图像；对二值化图像进行边缘检测，以提取二值化图像的初步边缘轮廓；对初步边缘轮廓进行拟合处理，以得到多个参照圆轮廓。
其中，二值化处理就是将图像上的像素点的灰度值设置为0或255，也就是将整个图像呈现出明显的只有黑和白的视觉效果。
边缘检测，是图像处理和计算机视觉中的基本问题，边缘检测的目的是标识数字图像中亮度变化明显的点。
拟合，就是把平面上一系列的点，用一条光滑的曲线连接起来。
具体地，计算机设备对参照对象进行二值化处理，以得到只有黑和白的视觉效果的二值化图像。接着，计算机设备对二值化图像进行边缘检测，得到多个边缘点所形成的初步边缘轮廓。此外，计算机设备进一步对进行边缘检测得到的多个边缘点进行拟合处理，以得到更为清晰的多个参照圆轮廓。
在一些实施例中，计算机设备对二值化图像进行边缘检测之后，还可以仅提取亚像素级边缘，对提取得到的亚像素级边缘进行最小二乘拟合，剔除误差大于阈值的拟合点，以 得到拟合结果，并将拟合结果聚类成多个参照圆轮廓。其中，亚像素是将像素这个基本单位再进行细分后得到的比像素还小的单位，可以提高图像分辨率。通常情况下，亚像素级边缘存在于图像中逐渐发生过度变化的区域。
在一些实施例中，步骤106具体还包括但不限于包括：从混合标定板中确定与参照圆轮廓相匹配的标定圆轮廓；根据初始标定点位置和标定圆轮廓在混合标定板对应的物理半径进行偏心误差校正，得到校正后的标定点位置。
其中，标定圆轮廓指的是在混合标定板的各个同心圆图案中，与同心圆的圆心相同且半径不同的圆所对应的轮廓。
标定圆轮廓在混合标定板对应的物理半径指的是，在混合标定板中实际测量出的标定圆轮廓的半径值。
偏心误差，指的是空间中的圆的圆心投影不等于投影出的椭圆的圆心，两个圆心之间存在误差。
具体地，计算机设备从标定混合板中确定与参照圆轮廓相匹配的标定圆轮廓，相匹配指的是参照圆轮廓的面积与标定圆轮廓的面积相同、参照圆轮廓的半径与标定圆轮廓的半径相同或者参照圆轮廓的直径与标定圆轮廓的直径相同中的至少一种。接着，计算机设备根据初始标定点位置和标定圆轮廓在混合标定板对应的物理半径进行偏心误差校正，得到校正后的标定点位置。本申请通过对初始标定点位置进行偏心误差校准，能够有效避免同心圆图案的圆心在经过透视变换所产生的偏移误差，以提高标定精度。
在一些实施例中，标定点位置指的是标定点在标定图像中对应的坐标。若参照对象中的同心圆图案包括三个与同心圆的圆心相同且半径不同的椭圆，记为椭圆1、椭圆2和椭圆3，且椭圆1、椭圆2和椭圆3的圆心坐标分别为：(u B1，v B1)、(u B2，v B2)、(u B3，v B3)，且椭圆1、椭圆2和椭圆3在混合标定板中对应的物理半径分别为r1、r2、r3，根据圆在透视变换的偏心误差原理确定出标定点在标定图像中的横坐标和纵坐标，根据标定点的横坐标和纵坐标确定标定点在标定图像中对应的坐标。
在一些实施例中，首先，计算机设备通过椭圆1在标定图像中的横坐标、椭圆2在标定图像中的横坐标、椭圆3在标定图像中的横坐标、椭圆1在混合标定板中对应的物理半径、椭圆2在混合标定板中对应的物理半径和椭圆3在混合标定板中对应的物理半径，计算出第一中间变量，即P 2。其次，计算机设备通过椭圆1在标定图像中的横坐标、椭圆2在标定图像中的横坐标、椭圆1在混合标定板中对应的物理半径、椭圆2在混合标定板中对应的物理半径和第一中间变量，计算出第二中间变量，即K。接着，计算机设备根据椭 圆1在标定图像中的纵坐标、椭圆2在标定图像中的纵坐标、椭圆1在混合标定板中对应的物理半径、椭圆2在混合标定板中对应的物理半径和第一中间变量，计算出第三中间变量，即L。需要说明的是，本申请可通过公式(1)、公式(2)和公式(3)分别计算出第一中间变量、第二中间变量和第三中间变量。
在计算机设备计算出第一中间变量、第二中间变量和第三中间变量之后，根据椭圆1在标定图像中的横坐标、椭圆1在混合标定板中对应的物理半径、第一中间变量和第二中间变量，计算出标定点在标定图像中的横坐标。计算机设备根据椭圆1在标定图像中的纵坐标、椭圆1在混合标定板中对应的物理半径、第一中间变量和第三中间变量，计算出标定点在标定图像中的纵坐标。需要说明的是，本申请可通过公式(4)和公式(5)分别计算出标定点在标定图像中的横坐标和标定点在标定图像中的纵坐标。
Figure PCTCN2022123363-appb-000001
Figure PCTCN2022123363-appb-000002
Figure PCTCN2022123363-appb-000003
Figure PCTCN2022123363-appb-000004
Figure PCTCN2022123363-appb-000005
其中，P 2、K和L都是用于计算校正后的标定点的坐标对应的中间变量，u c指的是校正后的标定点的横坐标，v c指的是校正后的标定点的纵坐标，(u c，v c)指的是校正后的标定点的坐标。
在一些实施例中，本申请的标定参数确定方法具体还包括但不限于包括：根据图像采集标定参数计算标定图像的重投影误差；基于重投影误差和图像采集标定参数构建代价函数；以最小化代价函数为优化目标，对图像采集标定参数进行优化，得到优化后的图像采集标定参数。
其中，重投影误差指的真实三维空间点在标定图像的平面上的投影和重投影的差值，真实三维空间点在标定图像的平面上的投影指的是标定图像上的像素点，重投影指的是基于图像采集标定参数得到的虚拟的像素点。
具体地，计算机设备提取标定图像中校正后的标定点位置的像素坐标，根据该标定点在混合标定板所在的标定坐标系下的坐标以及图像采集标定参数进行反计算得到新的像 素坐标。接着，计算机设备计算校正后的标定点位置的像素坐标和新的像素坐标的二范数求和并取平均值，即可得到标定图像的重投影误差。计算机设备在计算出重投影误差之后，基于重投影误差和图像采集标定参数构建代价函数，然后最小化该代价函数，以优化单映射变换矩阵，从而进一步优化基于该单映射变换矩阵所计算得到的图像采集标定参数，得到优化后的图像采集标定参数。本申请通过重投影误差优化图像采集标定参数，能够同时考虑单映射变换矩阵的计算误差和标定点位置的测量误差，所以其精度会更高。
在一些实施例中，设图像采集设备的设备内参为K，标定图像所在的坐标系到混合标定板所在的坐标系的设备外参R kca和t kca，可以利用标定参数的标定点位置得到坐标作为约束进行一次优化，具体的代价函数如公式(6)所示：
r rpj＝K(R kcaX ia+t kca)-m ik   (6)
其中，K()表示将图像采集设备所在的设备坐标系中的三维空间坐标通过设备内参投影到标定图像的二维图像坐标，X ia表示待优化的第i个标定点位置的三维坐标，m ik表示第i个标定点位置在第k张标定图像上检测出的二维影像坐标位置。
在另一些实施例中，还可以利用混合标定板的运动变化作为约束进行二次优化，即世界坐标系中的三维点为X w，图像采集设备所在的设备坐标系中的三维点为X c，则设备坐标系与世界坐标系的关系如公式(7)所示：
X c＝R cw(X w-c cw)   (7)
其中，R cw为图像采集设备的朝向，c cw为图像采集设备的中心在世界坐标系的坐标。
在一些实施例中，假设混合标定板的标定板坐标系下的三维点为X a，则在图像采集设备针对混合标定板的每个拍照的k时刻，X w与X a的关系为如公式(8)所示：
X w＝R kwa(X a-c wa)   (8)
其中，R kwa为标定板相对于世界坐标系的旋转，c wa为标定板坐标系相对于世界坐标系的偏移。此时，可推出图像采集设备所在的设备坐标系与每个拍照的k时刻的标定板坐标系之间的变换关系，如公式(9)所示：
X c＝R cwR kwaX a-R cw(c cw+R kwac wa)  (9)
对公式(6)至公式(10)进行分析转换，将最终的重投影误差改为r rpjt，具体的计算过程如公式(10)所示，根据最终计算出的重投影误差，可以进一步对图像采集标定参数进行优化，以提高标定精度。
r rpjt＝K(R cwR kwaX ia-R cw(c cw+R kwac wa))-m ik   (10)
在一些实施例中，混合标定板中的第一标定参照对象为ArUco图案，第二标定参照对 象为同心圆图案，如图4所示，本申请的标定参数确定方法具体还包括但不限于包括以下步骤402至步骤412。
步骤402，获取对混合标定板进行图像采集得到的标定图像。
步骤404，对同心圆图案在标定图像中的位置进行粗定位，得到参照对象图像。
在一些实施例中，计算机设备可以获取ArUco图案的角点在混合标定板所在的标定坐标系下的角点标定坐标。根据ArUco图案的角点在标定图像所在的图像坐标系下的角点图像坐标与角点标定坐标之间的位置关系，确定映射转换关系。根据映射转换关系，将标定图像映射到标定坐标系下的标定板模板图所在平面，以得到参考图像。将参考图像中同心圆图案与标定板模板图中的同心圆图案进行位置比对，以确定参考图像中同心圆图案的编码标识符。基于映射转换关系，将编码标识符所对应的位置信息反向映射至标定图像中，以从标定图像中定位同心圆图案对应的参照对象图像。
步骤406，对参照对象图像进行椭圆边缘检测，得到参照对象图像中参照圆轮廓的初始标定点位置。
在一些实施例中，对参照对象图像进行二值化处理，得到二值化图像。对二值化图像进行边缘检测，以提取二值化图像的初步边缘轮廓。对初步边缘轮廓进行拟合处理，以得到多个参照圆轮廓。获取参照圆轮廓在参照对象图像中对应的圆心坐标，作为初始标定点位置。
步骤408，对初始标定点位置进行偏心误差校正，得到校正后的标定点位置。
在一些实施例中，从混合标定板中确定与参照圆轮廓相匹配的标定圆轮廓。根据初始标定点位置和标定圆轮廓在混合标定板对应的物理半径进行偏心误差校正，得到校正后的标定点位置。
步骤410，基于校正后的标定点位置，确定混合标定板对应的图像采集标定参数。
步骤412，对图像采集标定参数进行优化。
在一些实施例中，根据图像采集标定参数计算混合标定板的重投影误差。基于重投影误差和图像采集标定参数构建代价函数。以最小化代价函数为优化目标，对图像采集标定参数进行优化，得到优化后的图像采集标定参数。
应该理解的是，虽然如上所述的各实施例所涉及的流程图中的各个步骤按照箭头的指示依次显示，但是这些步骤并不是必然按照箭头指示的顺序依次执行。除非本文中有明确的说明，这些步骤的执行并没有严格的顺序限制，这些步骤可以以其它的顺序执行。而且，如上所述的各实施例所涉及的流程图中的至少一部分步骤可以包括多个步骤或者多个阶 段，这些步骤或者阶段并不必然是在同一时刻执行完成，而是可以在不同的时刻执行，这些步骤或者阶段的执行顺序也不必然是依次进行，而是可以与其它步骤或者其它步骤中的步骤或者阶段的至少一部分轮流或者交替地执行。
基于同样的发明构思，本申请实施例还提供了一种用于实现上述所涉及的标定参数确定方法的标定参数确定装置。该装置所提供的解决问题的实现方案与上述方法中所记载的实现方案相似，故下面所提供的一个或多个标定参数确定装置实施例中的具体限定可以参见上文中对于标定参数确定方法的限定，在此不再赘述。
在一些实施例中，本申请还提供一种混合标定板，该混合标定板包括板体。其中，板体为平面结构，板体的表面上设置有至少一个第一标定参照对象和至少一个第二标定参照对象；第一标定参照对象是角点的位置不发生透视形变的对象，且第一标定参照对象和第二标定参照对象在板体的表面上呈间隔排布。
在一些实施例中，第一标定参照对象和第二标定参照对象皆为多个；多个第一标定参照对象在板体的表面上呈中心对称排布；多个第二标定参照对象在板体的表面上呈中心对称排布。
在一些实施例中，第一标定参照对象为矩形图案，第二标定参照对象为同心圆图案。
在一些实施例中，如图5所示，提供了一种标定参数确定装置，包括：图像获取模块502、图像确定模块504、位置校正模块506和参数确定模块508。
图像获取模块502，用于获取对混合标定板进行图像采集得到的标定图像；混合标定板中包括第一标定参照对象和第二标定参照对象；第一标定参照对象的角点的位置不发生透视形变。
图像确定模块504，用于根据第一标定参照对象的角点在标定图像中的角点位置，从标定图像中确定第二标定参照对象对应的参照对象图像。
位置校正模块506，用于根据参照对象图像确定第二标定参照对象的初始标定点位置，并对初始标定点位置进行校正。
参数确定模块508，用于基于校正后的标定点位置，确定混合标定板对应的图像采集标定参数。
上述标定参数确定装置中，通过获取对混合标定板进行图像采集得到的标定图像；混合标定板中包括第一标定参照对象和第二标定参照对象；第一标定参照对象的角点的位置不发生透视形变；根据第一标定参照对象的角点在标定图像中的位置，从标定图像中确定第二标定参照对象对应的参照对象图像；根据参照对象图像确定第二标定参照对 象的初始标定点坐标，并对初始标定点坐标进行校正；基于校正后的标定点坐标，确定混合标定板对应的图像采集标定参数。本申请设计包括不同的标定参照对象的混合标定板，并基于该混合标定板提出了一种全新的标定参数确定方法，即通过角点位置不会发生透视形变的第一标定参照对象准确地从标定图像中确定第二标定参照对象图像，进而保证基于该第二标定参照对象中获取的初始标定点位置的准确性；对初始标定点位置进行校正，还能够有效提高校正后的标定点位置的精度，进而提高标定参数准确性。
在一些实施例中，角点位置是第一标定参照对象的角点在标定图像所在的图像坐标系下的角点图像坐标。图像确定模块504包括坐标获取单元、关系确定单元和图像定位单元。坐标获取单元用于获取第一标定参照对象的角点在混合标定板所在的标定坐标系下的角点标定坐标；关系确定单元用于根据角点图像坐标与角点标定坐标之间的位置关系，确定映射转换关系；映射转换单元，用于实现标定图像所在平面与标定坐标系下的标定板模板图所在平面之间的映射转换；图像定位单元用于根据映射转换关系和标定板模板图，从标定图像中定位第二标定参照对象对应的参照对象图像。
在一些实施例中，图像定位单元还用于根据映射转换关系，将标定图像映射到标定坐标系下的标定板模板图所在平面，以得到参考图像；将参考图像中第二标定参照对象与标定板模板图中的第二标定参照对象进行位置比对，以确定参考图像中第二标定参照对象的编码标识符；编码标识符具有唯一对应的位置信息；基于映射转换关系，将编码标识符所对应的位置信息反向映射至标定图像中，以从标定图像中定位第二标定参照对象对应的参照对象图像。
在一些实施例中，第二标定参照对象为同心圆图案，位置校正模块包括轮廓提取单元和位置确定单元。轮廓提取单元用于从参照对象图像中提取同心圆图案的多个参照圆轮廓；位置确定单元用于获取参照圆轮廓在参照对象图像中对应的圆心坐标，作为初始标定点位置。
在一些实施例中，轮廓提取单元还用于对参照对象图像进行二值化处理，得到二值化图像；对二值化图像进行边缘检测，以提取二值化图像的初步边缘轮廓；对初步边缘轮廓进行拟合处理，以得到多个参照圆轮廓。
在一些实施例中，位置校正模块还用于从混合标定板中确定与参照圆轮廓相匹配的标定圆轮廓；根据初始标定点位置和标定圆轮廓在混合标定板对应的物理半径进行偏心误差校正，得到校正后的标定点位置。
在一些实施例中，标定参数确定装置还包括参数优化模块，参数优化模块用于根据 图像采集标定参数计算标定图像的重投影误差；基于重投影误差和图像采集标定参数构建代价函数；以最小化代价函数为优化目标，对图像采集标定参数进行优化，得到优化后的图像采集标定参数。
上述标定参数确定装置中的各个模块可全部或部分通过软件、硬件及其组合来实现。上述各模块可以硬件形式内嵌于或独立于计算机设备中的处理器中，也可以以软件形式存储于计算机设备中的存储器中，以便于处理器调用执行以上各个模块对应的操作。
在一些实施例中，提供了一种计算机设备，该计算机设备可以是服务器，其内部结构图可以如图6所示。该计算机设备包括通过系统总线连接的处理器、存储器和网络接口。其中，该计算机设备的处理器用于提供计算和控制能力。该计算机设备的存储器包括非易失性存储介质和内存储器。该非易失性存储介质存储有操作系统、计算机程序和数据库。该内存储器为非易失性存储介质中的操作系统和计算机程序的运行提供环境。该计算机设备的数据库用于存储与标定参数相关的数据。该计算机设备的网络接口用于与外部的终端通过网络连接通信。该计算机程序被处理器执行时以实现一种标定参数确定方法。
本领域技术人员可以理解，图6中示出的结构，仅仅是与本申请方案相关的部分结构的框图，并不构成对本申请方案所应用于其上的计算机设备的限定，具体的计算机设备可以包括比图中所示更多或更少的部件，或者组合某些部件，或者具有不同的部件布置。
在一些实施例中，还提供了一种计算机设备，包括存储器和处理器，存储器中存储有计算机程序，该处理器执行计算机程序时实现上述各方法实施例中的步骤。
在一些实施例中，提供了一种计算机可读存储介质，其上存储有计算机程序，该计算机程序被处理器执行时实现上述各方法实施例中的步骤。
在一些实施例中，提供了一种计算机程序产品，包括计算机程序，该计算机程序被处理器执行时实现上述各方法实施例中的步骤。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程，是可以通过计算机程序来指令相关的硬件来完成，所述的计算机程序可存储于一非易失性计算机可读取存储介质中，该计算机程序在执行时，可包括如上述各方法的实施例的流程。其中，本申请所提供的各实施例中所使用的对存储器、数据库或其它介质的任何引用，均可包括非易失性和易失性存储器中的至少一种。非易失性存储器可包括只读存储器(Read-Only Memory，ROM)、磁带、软盘、闪存、光存储器、高密度嵌入式非易失性存储器、阻变存储器(ReRAM)、磁变存储器(Magnetoresistive Random Access Memory，MRAM)、铁电存 储器(Ferroelectric Random Access Memory，FRAM)、相变存储器(Phase Change Memory，PCM)、石墨烯存储器等。易失性存储器可包括随机存取存储器(Random Access Memory，RAM)或外部高速缓冲存储器等。作为说明而非局限，RAM可以是多种形式，比如静态随机存取存储器(Static Random Access Memory，SRAM)或动态随机存取存储器(Dynamic Random Access Memory，DRAM)等。本申请所提供的各实施例中所涉及的数据库可包括关系型数据库和非关系型数据库中至少一种。非关系型数据库可包括基于区块链的分布式数据库等，不限于此。本申请所提供的各实施例中所涉及的处理器可为通用处理器、中央处理器、图形处理器、数字信号处理器、可编程逻辑器、基于量子计算的数据处理逻辑器等，不限于此。
以上实施例的各技术特征可以进行任意的组合，为使描述简洁，未对上述实施例中的各个技术特征所有可能的组合都进行描述，然而，只要这些技术特征的组合不存在矛盾，都应当认为是本说明书记载的范围。
以上所述实施例仅表达了本申请的几种实施方式，其描述较为具体和详细，但并不能因此而理解为对本申请专利范围的限制。应当指出的是，对于本领域的普通技术人员来说，在不脱离本申请构思的前提下，还可以做出若干变形和改进，这些都属于本申请的保护范围。因此，本申请的保护范围应以所附权利要求为准。

Claims (15)

  1. 一种标定参数确定方法，其特征在于，所述方法包括：
    获取对混合标定板进行图像采集得到的标定图像；所述混合标定板中包括第一标定参照对象和第二标定参照对象；第一标定参照对象的角点的位置不发生透视形变；
    根据所述第一标定参照对象的角点在所述标定图像中的角点位置，从所述标定图像中确定所述第二标定参照对象对应的参照对象图像；
    根据所述参照对象图像确定所述第二标定参照对象的初始标定点位置，并对所述初始标定点位置进行校正；
    基于校正后的标定点位置，确定所述混合标定板对应的图像采集标定参数。
  2. 根据权利要求1所述的方法，其特征在于，所述角点位置，是所述第一标定参照对象的角点在所述标定图像所在的图像坐标系下的角点图像坐标；
    所述根据所述第一标定参照对象的角点在所述标定图像中的角点位置，从所述标定图像中确定所述第二标定参照对象对应的参照对象图像包括：
    获取所述第一标定参照对象的角点在所述混合标定板所在的标定坐标系下的角点标定坐标；
    根据所述角点图像坐标与所述角点标定坐标之间的位置关系，确定映射转换关系；所述映射转换关系，用于实现所述标定图像所在平面与所述标定坐标系下的标定板模板图所在平面之间的映射转换；
    根据所述映射转换关系和所述标定板模板图，从所述标定图像中定位第二标定参照对象对应的参照对象图像。
  3. 根据权利要求2所述的方法，其特征在于，所述根据所述映射转换关系和所述标定板模板图，从所述标定图像中定位第二标定参照对象对应的参照对象图像包括：
    根据所述映射转换关系，将所述标定图像映射到所述标定坐标系下的标定板模板图所在平面，以得到参考图像；
    将参考图像中第二标定参照对象与所述标定板模板图中的第二标定参照对象进行位置比对，以确定所述参考图像中第二标定参照对象的编码标识符；所述编码标识符具有唯一对应的位置信息；
    基于所述映射转换关系，将所述编码标识符所对应的位置信息反向映射至所述标定图像中，以从所述标定图像中定位第二标定参照对象对应的参照对象图像。
  4. 根据权利要求1所述的方法，其特征在于，所述第二标定参照对象为同心圆图案； 所述根据所述参照对象图像确定所述第二标定参照对象的初始标定点位置，包括：
    从所述参照对象图像中提取同心圆图案的多个参照圆轮廓；
    获取所述参照圆轮廓在所述参照对象图像中对应的圆心坐标，作为所述初始标定点位置。
  5. 根据权利要求4所述的方法，其特征在于，所述从所述参照对象图像中提取同心圆图案的多个参照圆轮廓，包括：
    对所述参照对象图像进行二值化处理，得到二值化图像；
    对所述二值化图像进行边缘检测，以提取所述二值化图像的初步边缘轮廓；
    对所述初步边缘轮廓进行拟合处理，以得到多个参照圆轮廓。
  6. 根据权利要求4所述的方法，其特征在于，所述对所述初始标定点位置进行校正，包括：
    从所述混合标定板中确定与所述参照圆轮廓相匹配的标定圆轮廓；
    根据所述初始标定点位置和所述标定圆轮廓在所述混合标定板对应的物理半径进行偏心误差校正，得到校正后的标定点位置。
  7. 根据权利要求1至6任一项所述的方法，其特征在于，所述方法还包括：
    根据所述图像采集标定参数计算所述标定图像的重投影误差；
    基于所述重投影误差和所述图像采集标定参数构建代价函数；
    以最小化所述代价函数为优化目标，对所述图像采集标定参数进行优化，得到优化后的图像采集标定参数。
  8. 一种如权利要求1至7任一方法中的混合标定板，包括板体，其中：
    所述板体为平面结构；所述板体的表面上设置有至少一个第一标定参照对象和至少一个第二标定参照对象；所述第一标定参照对象是角点的位置不发生透视形变的对象；
    所述第一标定参照对象和所述第二标定参照对象在所述板体的表面上呈间隔排布。
  9. 根据权利要求8所述的混合标定板，其特征在于，所述第一标定参照对象和所述第二标定参照对象皆为多个；多个第一标定参照对象在所述板体的表面上呈中心对称排布；多个第二标定参照对象在所述板体的表面上呈中心对称排布。
  10. 根据权利要求9所述的混合标定板，其特征在于，所述第一标定参照对象为矩形图案，所述第二标定参照对象为同心圆图案。
  11. 一种标定参数确定装置，包括：
    图像获取模块，用于获取对混合标定板进行图像采集得到的标定图像；所述混合标定 板中包括第一标定参照对象和第二标定参照对象；第一标定参照对象的角点的位置不发生透视形变；
    图像确定模块，用于根据所述第一标定参照对象的角点在所述标定图像中的角点位置，从所述标定图像中确定所述第二标定参照对象对应的参照对象图像；
    位置校正模块，用于根据所述参照对象图像确定所述第二标定参照对象的初始标定点位置，并对所述初始标定点位置进行校正；
    参数确定模块，用于基于校正后的标定点位置，确定所述混合标定板对应的图像采集标定参数。
  12. 根据权利要求11所述的标定参数确定装置，其特征在于，所述角点位置，是所述第一标定参照对象的角点在所述标定图像所在的图像坐标系下的角点图像坐标；
    所述图像确定模块，包括：
    坐标获取单元，用于获取所述第一标定参照对象的角点在所述混合标定板所在的标定坐标系下的角点标定坐标；
    关系确定单元,根据所述角点图像坐标与所述角点标定坐标之间的位置关系，确定映射转换关系；
    映射转换单元，用于实现所述标定图像所在平面与所述标定坐标系下的标定板模板图所在平面之间的映射转换；和
    图像定位单元，用于根据所述映射转换关系和所述标定板模板图，从所述标定图像中定位第二标定参照对象对应的参照对象图像。
  13. 根据权利要求12所述的标定参数确定装置，其特征在于，所述图像定位单元还用于：
    根据所述映射转换关系，将所述标定图像映射到所述标定坐标系下的标定板模板图所在平面，以得到参考图像；
    将参考图像中第二标定参照对象与所述标定板模板图中的第二标定参照对象进行位置比对，以确定所述参考图像中第二标定参照对象的编码标识符；所述编码标识符具有唯一对应的位置信息；
    基于所述映射转换关系，将所述编码标识符所对应的位置信息反向映射至所述标定图像中，以从所述标定图像中定位第二标定参照对象对应的参照对象图像。
  14. 一种计算机设备，包括存储器和处理器，所述存储器存储有计算机程序，其特征在于，所述处理器执行所述计算机程序时实现权利要求1至7中任一项所述的方法的步骤。
  15. 一种计算机可读存储介质，其上存储有计算机程序，其特征在于，所述计算机程序被处理器执行时实现权利要求1至7中任一项所述的方法的步骤。
PCT/CN2022/123363 2022-07-11 2022-09-30 标定参数确定方法、混合标定板、装置、设备和介质 WO2024011764A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210809183.6A CN115147499A (zh) 2022-07-11 2022-07-11 标定参数确定方法、混合标定板、装置、设备和介质
CN202210809183.6 2022-07-11

Publications (1)

Publication Number Publication Date
WO2024011764A1 true WO2024011764A1 (zh) 2024-01-18

Family

ID=83412498

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/123363 WO2024011764A1 (zh) 2022-07-11 2022-09-30 标定参数确定方法、混合标定板、装置、设备和介质

Country Status (2)

Country Link
CN (1) CN115147499A (zh)
WO (1) WO2024011764A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116152809B (zh) * 2023-02-02 2023-07-28 深圳思谋信息科技有限公司 设备控制方法、系统、计算机设备和存储介质
CN115861429B (zh) * 2023-02-28 2023-06-16 深圳思谋信息科技有限公司 图像采集设备标定方法、装置、计算机设备和存储介质
CN117437304B (zh) * 2023-12-18 2024-04-16 科大讯飞(苏州)科技有限公司 安检机标定方法及相关方法、装置、设备和存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107633536A (zh) * 2017-08-09 2018-01-26 武汉科技大学 一种基于二维平面模板的相机标定方法及系统
CN112132907A (zh) * 2020-09-22 2020-12-25 北京的卢深视科技有限公司 一种相机标定方法、装置、电子设备及存储介质
CN112330755A (zh) * 2020-11-26 2021-02-05 展讯通信（上海）有限公司 环视系统的标定评价方法及装置、存储介质、终端
CN113012234A (zh) * 2021-03-16 2021-06-22 中国人民解放军火箭军工程大学 基于平面变换的高精度相机标定方法
WO2022108449A1 (en) * 2020-11-20 2022-05-27 Nearfield Instruments B.V. Alignment system and method for aligning an object having an alignment mark

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107633536A (zh) * 2017-08-09 2018-01-26 武汉科技大学 一种基于二维平面模板的相机标定方法及系统
CN112132907A (zh) * 2020-09-22 2020-12-25 北京的卢深视科技有限公司 一种相机标定方法、装置、电子设备及存储介质
WO2022108449A1 (en) * 2020-11-20 2022-05-27 Nearfield Instruments B.V. Alignment system and method for aligning an object having an alignment mark
CN112330755A (zh) * 2020-11-26 2021-02-05 展讯通信（上海）有限公司 环视系统的标定评价方法及装置、存储介质、终端
CN113012234A (zh) * 2021-03-16 2021-06-22 中国人民解放军火箭军工程大学 基于平面变换的高精度相机标定方法

Also Published As

Publication number Publication date
CN115147499A (zh) 2022-10-04

Similar Documents

Publication Publication Date Title
WO2024011764A1 (zh) 标定参数确定方法、混合标定板、装置、设备和介质
CN109949899B (zh) 图像三维测量方法、电子设备、存储介质及程序产品
US11321937B1 (en) Visual localization method and apparatus based on semantic error image
US10810718B2 (en) Method and device for three-dimensional reconstruction
CN101311963B (zh) 摄像机定位用圆标志点圆心的图像投影点位置的获取方法
CN111123242B (zh) 一种基于激光雷达和相机的联合标定方法及计算机可读存储介质
CN112132907B (zh) 一种相机标定方法、装置、电子设备及存储介质
CN110443879B (zh) 一种基于神经网络的透视误差补偿方法
CN102750697A (zh) 一种参数标定方法及装置
CN110738273B (zh) 图像特征点的匹配方法、装置、设备及存储介质
CN112396640A (zh) 图像配准方法、装置、电子设备及存储介质
CN113012234B (zh) 基于平面变换的高精度相机标定方法
CN116129037B (zh) 视触觉传感器及其三维重建方法、系统、设备及存储介质
CN115830135A (zh) 一种图像处理方法、装置及电子设备
CN114792345A (zh) 一种基于单目结构光系统的标定方法
Song et al. Flexible line-scan camera calibration method using a coded eight trigrams pattern
CN114862866B (zh) 标定板的检测方法、装置、计算机设备和存储介质
CN110458951A (zh) 一种电网杆塔的建模数据获取方法及相关装置
CN116524041A (zh) 一种相机标定方法、装置、设备及介质
CN116385575A (zh) 图像重构方法、装置、计算机设备和存储介质
CN114897990A (zh) 一种基于神经网络的相机畸变标定方法、系统和存储介质
CN115601451B (zh) 外参数据标定方法、装置、计算机设备和存储介质
CN116500851B (zh) 3d成像系统及其标定方法、装置、设备和存储介质
CN116844058B (zh) 一种指针式仪表示数的识别方法、装置、设备及存储介质
RU2538319C1 (ru) Устройство поиска дубликатов изображений

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22950874

Country of ref document: EP

Kind code of ref document: A1