A vision system positioning method based on a mirror-type stereoscopic target

Abstract

The application discloses a kind of vision system positioning methods based on mirror type stereoscopic target, it is related to visual measurement technical field, for the problems such as high-precision stereoscopic target structure complex, processing difficult, yield low, not easy to save, high cost, mirror type stereoscopic target is constructed using plane target and plane mirror, combined with computer vision and image processing technology, mirror type stereoscopic target is fixed on optical three-dimensional sensor, by stereoscopic target feature point coordinate unification, pose solving of stereoscopic target and optical three-dimensional sensor, optical three-dimensional sensor positioning, three-dimensional point cloud obtained by optical three-dimensional sensor in different positions is converted to global coordinate system, realize the global unification of multi-view measurement data, and can be used for the integrity measurement of object.The application has the advantages of low cost, simple processing and strong versatility.

Description

Technical Field

[0001] This invention belongs to the field of visual measurement technology, specifically relating to a visual system localization method based on a mirror-type stereo target. Background Technology

[0002] Visual measurement technology uses clear images captured by cameras as information carriers and leverages the powerful data processing capabilities of computers to achieve non-contact, high-precision, and automated measurement, leading to its widespread application in industrial inspection. However, a single optical 3D sensor is limited by its field of view, resulting in a limited measurement range; a single measurement can only obtain local 3D information about the surface of the object being measured. To acquire complete 3D information, multiple movements of the optical 3D sensor are typically required to complete multi-angle measurements. While optical 3D sensors can perform independent measurements within their own measurement coordinate system, the measurement data acquired at different locations lacks uniformity. To unify the data in the local measurement coordinate system, a global coordinate system needs to be established for positioning the optical 3D sensor.

[0003] Currently, several common methods exist: 1. Combining a laser tracker with a vision measurement system, using the laser tracker's own coordinate system as the global coordinate system, and calculating the pose of the optical 3D sensor to unify multi-viewpoint measurement data. This method involves fewer coordinate transformation steps and offers high measurement accuracy, but it is costly and complex to operate. 2. Attaching a certain number of marker points to the surface of the object being measured, using the measurement coordinate system of the optical 3D sensor at its first position as the global coordinate system, and stitching together 3D data in the local measurement coordinate system through feature point recognition and matching. This method is simple and easy to operate, but there are many limitations on the placement of marker points, leading to problems such as large cumulative errors and missing data. 3. Planar targets have the advantages of simple fabrication, high accuracy, and easy storage. In the global calibration of vision systems, multiple targets in different poses need to be captured. Due to occlusion issues with planar targets, it is often necessary to construct a 3D target by combining multiple planar targets. However, common 3D targets suffer from complex structures, difficult processing, low yield rates, difficulty in storage, and high costs. Therefore, a new type of 3D target needs to be designed to minimize processing costs while achieving accurate positioning by the vision system. Summary of the Invention

[0004] To address the problems of traditional stereo targets, such as complex structure, difficult processing, low yield, difficulty in preservation, and high cost, this invention proposes a vision system positioning method based on a mirror-type stereo target. The mirror-type stereo target is composed of a planar target connected to a planar reflector, eliminating the need for complex mechanical processing. Combining computer vision and image processing technologies, the mirror-type stereo target is fixed on an optical 3D sensor. By unifying the coordinates of the stereo target's feature points, solving the pose of the stereo target and the optical 3D sensor, and positioning the optical 3D sensor, the 3D point cloud acquired by the optical 3D sensor at different positions is transformed into a global coordinate system, achieving global unification of multi-viewpoint measurement data. This method can be used for the integrity measurement of objects.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A vision system localization method based on a mirrored stereo target includes the following steps:

[0007] S1: Design a mirror-type stereo target, which consists of a planar target and a planar reflector. The planar target has four different feature patterns.

[0008] S2: Establish a three-dimensional target coordinate system and a virtual planar target coordinate system respectively, and determine the origin position and coordinate axis direction of each coordinate system;

[0009] S3: Solve the relative pose between the virtual planar target and the planar target to obtain the coordinates of each feature point on the mirrored 3D target in the 3D target coordinate system;

[0010] S4: Fix the mirror-type stereo target on the optical 3D sensor and calculate the relative pose between the optical 3D sensor and the mirror-type stereo target;

[0011] S5: Fix a pre-calibrated camera above the experimental platform, establish a global coordinate system, freely move the optical 3D sensor, use the camera to photograph a mirrored stereo target, and ensure that there are at least two feature patterns in the camera's field of view. Use the stereo target image acquired by the camera to locate the optical 3D sensor, and globally unify the 3D data acquired by the optical 3D sensor at different locations.

[0012] Furthermore, in S1, the design of the mirror-type three-dimensional target specifically involves:

[0013] S101: The planar target is made of ceramic material. The four feature patterns on the planar target are circular patterns with coded features, namely one circle, two concentric circles, three concentric circles, and four concentric circles. The feature points on the planar target are the centers of the four circular patterns and the common tangent points of adjacent circular patterns.

[0014] S102: Set the plane mirror as an aluminum-plated front mirror;

[0015] S103: A planar target is reflected by a planar mirror to form a virtual planar target. The planar target and the virtual planar target are symmetrical about the planar mirror. The line connecting the corresponding feature points on the two targets is perpendicular to the planar mirror, and the distances from the corresponding feature points to the planar mirror are equal.

[0016] Furthermore, S2 includes:

[0017] S201: Establishing a three-dimensional target coordinate system based on a planar target O st -X st Y st Z st , origin O st Let X be the center of the three concentric circles on the planar target. st The axis points to the center of the four concentric circles on the planar target, Y. st The axis points to the center of the two concentric circles on the planar target, Z. st The axis is perpendicular to the target plane, Z st The positive direction of the axis is determined according to the right-hand rule.

[0018] S202: Establishing a virtual plane target coordinate system based on a virtual plane target O stv -X stv Y stv Z stv , origin O stv Let X be the center of the three concentric circles on the virtual planar target. stv The axis points to the center of the four concentric circles on the virtual planar target, Y. stv The axis points to the center of two concentric circles on the virtual planar target, Z. stv The Z axis is perpendicular to the virtual target plane. stv The positive direction of the axis is determined according to the right-hand rule.

[0019] Furthermore, S3 includes:

[0020] S301: Place the mirror-type stereo target within the pre-calibrated field of view of the camera, and have the camera capture a clear stereo target image;

[0021] S302: Establish camera coordinate system O c -X c Y c Z c , origin O c For the camera's optical center, X c The axis is parallel to the horizontal direction of the image and points to the right, Y c The axis is parallel to the vertical direction of the image and points downwards, Z. cThe axis is the optical axis, and the camera coordinate system conforms to the right-hand rule;

[0022] S303: Solve for O using the coordinates of feature points on the planar target in the image and their coordinates in the three-dimensional target coordinate system. st -X st Y st Z st coordinate system to O c -X c Y c Z c Rotation matrix of coordinate system Translation vector

[0023] S304: Establish virtual camera coordinate system O v -X v Y v Z v The virtual camera coordinate system is symmetrical to the camera coordinate system about the plane mirror, and the virtual camera coordinate system conforms to the left-hand rule;

[0024] S305: Establish a virtual camera coordinate system O that conforms to the right-hand rule. vr -X vr Y vr Z vr , origin O vr Y vr Axis, Z vr The axis and the origin O in the virtual camera coordinate system v Y v Axis, Z v Axis consistent, X vr Axial direction and X v The axes are in opposite directions;

[0025] S306: Flip the image horizontally, and use the coordinates of feature points on the virtual plane target in the flipped image to calculate O. st -X st Y st Z st coordinate system to O vr -X vr Y vr Z vr Rotation matrix of coordinate system Translation vector

[0026] S307: Based on the relative poses of the planar target and the virtual camera, and the fact that the relative poses of the virtual planar target and the camera are the same, O stv -X stv Y stv Z stv coordinate system to O c -Xc Y c Z c Rotation matrix of coordinate system Translation vector and and same;

[0027] S308: Solve for O according to formula (1) stv -X stv Y stv Z stv coordinate system to O st -X st Y st Z st Rotation matrix of coordinate system Translation vector

[0028]

[0029] Solve the coordinates of feature points on the virtual planar target in the coordinate system of the three-dimensional target.

[0030] Furthermore, S4 includes:

[0031] S401: Using a dot array planar target and a pre-calibrated camera, by adjusting the position of the optical 3D sensor and the dot array planar target, the dot array planar target is placed within the field of view of the optical 3D sensor and the camera, so that at least two feature patterns on the mirrored stereo target are within the field of view of the camera, and the optical 3D sensor and the camera respectively acquire a clear image.

[0032] S402: Establish the coordinate system of the dot array planar target O t -X t Y t Z t , origin O t X is the center of the upper left feature point on the dot array planar target. t Axis, Y t The axes point to the horizontal and vertical directions of the dot array planar target, respectively, Z. t The axis is perpendicular to the plane of the dot array target, and the coordinate system of the dot array target plane conforms to the right-hand rule.

[0033] S403: Establish the coordinate system of the optical three-dimensional sensor O s -X s Y s Z s , origin O s X is the optical center of the camera in the optical 3D sensor. s The axis is parallel to the horizontal direction of the image and points to the right, Y sThe axis is parallel to the vertical direction of the image and points downwards, Z. s The axis is the optical axis direction, and the coordinate system of the optical three-dimensional sensor conforms to the right-hand rule;

[0034] S404: Calculate O using the coordinates of feature points on the dot array planar target in the image acquired by the optical 3D sensor. t -X t Y t Z t coordinate system to O s -X s Y s Z s Rotation matrix of coordinate system Translation vector

[0035] S405: Calculate O using the coordinates of feature points on the dot array planar target in the camera-captured image. t -X t Y t Z t coordinate system to O c -X c Y c Z c Rotation matrix of coordinate system Translation vector

[0036] S406: Calculate O using the coordinates of feature points on a mirrored stereo target in the camera-captured image. st -X st Y st Z st coordinate system to O c -X c Y c Z c Rotation matrix of coordinate system Translation vector

[0037] S407: From formula (2):

[0038]

[0039] Solve O st -X st Y st Z st coordinate system to O s -X s Y s Z s Rotation matrix of coordinate system Translation vector

[0040] Furthermore, S5 includes:

[0041] S501: Establish global coordinate system O g -X g Y g Z g , origin O g For the camera's optical center, X g The axis is parallel to the horizontal direction of the image and points to the right, Y g The axis is parallel to the vertical direction of the image and points downwards, Z. g The axis is the optical axis, and the global coordinate system conforms to the right-hand rule.

[0042] S502: Calculate O using the coordinates of feature points on a mirrored stereo target in the camera-captured image. st -X st Y st Z st coordinate system to O g -X g Y g Z g Rotation matrix of coordinate system Translation vector

[0043] S503: Based on the relative pose of the mirror-type stereo target and the optical three-dimensional sensor, from formula (3):

[0044]

[0045] Solve O s -X s Y s Z s coordinate system to O g -X g Y g Z g Rotation matrix of coordinate system Translation vector The vision system can be used to locate the target data and the 3D data acquired by the optical 3D sensor at different locations can be globally unified.

[0046] The beneficial effects of this invention are:

[0047] (1) This invention designs a novel three-dimensional target, which is composed of a planar target and a planar mirror. By utilizing the principle of plane mirror reflection imaging, a single planar target can be expanded into a three-dimensional target, which greatly reduces the manufacturing cost. The mirror-type three-dimensional target is easy to assemble and disassemble, and does not require complex machining, thus solving the problems of complex structure, difficult processing, low yield, difficulty in storage, and high cost of traditional three-dimensional targets.

[0048] (2) This invention combines computer vision and image processing technology to propose a vision system localization method based on a mirror-type stereo target. By unifying the coordinates of feature points of the stereo target, solving the pose of the stereo target and the optical 3D sensor, and localizing the optical 3D sensor, the 3D point cloud obtained by the optical 3D sensor at different positions can be transformed into the global coordinate system, realizing the global unification of multi-view measurement data. This vision system localization method is universal for any stereo target and is not limited to a specific shape. It is applicable to a variety of measurement scenarios.

[0049] (3) The present invention fixes a mirror-type stereo target on an optical three-dimensional sensor. Combined with the designed stereo target and the vision system positioning algorithm, it can be used for the integrity measurement of objects. Attached Figure Description

[0050] 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 embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0051] Figure 1 The overall flowchart of the vision system localization method based on a mirrored stereo target provided by the present invention;

[0052] Figure 2 This is a schematic diagram of a planar target provided by the present invention;

[0053] Figure 3 This is a schematic diagram of a mirror-type three-dimensional target provided by the present invention;

[0054] Figure 4 A schematic diagram illustrating the process of transforming feature points on a virtual planar target from the virtual planar target coordinate system to the three-dimensional target coordinate system provided by the present invention;

[0055] Figure 5 This is a schematic diagram of the assembly of the mirror-type stereo target and the optical three-dimensional sensor provided by the present invention;

[0056] Figure 6 A schematic diagram illustrating the process of transforming feature points on a mirrored stereo target from the stereo target coordinate system to the optical three-dimensional sensor coordinate system provided by the present invention;

[0057] Figure 7 This is a schematic diagram illustrating the process of converting three-dimensional data from different measurement coordinate systems to the global coordinate system, as provided by the present invention. Detailed Implementation

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

[0059] like Figure 1 As shown, this invention discloses a visual system localization method based on a mirrored stereo target, comprising the following steps:

[0060] Step 1: Design a mirror-shaped 3D target:

[0061] like Figure 2 As shown, the planar target is made of ceramic material. The four feature patterns on it are circular patterns with coded features, namely one circle, two concentric circles, three concentric circles, and four concentric circles. The feature points are the centers of the four circular patterns and the common tangent points of adjacent circular patterns.

[0062] like Figure 3 As shown, the mirror-type stereo target consists of a planar target and a planar reflector, where the planar reflector is an aluminum-coated forward-facing mirror. The planar target is reflected by the planar reflector to form a virtual planar target. The planar target and the virtual planar target are symmetrical about the planar reflector. The line connecting corresponding feature points on the two targets is perpendicular to the planar reflector, and the distances from the corresponding feature points to the planar reflector are equal.

[0063] Step 2: Establish the 3D target coordinate system and the virtual planar target coordinate system:

[0064] Establish a three-dimensional target coordinate system O based on the planar target. st -X st Y st Z st , origin O st Let X be the center of the three concentric circles on the planar target. st The axis points to the center of the four concentric circles on the planar target, Y. st The axis points to the center of the two concentric circles on the planar target, Z. st The axis is perpendicular to the target plane, Z st The positive direction of the axis is determined according to the right-hand rule.

[0065] A virtual plane target coordinate system O is established based on the virtual plane target. stv -X stv Y stv Z stv , origin Os tv Let X be the center of the three concentric circles on the virtual planar target. stvThe axis points to the center of the four concentric circles on the virtual planar target, Y. stv The axis points to the center of two concentric circles on the virtual planar target, Z. stv The Z axis is perpendicular to the virtual target plane. stv The positive direction of the axis is determined according to the right-hand rule.

[0066] Step 3: Unify the coordinates of the feature points of the 3D target:

[0067] like Figure 4 As shown, a mirrored stereo target is placed within the pre-calibrated camera's field of view, and the camera captures a clear image of the stereo target. A camera coordinate system O is established. c -X c Y c Z c , origin O c For the camera's optical center, X c The axis is parallel to the horizontal direction of the image and points to the right, Y c The axis is parallel to the vertical direction of the image and points downwards, Z. c The axis is the optical axis, and the camera coordinate system conforms to the right-hand rule.

[0068] Using the coordinates m of feature points on a planar target in the image c And the coordinates M in the three-dimensional target coordinate system st Solve for O according to formula (1) st -X st Y st Z st coordinate system to O c -X c Y c Z c Rotation matrix of coordinate system Translation vector

[0069]

[0070]

[0071] Where λ is the scale factor. and It is m c and M st homogeneous coordinates, A c These are internal parameters of the camera, determined by the effective focal length (f). x f y It consists of the coordinates of the principal point (u0, u0) and the principal point coordinates (u0, u0).

[0072] Establish virtual camera coordinate system O v -X v Yv Z v The virtual camera coordinate system is symmetrical to the camera coordinate system about the plane mirror, and the virtual camera coordinate system conforms to the left-hand rule.

[0073] Establish a virtual camera coordinate system O that conforms to the right-hand rule. vr -X vr Y vr Z vr , origin O vr Y vr Axis, Z vr The axis and the origin O in the virtual camera coordinate system v Y v Axis, Z v Axis consistent, X vr Axial direction and X v The axes are in opposite directions.

[0074] The image is horizontally flipped, and the coordinates m of the feature points on the virtual plane target in the flipped image are used. vr According to formula (3), calculate O st -X st Y st Z st coordinate system to O vr -X vr Y vr Z vr Rotation matrix of coordinate system Translation vector

[0075]

[0076]

[0077] in, It is m vr homogeneous coordinates, A vr These are the internal parameters of the virtual camera, where W is the width of the image pixels.

[0078] Based on the relative poses of the planar target and the virtual camera, and the fact that the relative poses of the virtual planar target and the camera are the same, O stv -X stv Y stv Z stv coordinate system to O c -X c Y c Z c Rotation matrix of coordinate system Translation vector and and same.

[0079] Solve for O according to formula (5) stv -X stv Y stv Z stv coordinate system to O st -X st Y st Z st Rotation matrix of coordinate system Translation vector The feature points M on the virtual plane target are calculated according to formula (6). stv Coordinate M in the three-dimensional target coordinate system st2 Thus, the coordinates {M} of each feature point on the mirrored 3D target in the 3D target coordinate system are obtained. st M st2}

[0080]

[0081]

[0082] Step 4: Solve for the pose of the stereo target and the optical 3D sensor:

[0083] like Figure 5 As shown, a mirror-shaped stereo target is fixed onto an optical 3D sensor. Figure 6 As shown, a dot array planar target and a pre-calibrated camera are used. By adjusting the positions of the optical 3D sensor and the dot array planar target, the dot array planar target is placed within the field of view of the optical 3D sensor and the camera, so that at least two feature patterns on the mirrored stereo target are within the field of view of the camera. The optical 3D sensor and the camera each acquire a clear image.

[0084] Establish the coordinate system O of the dot array planar target t -X t Y t Z t , origin O t X is the center of the upper left feature point on the dot array planar target. t Axis, Y t The axes point to the horizontal and vertical directions of the dot array planar target, respectively, Z. t The axis is perpendicular to the plane of the dot array target, and the coordinate system of the dot array target plane conforms to the right-hand rule.

[0085] Establish the optical three-dimensional sensor coordinate system O s -X s Y s Z s , origin O s X is the optical center of the camera in the optical 3D sensor.s The axis is parallel to the horizontal direction of the image and points to the right, Y s The axis is parallel to the vertical direction of the image and points downwards, Z. s The axis is the optical axis direction, and the coordinate system of the optical three-dimensional sensor conforms to the right-hand rule.

[0086] By utilizing the coordinates of feature points on a dot array planar target in the image acquired by an optical 3D sensor, the O-values ​​can be calculated. t -X t Y t Z t coordinate system to O s -X s Y s Z s Rotation matrix of coordinate system Translation vector

[0087] Using the coordinates of feature points on the dot array planar target in the camera-captured image, solve for O t -X t Y t Z t coordinate system to O c -X c Y c Z c Rotation matrix of coordinate system Translation vector

[0088] Using the coordinates of feature points on a mirrored stereo target in the camera-captured image, the O is calculated. st -X st Y st Z st coordinate system to O c -X c Y c Z c Rotation matrix of coordinate system Translation vector

[0089] Solve for O according to formula (7) st -X st Y st Z st coordinate system to O s -X s Y s Z s Rotation matrix of coordinate system Translation vector

[0090]

[0091] Step 5: Position the optical 3D sensor:

[0092] like Figure 7 As shown, a pre-calibrated camera is fixed above the experimental platform, and a global coordinate system O is established. g -X g Y g Z g , origin O g For the camera's optical center, X g The axis is parallel to the horizontal direction of the image and points to the right, Y g The axis is parallel to the vertical direction of the image and points downwards, Z. g The axis is the optical axis, and the global coordinate system conforms to the right-hand rule.

[0093] A freely movable optical 3D sensor is used to capture images of a mirror-image stereo target using a camera. At least two feature patterns must exist within the camera's field of view. The coordinates of the feature points on the mirror-image stereo target within the camera's captured image are used to calculate O. st -X st Y st Z st coordinate system to O g -X g Y g Z g Rotation matrix of coordinate system Translation vector

[0094] Based on the relative pose of the mirror-type stereo target and the optical three-dimensional sensor, O is calculated according to formula (8). s -X s Y s Z s coordinate system to O g -X g Y g Z g Rotation matrix of coordinate system Translation vector By unifying the 3D point clouds acquired by optical 3D sensors at different locations into a global coordinate system, the integrity of objects can be measured.

[0095]

[0096] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A visual system localization method based on a mirror-type stereo target, characterized in that, Includes the following steps: S1: Design a mirror-type stereo target, which consists of a planar target and a planar reflector. The planar target has four different feature patterns. The design of a mirror-type 3D target specifically includes: S101: The planar target is made of ceramic material. The four feature patterns on the planar target are circular patterns with coded features, namely one circle, two concentric circles, three concentric circles, and four concentric circles. The feature points on the planar target are the centers of the four circular patterns and the common tangent points of adjacent circular patterns. S102: Set the plane mirror as an aluminum-plated front mirror; S103: A planar target is reflected by a planar mirror to form a virtual planar target. The planar target and the virtual planar target are symmetrical about the planar mirror. The line connecting the corresponding feature points on the two targets is perpendicular to the planar mirror, and the distances from the corresponding feature points to the planar mirror are equal. S2: Establish a three-dimensional target coordinate system and a virtual planar target coordinate system respectively, and determine the origin position and coordinate axis direction of each coordinate system; S3: Solve the relative pose between the virtual planar target and the planar target to obtain the coordinates of each feature point on the mirrored 3D target in the 3D target coordinate system, including: S301: Place the mirror-type stereo target within the pre-calibrated field of view of the camera, and have the camera capture a clear stereo target image; S302: Establish camera coordinate system ,origin For the camera's optical center, The axis is parallel to the horizontal direction of the image and points to the right. The axis is parallel to the vertical direction of the image and points downwards. The axis is the optical axis, and the camera coordinate system conforms to the right-hand rule; S303: Solve the following by using the coordinates of feature points on the planar target in the image and their coordinates in the three-dimensional target coordinate system. coordinate system to Rotation matrix of coordinate system Translation vector ; S304: Establish the virtual camera coordinate system The virtual camera coordinate system is symmetrical to the camera coordinate system about the plane mirror, and the virtual camera coordinate system conforms to the left-hand rule; S305: Establish a virtual camera coordinate system that conforms to the right-hand rule. ,origin , axis, The axis and the origin in the virtual camera coordinate system , axis, Axis alignment Axial direction and The axes are in opposite directions; S306: Flip the image horizontally, and use the coordinates of feature points on the virtual plane target in the flipped image to calculate... coordinate system to Rotation matrix of coordinate system Translation vector ; S307: Based on the relative pose of the planar target and the virtual camera, and the fact that the relative pose of the virtual planar target and the camera are the same, coordinate system to Rotation matrix of coordinate system Translation vector and and same; S308: Solve according to formula (1) coordinate system to Rotation matrix of coordinate system Translation vector : (1) Solve the coordinates of feature points on the virtual planar target in the coordinate system of the three-dimensional target; S4: Fix the mirror-type stereo target on the optical 3D sensor and calculate the relative pose between the optical 3D sensor and the mirror-type stereo target; S5: Fix a calibrated camera above the experimental platform, establish a global coordinate system, freely move the optical 3D sensor, use the camera to photograph a mirrored stereo target, and ensure that there are at least two feature patterns in the camera's field of view. Use the stereo target image acquired by the camera to locate the optical 3D sensor, and globally unify the 3D data acquired by the optical 3D sensor at different locations.

2. The visual system localization method based on a mirror-type stereo target according to claim 1, characterized in that, In step S2, establishing the three-dimensional target coordinate system and the virtual planar target coordinate system specifically includes: S201: Establishing a three-dimensional target coordinate system based on a planar target ,origin The center of the three concentric circles on the planar target. The axis points to the center of the four concentric circles on the planar target. The axis points to the center of the two concentric circles on the planar target. The axis is perpendicular to the target plane. The positive direction of the axis is determined according to the right-hand rule; S202: Establishing a virtual plane target coordinate system based on a virtual plane target. ,origin The center of the three concentric circles on the virtual planar target. The axis points to the center of the four concentric circles on the virtual planar target. The axis points to the center of the two concentric circles on the virtual planar target. The axis is perpendicular to the virtual target plane. The positive direction of the axis is determined according to the right-hand rule.

3. The visual system localization method based on a mirrored stereo target according to claim 1, characterized in that, S4 includes: S401: Using a dot array planar target and a pre-calibrated camera, by adjusting the position of the optical 3D sensor and the dot array planar target, the dot array planar target is placed within the field of view of the optical 3D sensor and the camera, so that at least two feature patterns on the mirrored stereo target are within the field of view of the camera, and the optical 3D sensor and the camera respectively acquire a clear image. S402: Establish the coordinate system of the dot array planar target ,origin The center of the feature point at the upper left corner of the dot array planar target. axis, The axes point to the horizontal and vertical directions of the dot array planar target, respectively. The axis is perpendicular to the plane of the dot array target, and the coordinate system of the dot array target plane conforms to the right-hand rule. S403: Establish the coordinate system of the optical three-dimensional sensor ,origin The optical center of the camera in the optical 3D sensor. The axis is parallel to the horizontal direction of the image and points to the right. The axis is parallel to the vertical direction of the image and points downwards. The axis is the optical axis direction, and the coordinate system of the optical three-dimensional sensor conforms to the right-hand rule; S404: Calculate the coordinates of feature points on the dot array planar target in the image acquired by the optical 3D sensor. coordinate system to Rotation matrix of coordinate system Translation vector ; S405: Calculate the coordinates of feature points on the dot array planar target within the camera-captured image. coordinate system to Rotation matrix of coordinate system Translation vector ; S406: Solve the coordinates of feature points on a mirrored stereo target in the camera-captured image by utilizing the coordinates of these points. coordinate system to Rotation matrix of coordinate system Translation vector ; S407: From formula (2): (2) Solution coordinate system to Rotation matrix of coordinate system Translation vector .

4. The visual system localization method based on a mirror-type stereo target according to claim 1, characterized in that, S5 includes: S501: Establish a global coordinate system ,origin For the camera's optical core, The axis is parallel to the horizontal direction of the image and points to the right. The axis is parallel to the vertical direction of the image and points downwards. The axis is the optical axis, and the global coordinate system conforms to the right-hand rule. S502: Calculate the coordinates of feature points on a mirrored stereo target within the camera-captured image. coordinate system to Rotation matrix of coordinate system Translation vector ; S503: Based on the relative pose of the mirror-type stereo target and the optical three-dimensional sensor, from formula (3): (3) Solution coordinate system to Rotation matrix of coordinate system Translation vector This enables the positioning of the vision system and globally unifies the 3D data acquired by the optical 3D sensor at different locations.

Images