A circular ring mark for visual measurement and high-precision extraction and coding method

Abstract

This invention provides a circular marker for visual measurement and a high-precision extraction and encoding / decoding method. The circular marker includes a square outer frame and a circular ring. The circular ring is disposed inside the square outer frame, and the outer frame and the circular ring have the same center. The ratio of the side length of the square outer frame to the radius of the larger circle and the radius of the smaller circle of the circular ring is r. 1 :r 2 :r 3 The circular ring contains a clockwise embedded binary code. The circular marker and its high-precision extraction and encoding / decoding method for visual measurement provided by this invention balances the accurate detection of control points on images with the issue of self-carrying encoding, and can be used to solve camera pose with high precision.

Description

A circular marker for visual measurement and a high-precision extraction and encoding / decoding method. Technical Field

[0001] This invention relates to the field of computer vision technology, and in particular to a circular marker for visual measurement and a high-precision extraction and encoding / decoding method. Background Technology

[0002] Camera pose estimation is a fundamental task in computer vision, widely applied in fields such as autonomous driving, remote sensing and mapping, and intelligent robotics. By obtaining the correspondence between the coordinates of known landmarks in the environment and their projections onto an image, the current pose of the camera can be estimated. Landmark designs can be categorized into two types based on whether they possess coded identifiers: one type, represented by TagCC, is a ring-shaped marker with control points at its center. The projection of these control points onto the image can be accurately obtained through least-squares ring fitting. However, due to the lack of coded information, other prior knowledge is needed to determine the correspondence between the center and control points in the image. The other type, represented by Aruco, is a rotation-resistant square marker with the code embedded within the square. It provides the coordinates of four vertices, but the projected pixel coordinates of the vertices are obtained using a corner sub-pixel refinement algorithm, resulting in relatively low accuracy compared to the center fitting algorithm. Therefore, designing a ring-shaped marker for visual measurement and a high-precision extraction and encoding / decoding method is essential. Summary of the Invention

[0003] The purpose of this invention is to provide a circular marker for visual measurement and a high-precision extraction and encoding / decoding method, which takes into account both the accurate detection of control points on images and the problem of self-carrying encoding, and can be used to solve camera pose with high precision.

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

[0005] A circular marker for visual measurement includes: a square outer frame and a circular ring, wherein the circular ring is disposed inside the square outer frame, and the outer frame and the circular ring have the same center. The ratio of the side length of the square outer frame to the radius of the large circle and the radius of the small circle of the circular ring is preset to r1:r2:r3. The circular ring has a clockwise embedded binary code.

[0006] This invention also provides a high-precision extraction and encoding / decoding method for circular markers used in visual measurement, applied to the aforementioned circular markers used in visual measurement, comprising the following steps:

[0007] Step 1: Obtain an image of the circular logo and extract its approximate quadrilateral outline;

[0008] Step 2: Correct the extracted quadrilateral outline to resemble a regular quadrilateral;

[0009] Step 3: Perform ring detection on the corrected contour;

[0010] Step 4: After the test is completed, refine the parameters of the annulus ring;

[0011] Step 5: After refinement, unfold the ring.

[0012] Step 6: After expansion, perform cyclic binary encoding extraction;

[0013] Step 7: Determine the identity of the circular marker based on the extracted cyclic binary code.

[0014] Optionally, in step 1, the image of the circular marker is acquired, and its approximate quadrilateral contour is extracted. This specifically includes the following steps:

[0015] Step 101: Obtain the edge information of image I of the circular marker. Use the Canny operator to extract the edge information in image I. Based on the edge detection results, use a quadrilateral contour fitting algorithm to store the detected contours in the set C = {c1, c2, ..., c...} n Within}, where c i This represents a closed quadrilateral outline, containing the pixel coordinates of its four vertices: p1, p2, p3, and p4.

[0016] Step 102: For each quadrilateral profile c i Calculate the side length of the quadrilateral using the following formula. in, Represents vertex p i With p j Pixel distance on image I;

[0017]

[0018] In the formula, x i With y i They are points p i The x and y coordinate components in the pixel coordinate system;

[0019] Step 103: Calculate the ratio of the longest side length to the shortest side length of the quadrilateral. If ρ is greater than the threshold of 0.8, the quadrilateral is considered a regular quadrilateral and the extracted regular quadrilateral outline is obtained; otherwise, the quadrilateral is discarded.

[0020] Optionally, in step 2, the extracted quadrilateral contour is corrected to resemble a regular quadrilateral, specifically as follows:

[0021] Create a square image block I′ with width W. The gray value I′(p′) of I′ at pixel coordinate p′ is taken from the gray value I(p) of image I at pixel coordinate p. The relationship between pixel coordinates p′ and p is established through homography transformation p = λHp′, where H is calculated using the formula:

[0022]

[0023] Among them, (x i ,y i (x′) represents the pixel coordinates of vertex p′ of the quadrilateral outline in image I. i ,y′ i ) is the pixel coordinate (x) of the corresponding vertex in the square image block I′. i ,y i ), where λ is the normalization scaling factor, h ij It is the element in the i-th row and j-th column of the homography matrix.

[0024] Optionally, in step 3, a ring detection is performed on the corrected contour, specifically including the following steps:

[0025] Step 301: Based on the geometric prior of the circular symbol, namely that the square border and the circular ring have the same geometric center and the ratio of the side length of the square border to the outer diameter to the inner diameter of the circular ring is r1:r2:r3, then... Let I be the pixel coordinates of the center of the annulus in the square image block I′, and the initial values ​​of the outer and inner diameters of the annulus, respectively.

[0026] Step 302: Use the Canny operator to extract the edge information in the square image block I′, and statistically analyze the response values ​​of the large and small circles using the following formula:

[0027] E(x,y,a,b,r)=(xa) 2 +(yb) 2 -r 2 (4) (3)

[0028]

[0029]

[0030] In the formula, s1 and s2 are the response values ​​of the large circle and the small circle, respectively, and φ and Let be the radii of the large circle and the small circle, respectively. Here, I′(x,y)∈D indicates that the square image patch I′ is detected as edge information by the Canny operator at pixel coordinates (x,y). T[condition] is 1 if the condition is true, and 0 otherwise. The coordinates (x, y) on the image were measured to The error of a circle with center φ and radius φ;

[0031] Step 303: Estimate the probability p of the presence of a ring in the image as follows:

[0032]

[0033] When p is greater than the threshold of 0.8, the pattern is considered to be a circular symbol.

[0034] Optionally, in step 4, after the detection is completed, the ring parameters are refined, specifically as follows:

[0035] The parameters of the annulus are refined based on the loss function to obtain the precise pixel coordinates of the center of the annulus marker and the inner and outer diameters of the annulus. The loss function is:

[0036]

[0037] In the formula, (a,b), φ, These are the values ​​to be optimized for the center, outer diameter, and inner diameter of the annulus.

[0038] Optionally, in step 5, after refinement, the ring is unfolded, specifically as follows:

[0039] For the refined square image block I′, establish a coordinate system xoy parallel to the pixel coordinate system with (a,b) as the origin, and expand the annulus into a bar chart based on the following formula with the x-axis as the starting point:

[0040]

[0041] The expanded bar image is represented by a matrix M with W columns and H rows, where M is the element of the matrix with W columns and h rows. hw This represents the grayscale value of the bar image at (h,w).

[0042] Optionally, in step 6, after expansion, a cyclic binary encoding extraction is performed, specifically including the following steps:

[0043] Step 601: Summate matrix M along the vertical direction to obtain row vector m, where m i This represents the i-th element of vector m;

[0044] Step 602: Solve for the gradient vector m′=(m2-m1,m3-m2,...,m1-m n ), where m′ i =m i+1 -m iBreak m at the t-th element and reassemble it to obtain m * =(m t ,m t+1 ,...,m n ,m1,m2,...,m t-1 ), where t is the index of the largest element in m′;

[0045] Step 603: Construct a row vector c = (c1, c2, ..., cn) of size N. N ), where c i The calculation formula is:

[0046]

[0047] Optionally, in step 7, the identity identifier of the circular marker is determined based on the extracted cyclic binary code, specifically as follows:

[0048] Perform Otsu binarization on the row vector c to obtain the encoding a1a2...a within the flag. N And obtain the identifier of the flag id = min(p0, p1, ..., p N-1 ), where p0 = a1a2...a N p i This indicates that p0 is shifted left by i bits. i =a i+1 a i+ 2...a n a1a2...a i .

[0049] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects: The circular marker for visual measurement and its high-precision extraction and encoding / decoding method provided by the present invention uses a circular ring as a carrier, embeds a binary cyclic code, has rotation invariance, and sets the control point as the center of the ring. It can be extracted with high precision in the image, and each marker has a unique identity. Multiple markers can be placed in complex scenes, and the extraction and matching of control points can be completed solely through the marker itself. Compared with traditional markers, it takes into account both the accurate detection of control points in the image and the problem of carrying its own encoding. The method includes acquiring the image of the circular marker and approximating its quadrilateral contour, correcting the extracted quadrilateral contour, performing circular detection on the corrected contour, refining the circular parameters after detection, unfolding the circular marker after refinement, extracting the cyclic binary code after unfolding, and determining the identity of the circular marker based on the extracted cyclic binary code. It can solve the camera pose with high precision in complex environments. 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 embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0051] Figure 1 shows the circular symbol with ID 00010101001;

[0052] Figure 2 is a schematic flowchart of a high-precision extraction and encoding / decoding method for circular marks used in visual measurement according to an embodiment of the present invention.

[0053] Figure 3 is a schematic diagram of homography transformation point pairs;

[0054] Figure 4a is a schematic diagram of the physical unfolding of the marker;

[0055] Figure 4b is a schematic diagram of the logo expansion logic. Detailed Implementation

[0056] The purpose of this invention is to provide a circular marker for visual measurement and a high-precision extraction and encoding / decoding method, which takes into account both the accurate detection of control points on images and the problem of self-carrying encoding, and can be used to solve camera pose with high precision.

[0057] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0058] The circular marker for visual measurement provided in this embodiment of the invention includes: a square outer frame and a circular ring. The circular ring is disposed inside the square outer frame, and the outer frame and the circular ring have the same center. The ratio of the side length of the square outer frame to the radius of the large circle and the radius of the small circle of the circular ring is r1:r2:r3. The circular ring has a clockwise embedded binary code.

[0059] As shown in Figure 1, a self-encoded flag carrying a bit cyclic binary code is displayed, and its identity information id is 00010101001.

[0060] As shown in Figure 2, the present invention also provides a high-precision extraction and encoding / decoding method for circular markers used in visual measurement, applied to the aforementioned circular markers used in visual measurement, comprising the following steps:

[0061] Step 1: Obtain an image of the circular logo and extract its approximate quadrilateral outline;

[0062] Step 2: Correct the extracted quadrilateral outline to resemble a regular quadrilateral;

[0063] Step 3: Perform ring detection on the corrected contour;

[0064] Step 4: After the test is completed, refine the parameters of the annulus ring;

[0065] Step 5: After refinement, unfold the ring.

[0066] Step 6: After expansion, perform cyclic binary encoding extraction;

[0067] Step 7: Determine the identity of the circular marker based on the extracted cyclic binary code.

[0068] In step 1, the image of the circular marker is acquired, and its approximate quadrilateral contour is extracted. This includes the following steps:

[0069] Step 101: Obtain the edge information of image I of the circular marker. Use the Canny operator to extract the edge information in image I. Based on the edge detection results, use a quadrilateral contour fitting algorithm to store the detected contours in the set C = {c1, c2, ..., c...} n Within}, where c i This represents a closed quadrilateral outline, containing the pixel coordinates of its four vertices: p1, p2, p3, and p4.

[0070] Step 102: For each quadrilateral profile c i Calculate the side length of the quadrilateral using the following formula. in, Represents vertex p i With p j Pixel distance on image I;

[0071]

[0072] In the formula, x i With y i They are points p i The x and y coordinate components in the pixel coordinate system;

[0073] Step 103: Calculate the ratio of the longest side length to the shortest side length of the quadrilateral. If ρ is greater than the threshold of 0.8, the quadrilateral is considered a regular quadrilateral and the extracted regular quadrilateral outline is obtained; otherwise, the quadrilateral is discarded.

[0074] In step 2, the extracted square outline is corrected to resemble a regular square, specifically as follows:

[0075] Create a square image block I′ with width W. The gray value I′(p′) of I′ at pixel coordinate p′ is taken from the gray value I(p) of image I at pixel coordinate p. The relationship between pixel coordinates p′ and p is established through homography transformation p = λHp′, where H is calculated using the formula:

[0076]

[0077] Among them, (x i ,y i ) is the pixel coordinate of vertex p′ of the quadrilateral contour in image I, (x i ′,y i ′) is the pixel coordinate (x, y) of the corresponding vertex in the square image block I′. i ,y i ), where λ is the normalization scaling factor, h ij It is the element in the i-th row and j-th column of the homography matrix.

[0078] As shown in Figure 3, Figure 3 illustrates the contour c. i When performing geometric deformation correction on the pattern, c i The correspondence between the vertices of the square image block I′.

[0079] Step 3 involves performing ring detection on the corrected contour, specifically including the following steps:

[0080] Step 301: Based on the geometric prior of the circular symbol, namely that the square border and the circular ring have the same geometric center and the ratio of the side length of the square border to the outer diameter to the inner diameter of the circular ring is r1:r2:r3, then... Let I be the pixel coordinates of the center of the annulus in the square image block I′, and the initial values ​​of the outer and inner diameters of the annulus, respectively.

[0081] Step 302: Use the Canny operator to extract the edge information in the square image block I′, and statistically analyze the response values ​​of the large and small circles using the following formula:

[0082] E(x,y,a,b,r)=(xa) 2 +(yb) 2 -r 2 (4) (3)

[0083]

[0084]

[0085] In the formula, s1 and s2 are the response values ​​of the large circle and the small circle, respectively, and φ and Let be the radii of the large circle and the small circle, respectively. Here, I′(x,y)∈D indicates that the square image patch I′ is detected as edge information by the Canny operator at pixel coordinates (x,y). T[condition] is 1 if the condition is true, and 0 otherwise. The coordinates (x, y) on the image were measured to The error of a circle with center φ and radius φ;

[0086] Step 303: Estimate the probability p of the presence of a ring in the image as follows:

[0087]

[0088] When p is greater than the threshold of 0.8, the pattern is considered to be a circular symbol.

[0089] In step 4, after the inspection is completed, the parameters of the ring are refined, specifically as follows:

[0090] The parameters of the annulus are refined based on the loss function to obtain the precise pixel coordinates of the center of the annulus marker and the inner and outer diameters of the annulus. The loss function is:

[0091]

[0092] In the formula, (a,b), φ, These are the values ​​to be optimized for the center, outer diameter, and inner diameter of the annulus.

[0093] In step 5, after refinement, the ring is unfolded, specifically as follows:

[0094] For the refined square image block I′, a coordinate system xoy parallel to the pixel coordinate system is established with (a,b) as the origin. Based on the following formula, the annulus is unfolded into a bar chart with the x-axis as the starting point, as shown in Figures 4a and 4b:

[0095]

[0096] The expanded bar image is represented by a matrix M with W columns and H rows, where M is the element of the matrix with W columns and h rows. hw This represents the grayscale value of the bar image at (h,w).

[0097] In step 6, the expanded binary code is extracted cyclically, which includes the following steps:

[0098] Step 601: Summate matrix M along the vertical direction to obtain row vector m, where m i This represents the i-th element of vector m;

[0099] Step 602: Solve for the gradient vector m′=(m2-m1,m3-m2,...,m1-m n ), where m i ′=m i+1 -m i Break m at the t-th element and reassemble it to obtain m * =(m t ,m t+1 ,...,m n ,m1,m2,...,m t-1 ), where t is the index of the largest element in m′;

[0100] Step 603: Construct a row vector c = (c1, c2, ..., cn) of size N. N ), where c i The calculation formula is:

[0101]

[0102] In step 7, the identity identifier of the circular marker is determined based on the extracted cyclic binary code, specifically as follows:

[0103] Perform Otsu binarization on the row vector c to obtain the encoding a1a2...a within the flag. N And obtain the identifier of the flag id = min(p0, p1, ..., p N-1 ), where p0 = a1a2...a N p i This indicates that p0 is shifted left by i bits. i =a i+1 a i+ 2...a n a1a2...a i .

[0104] This invention provides an embodiment that details the above method. First, the Canny operator is used to extract edge information from image I. Based on the edge detection results, a quadrilateral contour fitting algorithm is used to store the detected contours in a set C = {c1, c2, ..., c...}. n Within}, where c i This represents a closed quadrilateral outline, containing the pixel coordinates of its four vertices: p1, p2, p3, and p4.

[0105] For each quadrilateral contour c i Calculate the side length of the quadrilateral using formula (2). in, Represents vertex p i With p jCalculate the ratio ρ of the maximum side length to the minimum side length of the quadrilateral based on the pixel distance in image I. If ρ is greater than the threshold of 0.8, the quadrilateral is considered to be a regular quadrilateral, and the extracted regular quadrilateral outline is obtained; otherwise, the quadrilateral is discarded.

[0106] Create a square image block I′ with width W. The gray value I′(p′) of I′ at pixel coordinate p′ is taken from the gray value I(p) of image I at pixel coordinate p. The relationship between pixel coordinate p′ and p is established by homography transformation p=λHp′. The calculation formula of H is shown in formula (2).

[0107] Will Let the pixel coordinates of the center of the ring in the square image block I′, the initial values ​​of the outer diameter and the inner diameter of the ring be assumed respectively; use the Canny operator to extract the edge information in the square image block I′, and use formulas (4) and (5) to statistically analyze the response values ​​of the large circle and the small circle, and use formula (6) to estimate the probability p of the ring in the image. When p is greater than the threshold 0.8, the pattern is considered to be a ring mark.

[0108] The parameters of the ring are refined according to formula (7) to obtain the accurate pixel coordinates of the center of the ring mark and the inner and outer diameters of the ring.

[0109] For the refined square image patch I′, a coordinate system xoy parallel to the pixel coordinate system is established with (a,b) as the origin. Based on the following formula, the annulus is expanded into a bar chart with the x-axis as the starting point. The expanded bar image is represented by a matrix M with W columns and H rows, where the elements M of the matrix are W columns and h rows. hw This represents the grayscale value of the bar image at (h,w).

[0110] Summing matrix M along the vertical direction yields row vector m, where m i This represents the i-th element of vector m;

[0111] Find the gradient vector m′=(m2-m1,m3-m2,...,m1-m n ), where m i ′=m i+1 -m i Break m at the t-th element and reassemble it to obtain m * =(m t ,m t+1 ,...,m n ,m1,m2,...,m t-1 ), where t is the index of the largest element in m′;

[0112] The row vector c = (c1, c2, ..., c3) is calculated using formula (9). N );

[0113] Perform Otsu binarization on the row vector c to obtain the encoding a1a2...a within the flag. N And obtain the identifier of the flag id = min(p0, p1, ..., p N-1 ), where p0 = a1a2...a N p i This indicates that p0 is shifted left by i bits. i =a i+1 a i+ 2...a n a1a2...a i The final decoding result of image I.

[0114] This invention provides a circular marker for visual measurement and a high-precision extraction and encoding / decoding method. The marker uses a circular ring as its carrier, embedding a binary cyclic code, exhibiting rotation invariance. The control point is set at the center of the ring, allowing for high-precision extraction from the image. Each marker has a unique identifier, enabling the placement of multiple markers in complex scenes. Control point extraction and matching are achieved solely through the marker itself. Compared to traditional markers, this method balances accurate detection of control points in the image with the issue of self-carried encoding. The method includes acquiring an image of the circular marker, approximating its quadrilateral contour, correcting the extracted quadrilateral contour, performing ring detection on the corrected contour, refining the ring parameters, unfolding the ring, extracting the cyclic binary code, and determining the identifier of the circular marker based on the extracted cyclic binary code. This allows for high-precision solution of camera pose in complex environments.

[0115] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A high-precision extraction and encoding / decoding method for circular markers used in visual measurement, characterized in that, The circular marker for visual measurement includes: a square outer frame and a circular ring. The circular ring is disposed inside the square outer frame, and the outer frame and the circular ring have the same center. The ratio of the side length of the square outer frame to the radius of the larger circle and the radius of the smaller circle of the circular ring is [value missing]. The inner ring contains a clockwise embedded binary code; the method includes the following steps: Step 1: Obtain an image of the circular mark and extract its approximate quadrilateral contour, specifically including the following steps: Step 101: Obtain the edge information of image I of the circular mark, use the Canny operator to extract the edge information in image I, and use a quadrilateral contour fitting algorithm based on the edge detection results to store the detected contours in a set. Inside, among which This represents the outline of a closed quadrilateral, containing the pixel coordinates of its four vertices. 、 、 、 Step 102: For each quadrilateral outline Calculate the side length of the quadrilateral using the following formula. 、 、 、 ,in, Represents vertices and Pixel distance on image I; (1) In the formula, and They are points The x and y coordinate components in the pixel coordinate system; Step 103: Calculate the ratio of the maximum side length to the minimum side length of the quadrilateral. ,like If the value is greater than the threshold of 0.8, the quadrilateral is considered a regular quadrilateral, and the extracted regular quadrilateral outline is obtained; otherwise, the quadrilateral is discarded. Step 2: Perform regular quadrilateral correction on the extracted regular quadrilateral outline, specifically: create a width... For square image blocks ， In pixel coordinates grayscale value at Taken from image I at pixel coordinates grayscale value at where pixel coordinates and The relationship between them is achieved through homography transformation Establish, The calculation formula is: (2) Among them, It is an image Vertex of the outline of a regular quadrilateral pixel coordinates, It is a square image block In and corresponding vertex pixel coordinates ， It is the normalized scaling factor. It is the first in the homography matrix line, number The elements of the column; Step 3: Perform ring detection on the corrected contour. In Step 3, the ring detection on the corrected contour specifically includes the following steps: Step 301: Based on the geometric prior of the ring mark, that is, the square border and the ring have the same geometric center and the ratio of the side length of the square border to the outer diameter of the ring to the inner diameter of the ring is . ,Will 、 、 Assuming they are square image blocks The pixel coordinates of the center of the ring, and the initial values ​​of the outer and inner diameters of the ring; Step 302: Extract the square image patch using the Canny operator. The edge information in the image is statistically analyzed using the following formula to calculate the response values ​​of the large and small circles: （3） （4） In formula (5), 、 These are the response values ​​for the large circle and the small circle, respectively. and Let be the radii of the great circle and the small circle, respectively. Represents a square image block In pixel coordinates The area was detected as edge information by the Canny operator. The value is 1 if the condition is true, and 0 otherwise. The coordinates on the image were measured. to With the center of the circle, Error of the circle with radius; Step 303: Estimate the probability of the presence of a ring in the image. for: (6) When If the value is greater than the threshold of 0.8, the pattern is considered a circular mark; Step 4: After detection, the circular mark parameters are refined, specifically: the circular mark parameters are refined according to the loss function to obtain the accurate center pixel coordinates of the circular mark and the inner and outer diameters of the circle, where the loss function is: In equation (7), ， ， These are the values ​​to be optimized for the center, outer diameter, and inner diameter of the ring, respectively; Step 5: After refinement, the ring is expanded, specifically: for the refined square image block... ,by Establish a coordinate system xoy parallel to the pixel coordinate system with the origin, and expand the annulus into a bar chart based on the following formula, with the x-axis as the starting point: (8) Use List Row matrix This represents the unfolded bar image, where the matrix... List row elements Representing bar images in The grayscale value at the location; Step 6: After expansion, perform cyclic binary encoding extraction, specifically including the following steps: Step 601: Along the vertical direction, the matrix Summing yields the row vectors ,in, Representing vectors The 1 element; Step 602: Solve gradient vector ,in, ,Will In the Each element is broken and reassembled to obtain... ,in, yes Find the index of the largest element in the middle; Step 603: Construct a row vector of size N. ,in, The calculation formula is: (9); Step 7: Determine the identity of the circular marker based on the extracted cyclic binary code.

2. The high-precision extraction and encoding / decoding method for circular marks used in visual measurement according to claim 1, characterized in that, In step 7, the identity identifier of the circular marker is determined based on the extracted cyclic binary code, specifically: for the row vector Perform Otsu binarization to obtain the code within the flag. and obtain the identification mark. ,in, ， Indicates will Shift left Bit, 。

Images