TW452644B - Measurement method of the spatial coordinates of objects - Google Patents

Abstract

The present invention provides a measurement method of the spatial coordinates of objects, especially a method to determine the accurate coordinates and traces of objects in three dimensional space, which has the following main features: measure a predetermined image data simultaneously by using a plural optical image sensing devices which have a known spatial relationship mutually to determine a reference coordinate, find the corresponding matrix of coordinates between the reference coordinates and the measurement device by iteration, obtain a transfer matrix by using each of the corresponding matrix and the position of the object to be tested projected to the visual plane, then calculate the image data obtained from these sensing devices with the transfer matrix during the measurement of the tested object, thereby the coordinate value of the tested object in the reference coordinate is obtained.

Description

A7 A7

V. Description of the invention (1) The present invention provides a method for measuring the coordinates of an object in space. In particular, it refers to a method for tracking the image data of an object in three-dimensional space, so as to correctly define the exact coordinates of the object to be measured and its tracker in three-dimensional space. Pressing, based on the characteristics of the light going straight and being not easily disturbed, the 5 2 image data of the object to be measured can be faithfully transmitted to a considerable distance through the optical signal under the condition of unobstructed. This kind of measurement of objects also currently has a computerized image tracking system, and the accuracy and quantitative measurement capabilities of the measurement are far superior to the physiological structure of the human body, but are limited by optical image sensing devices Both can only be set in a way parallel to the plane of the object to be measured. As shown in the first figure, the plane where the sensing device 10 is located is defined as the visual plane R, and the X axis and the gamma axis are parallel to the visual plane R. A Cartesian coordinate (X, γ, Z) indicates a three-dimensional three-dimensional space, and then the position Z-0 is defined as a focal plane F, and the distance between the focal plane F and the visual plane R is a focal length f. From this, it is clear that not only Zhao, which is located in the three-dimensional space, will be projected to the position m (U (), V ()) of the visual plane r, but all objects on the line connecting p and 15 m will be Projected at point m in a many-to-one manner, so ^ When this "depth" information is lost in the process of three-dimensional correspondence to two-dimensional, according to current computer vision technology, only image data from objects can be recorded in two dimensions. And analysis' can no longer reconstruct the 3D data. Therefore, the traditional image tracking system can only process the position information in the two -20-dimensional space through computer vision and image tracking technology, which obviously does not meet the actual needs. On the other hand, considering the mechanism of the human eye to observe the object, it can be found that when measuring the object in front, the optical image of the object to be measured is slightly different in the relevant positions of the fundus of the two eyes by the distance between the pupils of the two eyes. The perspective gap will be automatically triangulated through the visual area of the human brain, and it is estimated that the fourth paper size will be applicable to the national standard (CNS) A4 specification (210 X 297 mm) ------ ----- Install! — Order _! --- line (please read the notes on the back before filling this page)

45 2644 V. Description of the invention (2) Measure the distance between the object and itself. Therefore, if this mechanism can be simulated, the space coordinate value of the object to be measured can be successfully measured. However, generally, two sets of optical image settings are used to capture the image data of the object to be measured, and the three & positioning is performed by the coordinate conversion. Obtaining the reference coordinates of the object projection data to the corresponding matrix of the sensing device coordinate 5 is easier; but finding the inverse corresponding matrix from the corresponding matrix in the reverse direction may not necessarily have a solution. Not necessarily a solution 9 In view of this, the inventor has accumulated years of experience and actively researched to reveal that the present invention inputs pre-calibrated corrected image data into 10 sets of measurement devices that simulate this mechanism, so as to obtain relevant parameters and introduce the projection of the object to be measured To the position data of the visual plane, the spatial coordinate value of the object to be measured can be smoothly measured, and the three-dimensional spatial relationship of the original object to be measured can be reconstructed. Finally, the "method for measuring the coordinates of objects in space" of the present invention is finally produced. It is an object of the present invention to provide a method for accurately locating 15-dimensional coordinates of a three-dimensional object to be measured. Another object of the present invention is to provide a measurement method for real-time measurement of the three-dimensional space coordinates of a test object. Yet another object of the present invention is to provide a measurement method capable of measuring the trajectory of an object to be measured in a three-dimensional space. 20 The main feature of the present invention is that a predetermined image data is measured synchronously by using a plurality of optical image sensing devices whose spatial relationships are known to define a reference coordinate, and the reference coordinate and the measurement device are obtained by iteration. Coordinate matrix 'and use each of the corresponding matrix and the position where the object to be measured is projected to the visual plane to obtain a transfer matrix. When measuring an object to be measured, page 5 of this paper is fully compliant with the Chinese Painting National Standard (CNS) ) A4 Grid (210x297 Gongchu) I — Install -------- Order --------- Line I (Please read the notes on the back before filling this page) Intellectual Property Bureau of the Ministry of Economic Affairs Consumption cooperation with employees Du printed A? A? Intellectual property of the Ministry of Economic Affairs " Zhutu F " Zhatuhu Y 20 20 5. Description of the invention (3 Get the test object obtained by these sensing devices in this reference: 榡Shift matrix calculation. In order to achieve the above-mentioned purpose, the method for re-testing the goods month includes: a plurality of pills with a relationship between each other's jade known as the light ^ to the image sensing device to measure a predetermined image synchronously. Asset bucket, and decided based on the predetermined image data-reference Coordinates; defining a plurality of reference points of the predetermined image data; using the linear method to calculate differences from each reference point value-corresponding matrix parameters; given an initial value and performing iteration to obtain an optimization corresponding to each of the sensing devices Transfer matrix parameters; Finally, when measuring-the object to be measured, the position data measured by each sensing device is substituted into the transfer matrix, and the unique solution of the object to be measured relative to the reference coordinate is calculated in real time. The technical means adopted by the present invention to achieve the above-mentioned objects and features, and their effects, are described below with reference to the preferred embodiments and the drawings: The first diagram is a 15-dimensional image of an object in a three-dimensional space when projected on a visual plane Position diagram; The second diagram is a three-dimensional schematic diagram of the imaging position when an object in three-dimensional space is projected on the visual plane of two different optical image sensing devices; the third diagram is a flowchart of a preferred embodiment of the measurement method of the present invention Figure 4. The fourth picture is a three-dimensional schematic diagram of a standard object under test with predetermined image data. The fifth pictures A and Five B are shown when the boundary passes through a complex cell. And the image data obtained from the reduction of the electrical signals output by these unit cells are shown in a special circle. The sixth diagram is a schematic diagram of the sampling block distribution of the standard picture to be tested. And page 6 ---------- ---- Install --- (Please read the precautions on the back before filling this page)

-lSJI-Line · 芰 闼 襟 襟 襟 210 Specification (210g f) 452844 A7 Printed by the Consumer Property Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs 5. Description of the invention (4; The seventh picture shows that the visual system is not orthogonal Schematic diagram of drawing number component comparison table: 31, 32 ... CCD camera 9 ... paper jam 91, 92 ... square reference surface 911, 912, ... 914, 921, 924 ... square first please Referring to the second and third figures, the method for measuring the coordinates of an object in space according to the present invention includes the following steps: First, in step 20, a predetermined image data is simultaneously measured by a plurality of optical image sensing devices with known spatial relationships with each other. And determining a reference coordinate according to the predetermined image data; for the sake of explanation, in this embodiment, 'the complex optical image sensing device is based on two sets of CCD cameras 3 1 and 32 as an example, and the distance between the two cameras is β, Each camera has a visual plane Ri, R2, so one point P (x, y, z) to be measured in space will be projected to each of the visual plane Ri, & points, m2 (u2, v2). Pre-t / Image Information > The towel shown in the fourth picture is a paperboard 9 in the shape of Li Zhao, It mainly includes two square reference planes 91 and 92 that are perpendicular to each other, and each side has four completely black squares 911, 912, 914, 921, and 924 that are evenly distributed. Since the two reference planes 91 and 92 are perpendicular to each other, two reference planes are used. The parent line shared by w and 92 is the y-axis direction of the reference coordinate, and the lower edge of the reference plane μ is the X-axis direction. The direction of the lower edge of the reference plane 92 is the z-axis direction, and the intersection of the three coordinate axes is the origin of the reference coordinate system. Point 0. To estimate the corresponding matrix of the three-dimensional space object to be converted into the two-dimensional image from the images of the cardboard 9 taken by the two sets of CCD cameras 31 and 32, an appropriate reference point must be obtained as the initial input data. In this embodiment, , It is such as; 5 10 15 20 _______ Dental paper size 剌 + Mei Yi Pi (CNS) A4 Secret 2. Page (Please read the precautions on the back before filling in this page) Installation ----- -Order --------- The Intellectual Property Bureau of the Ministry of Economic Affairs and the Department of Work and Consumer Cooperation Du printed A7 _______ B7_____ V. A pixel operation described in the description of the invention (5) to identify each square 911, 912. ..924 each boundary line segment 'and defined by the intersection of the calculated boundary equations The plurality of reference points of the predetermined image data. As shown in the sixth figure, relative to the aforementioned reference coordinates, step 21 is to define a plurality of sampling areas by the coordinate values of the plurality of square line segments 911, 912 ... 924 on the jammed paper 59, ' Block, where each of the sampling blocks respectively corresponds to a boundary line segment of the square patterns 911, 912 ... 924 with a black solid pattern inside, and is far from the intersection of each boundary line segment, so as to avoid that the intersection point is not easy to be precise Measurement problems. 10 Secondly, if only the optical sensor element cell such as a CCD camera is allowed to convert electrical signals corresponding to the optical image on it, the cell with the square border crossed by it will be subjected to the above as the fifth The difference in the distribution of the black and white parts of Figure A 'produces different potential output signals between black and white' If these potential signals are used to reconstruct the image of this boundary region, 15 will produce different gray as shown in the fifth B 囷High-level image, forcing a threshold wth (thresh0ld) potential as the boundary to define the boundary line segment will inevitably cause the boundary line segment to appear uneven step shape distortion; therefore, in this embodiment, the In step 22, the data corresponding to each of the sampling blocks is extracted from the electrical signals converted by the entire picture, and each of the sampling blocks is a circle including 69 20 pixels, because the square patterns 911, 912. ..924 is not only a closed polygon, with a solid black pattern inside and a significant contrast with the high-contrast white portion of the exterior. Therefore, the unit cells corresponding to the 69 elements in each sampling block range can be replaced by The superposition operation of the overall unit cell output electrical signal of the row (or column) is to obtain the position of the intersection of the boundary line segment and the row (column), and this position can be referred to in §. 210 * 297 mm) ------------- Install i ·! · — Order ------- I line (Please read the precautions on the back before filling this page) 452644 A7 B7 May 10 15 Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 20 Invention description (6) Precise calculation to the sub-prime range, and perform such sub-pixel operations on adjacent rows (columns) in order to find the The line connecting the intersection of the row (column) with the boundary line segment should correspond to the equation of the reference coordinate, so that the ray equation containing each boundary line segment is less = ~ jc + 不, where t '= 1, 2, ..., 64 Then use the concept of linear algebra 'calculate the intersection of the equations corresponding to every two adjacent boundary line segments at step 23' as the four-corner 隅 coordinates of each square, and write and use the 64-corner 隅 coordinates of the eight squares as reference points. Of course, the two-dimensional space coordinates corresponding to these corner points will be detected as a reference point in two-dimensional space. As those skilled in the art can easily understand, other conventional numerical critical division methods can also be used as an alternative to obtain these reference points, and the shape, angle, and pattern for identification of the predetermined image data are not limited to this embodiment. The above description. Reviewing the above description from the perspective of linear algebra, each CCD camera can be regarded as a model with internal and external parameters. This system model has four internal parameters (pure, and the intersection of the optical axis and the image plane # 0 and ^ 0); At the same time, this system model also has six external parameters, including three parameters for camera rotation motion and three parameters for camera translation motion, thereby defining the conversion between the camera coordinate system and a reference coordinate system . This 3 乂 4 perspective projection matrix u can be written in step 23 as a function with internal and external parameters: V 'ai -1' -i into Λ. ―1. Λ. Page ------------ --------- I ----- If — Order --------- line ί ί Please read the notes on the back before filling out this page} AT 五 、 Description of Invention (

U ajy + ^ t = ^? 24 βΐ < iu (1) ίο ', ten 1 2 3 疋 column vector of rotation matrix inverse. Note: · "" 和 没> is not equal to 0, then this transparent Dongchaxin-open and radiant matrix? The rank number is 3, and then" Hexin "becomes a 1 X 3 vector. In three-dimensional space N reference points A Two-dimensional space midpoints%,%, ..., unknown < and? In the array 4? = 1, 2, 3) 得 n3 " a. + M = 0qip ,-^ lp ^ q2,-^ 34 = 0 ......... (2) where, I, ... 'N, and

Pl ’+ ·’ ···, ~ The linear equation corresponding to projection I " can be obtained from the corresponding husband: > Giant Please read the notes on the back first and then fill out this page.

Pi

A (3) The Ministry of Relief's Intellectual Condition " Booming Staff ^ 肖 费 < ^ Cha: ±.?-Β For N points, the system's 2N linear equations will be obtained, that is, Ag = 0 .. ........................................ .. (4) Here, 'matrix A is a 21s X 12 matrix containing three-dimensional space and two-dimensional spatial reference points, q is one; [2 X 1 matrix, and the component q of a 3 X 4 matrix can be expressed For: ql%? Ί L «? 34 Application of national standard-line · Ι2ί1 Sale 隼 丨 CNS) Ai specification U ratio (5) 45 264 4 10 15 A7 b / V. Description of the invention (8) where ' q is usually defined as a pure number and the rank of matrix A is 11. In order to exclude the meaningless solution of q = 0, the coefficients of the corresponding matrix must be limited, and the analysis of Eigei Analysis' the components of the corresponding matrix with respect to the smallest eigenvalue will be obtained, and the result will be expressed as U Linear solution; and u's parameter 5 can be obtained from q that satisfies II II 2 as the minimum and is limited by II q II = 1. Step 24 uses the non-linear least square method and gives an initial value. For the limitation of || q | 丨 = 1, after multiple iterations, the following non-linear equations will be optimized and reduced || | 丨The value of 2: 得到 The solution of q is obtained by Eigen Analysis. It can be used to give the initial value of the non-linear method, and then iterate several times to obtain the optimal solution of q. Then use it to rewrite a corresponding matrix U in the form of 3X4. . The internal and external parameters contained in this matrix can then be calculated. Of course, in fact 'in addition to the ten parameters listed in formula (1), the visual coordinate system shown in Figure 7 may need to be considered 14] primes' and are included in the two axial unit vectors of the visual plane; 7 and The missing angle θ between 7 causes the corresponding matrix u to be written as follows:---------- l · > 1111111 VI — — — — — — — * Please read the precautions on the back before filling (This page) Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs

U sin ^ tan ^ 'r3 and can be rewritten as a .. sin ^ page ii «page ^ 9,4? 24 (7) This paper size applies to China National Standard (CNS) A4 (210 χ 297 mm) 9 A7 B7 V. Description of the invention (a, a, u ύηθ sinff iyz? 34? 34 q: ^ 24 1 <? 34 Then find the equation of the aforementioned six internal parameters as follows. (8)? 34 a 'sin 夕 w3? 34iL 』? 34 八 αμ att--ί„ + uJ,? I4 _ ^ tan ^ '02? 34 a. ------------- 装 -------- Order- --------- line (please read the notes on the back before filling this page) Printed by the Consumer Cooperation Agreement of the Intellectual Property Bureau of the Ministry of Economic Affairs 10? 24 _ sin <? _? 34 h The reference coordinates of the 64 corners of the left and right images of Paper 9 can be used to obtain the two corresponding matrices 1 ^ and UR representing the left and right CCD cameras from three-dimensional space to two-dimensional space. When a reference point obtained by the two cameras is in two dimensions, When the coordinate values in the space are, ~) and (a, 4), in step 25, the points of these two-dimensional coordinates are projected back to the same point in the three-dimensional space using the triangular positioning method p = 1 > y 'z' t], p. u. · Paper ruler. ¾ National Standard π—Guide m〇 ...:) 97 mm) 4 5 2 6 4 4 A7 R7 V. invention is described (1Q) can be expressed as:

Mu vu

; And its center? £, 4 qL1A, ^ i3 ii34 can be derived from the above equation

Mlc

TJt P UTP u -Uhtu ~ UTp uT p y R2r jjt p ^ R3r

QP ^ O can be obtained by combining the foregoing equations

Ulx ~ ^ icUTLi UR Mrc VRC-(u R2 uT A 仞 4 ″ U R3 ^ Among them, 0 uL- ^ u: 3 ΛΙ ~ ^ rMk3

UT υΙι ~ νκΆ (Please read the precautions on the back before filling out this page) Install ----—! — Order —---- i -line 10 Printed by the Industrial Property Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs 15

After iterating given an initial value, · ^ find a set of optimal solutions ~ ― [~ '> V', then the wide-scale points in the three-dimensional space (· χ 'JV' χ = ^ ~] y ^ ^ EL and therefore, it is not necessary to force an inverse matrix, and the two sets of corresponding matrices and the related parameters obtained can be used directly to obtain the transfer matrix; and then in step 2 < Measurement I obtain the unique solution of the DUT's position relative to the reference coordinate through the real-time calculation of the transfer matrix. Based on the original _ page II-page, the paper & degree applies the Chinese National Standard (CNS) A4 specification ⑵〇χ (297 Gongchu) A7 A7 5 Read the > 1 item on the back and fill in page 10 V. Description of the invention (11 :: Can achieve the effect of real-time measurement, when the object to be measured moves, the object to be measured is calculated by the previous method = You can get the micrometer to be measured immediately by setting the coordinates of something else, and you can obtain the speed by differentiating time, or you can use time 77 "on speed and other related physical quantity information. In summary, in the space of the present invention Object coordinate measurement method ", can indeed borrow the structure , = M expected purpose and effect ', which was not seen in the publication or used before the application, and meets the requirements for novelty, progress, etc. of the invention patent. However, the figures and descriptions above are only for understanding the embodiments of the present invention. It is not intended to limit the implementation of the invention; anyone who is familiar with the art,

Other equivalent changes or modifications made in the characteristic category of the present invention shall be covered by the following patent applications in this case. U-line-Printed by the consumer cooperation of the Intellectual Property Bureau of the Ministry of Economic Affairs 14 |

Claims (1)

4 Xihuicai 4 Bureau of the Ministry of Economic Affairs 4 Consumer Cooperative Cooperatives f 5 2 6 4 4 as B8 C8 ______D8 VI. Application for Patent Scope 12 1. A method for measuring the coordinates of objects in space, including the following steps: a) by a plurality of each other's space The known optical image sensing device simultaneously measures a predetermined image data and converts it into an electrical signal, and determines a reference coordinate corresponding to the predetermined image data according to the electrical signal; 5 b) defines a plurality of references of the predetermined image data The coordinate value of the point relative to the reference coordinate; c) the corresponding matrix parameter is calculated by a linear method from each reference point value, thereby obtaining a complex correspondence matrix corresponding to the number of the optical image sensing devices; and 10 d) given a The initial value is iterated to obtain optimized transfer matrix parameters corresponding to each of the sensing devices, thereby obtaining a transfer matrix including one of the corresponding matrices. 2_ The measurement method as described in item 1 of the patent application scope, further comprising a measurement step after obtaining the optimized transfer matrix. When measuring an object to be measured, '15 each of the sensing devices are measured. The obtained position data is substituted into the transfer matrix, and a unique solution of the measured object relative to the reference coordinate position is obtained in real time. 3_ The measurement method described in item 丨 of the scope of patent application, wherein the step of defining the reference point of the predetermined image data is a numerical critical division process, including the following steps: 20 M) According to the pattern of the predetermined image data, Defining a plurality of sampling blocks, wherein the pattern of the predetermined image data includes at least one simple geometric figure, and each of the sampling blocks respectively corresponds to the boundaries of the at least one simple geometric figure and is far from the intersections of the boundaries; b2) Among the aforementioned electrical signals obtained from the measurement, the corresponding Chinese I 撷 {CNS) A4 specifications (210 X — — (please read the precautions on the back before filling this page) for each of the sampling I paper music standards are ordered. AS B8 C8 -_________ D8 VI. Patent application range 1 ~ " 13 Block data, calculate the equations in each sampling block where each boundary corresponds to the reference coordinates; and b3) Calculate the equation corresponding to each adjacent boundary The intersection point of the simple geometric pattern is used as the reference point of each adjacent boundary of the simple geometric pattern. 5 4. The measurement method described in item 3 of the patent scope, each of the optical image sensing devices in # has a complex unit cell, and the simple geometry is a closed polygon, and the inside of the polygon has The high-contrast color, ie, the optical image sensing devices t, corresponds to the far-distance unit cells of each sampling block range to obtain the equations of the corresponding boundary line segments by sub-pixel operations. ---------- ^ ------ ΐτ ------ 0 < Please read Minus Note on the back before filling this page) Printed by Employee Consumer Cooperative