KR100809541B1 - Calibration method for large scale panel measurement system - Google Patents

Abstract

A calibration method for a large-size panel measurement system is provided to calculate a precise position and posture relation between measurement units to improve measurement accuracy. A calibration method for a large-size panel measurement system includes a target unit(100), a measurement unit(200), and a main server unit(300). The target unit includes targets disposed at both ends and a scale bar disposed between the targets. The measurement unit detects the coordinates of the targets to calculate the distance between the targets and obtain angle information. The main server unit obtains intersection points of vectors to the targets using the angle information from the measurement unit to calculate the length of the target unit, and applies the length to a calibration algorithm and extracts calibration results.

Description

Calibration Method for Large Scale Panel Measurement System

1 is a block diagram schematically illustrating a method for measuring a large area panel measurement system and a target unit according to an exemplary embodiment of the present invention.

2 is a view showing a target unit for calibration of a large area panel measurement system.

3 is a module configuration diagram of a measurement unit according to an exemplary embodiment of the present invention.

4 is a module configuration diagram of an integrated management unit according to an exemplary embodiment of the present invention.

5 is a flowchart illustrating a calibration method using a system according to an exemplary embodiment of the present invention.

<Explanation of symbols for the main parts of the drawings>

100: target unit 101: target

102: scale bar 200: measurement unit

210: vision device 220: rotating device

300: main server unit

The present invention relates to a calibration method of a large area panel measurement system, and more particularly, a scale bar by recognizing a target with a measurement unit using a camera, a zoom lens, and a rotary stage, and calculating coordinates of the recognized target. It is a calibration method of a large-area panel measurement system that calculates the exact position and attitude relationship between each measurement unit by calculating the length information of.

Existing large-area panel measurement work relies on manual measurement by a tape measure by an operator.

Recently, when a worker determines that a more accurate measurement of a large area panel is necessary, the panel is accurately measured using a total station, which is a kind of surveying equipment.

Here, the total station is used to precisely measure the measurement angle or the measurement distance with respect to the object to be measured, and the electromagnetic range finder and theodolite function are combined to determine three-dimensional position coordinates between arbitrary points.

However, in order to use such a total station, the equipment must be operated by a professionally trained worker, so there is a considerable difficulty in measuring a large area panel using such equipment in the field except in special cases. .

In the current panel-based process of manufacturing large-area panels, it is necessary to measure the horizontal, vertical and diagonal lengths of large-area panels to check the alignment of each panel, and much effort of the workers is required. If the alignment of each panel of the large-area panel is wrong in the sheet-based process of manufacturing the large-area panel, the error of the panel to be aligned occurs a lot of cost and time to correct this in the post-processing process Can have

In fact, in the plate-based process of manufacturing the large-area panel, the process of checking the panel alignment as described above is repeated several times by the operator, thereby reducing the productivity efficiency of the large-area panel by manually aligning the panel for the large-area panel. I have a problem.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and an object of the present invention is to measure a target position information with a measurement system composed of a plurality of measurement units to manufacture a large area panel, and to measure the scale bar according to the coordinates of the recognized target. It is to provide the calibration method of large area panel measurement system that can improve the measurement accuracy of large area panel measurement system by calculating the exact position and posture relationship between each measurement unit by calculating the length information of scale bar).

In order to achieve the above objects, according to an aspect of the present invention, in the calibration method of a large-area panel measurement system, (a) using a plurality of measurement units to search for a target formed with a specific shape or identifier to obtain angle information Acquiring (S401); (b) calculating a position of a target using the angle information (S403); (c) calculating a length of a scale bar using the calculated position of the target (S405); (d) applying the angle information of the target and the length information of the scale bar to a calibration algorithm (S407); And (e) calculating coordinate values and attitude information from the reference measurement unit to the dependent measurement unit (S409); It can provide a calibration method of a large-area panel measurement system comprising a.

Next, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram schematically illustrating a method for measuring a large area panel measurement system and a target unit according to an exemplary embodiment of the present invention, and FIG. 2 is a view illustrating a target unit for calibration of a large area panel measurement system. .

1 and 2, a large area panel measuring system according to the present invention is composed of a target 101 for providing a known information for calibration and a scale bar 102 between the targets. A target unit 100, a measurement unit 200 that detects coordinates of targets at both ends of the target unit, calculates a length, and obtains angle information; And a main server unit 300 that calculates the intersection of the vectors to the target based on the angle information of the measurement unit, calculates the coordinates at which the target is located, calculates the length of the target unit, and applies the calibration algorithm to extract a calibration result. .

The target unit 100 performs a function of being located and measured at a plurality of positions in order to obtain the required number of length information for calibration.

The target unit 100 has a specific shape or identifier, and a scale bar 102 formed between a pair of targets 101 and a pair of targets to measure a plurality of panels. It is installed at each location.

The measurement unit 200 recognizes and rotates the target unit 100 at an arbitrary position and stores an angle thereof.

The measurement unit 200 obtains an angle with respect to one target and obtains coordinates by obtaining an intersection point of target vectors generated in each measurement unit.

The measurement unit 200 includes a vision device 210 including a zoom lens 205 for recognizing the target unit 100 having a specific shape or identifier, and the vision device 210 includes the target unit 100. It comprises a plurality of rotating device 220 for rotating to be aimed at).

The plurality of measurement units 200 may rotate based on a axis of a predetermined reference coordinate system using the vision device and the rotation device, and the rotation axis of the subordinate measurement units matched through the reference axis is calibrated along a target. Perform the function of calculating the angle.

In this case, the reference coordinate system arbitrarily selects one of the plurality of measurement units to determine a reference, and sets the origin of the reference coordinate system to the selected reference measurement unit.

Here, the plurality of measurement units 200, 200-1, 200-2, and 200-3 may be installed at predetermined distances on both sides of a large area panel as a measurement object.

The vision device 210 performs a function of recognizing the target unit 100 in which an identifier such as a specific shape, color, display, or the like is formed.

The vision device 210 outputs image information photographed using the CCD camera 210 and the zoom lens 205.

The image processing module 301 of the main server unit 300 that receives the image information through the vision device 210 receives target image information of the target unit 100 in which a specific form or identifier is formed, and targets each installed in a large area panel. The target of the unit 100 may be recognized.

The rotating device 220 performs the function of rotating the vision device 210 so as to aim at the target of the target unit 100 in which the specific shape or identifier is formed.

At this time, the method of aiming at the target of the target unit by using the rotating device 220 is to recognize the target using a vision device, calculate the degree of deviation from the center of the image and transmit it to the rotating device after the rotating device is Rotate so that the target is centered in the image.

Then, a plurality of measurement units aim at one target to obtain coordinates of the target, and then aim at a target on the opposite side to obtain coordinates, and then calculate distances using the two coordinates.

The rotating device 220 may be fixedly mounted to the vision device 210, and may be rotated and finely adjusted such that the fixed mounted vision device 210 is aimed at the target 101 having the specific shape or identifier.

The measurement unit 200 drives the vision device 210 and the rotation device 220, and transmits the rotation angle calculated by aiming the target unit 100 to the main server unit 300 to provide the large area target unit. Calculate the length of the scale bar.

The main server unit 300 calculates coordinates of the target in which the target unit 100 is located, and calculates length information of the scale bar of the target unit for calibration.

In this case, the length information is obtained by obtaining coordinates of both targets, calculating the length using the coordinates, and performing the calibration using the length to calculate the length information.

The main server unit 300 calculates the length of the scale bar of the target unit by using the calculated target coordinates and performs calibration using the length information to output coordinate information and attitude information between the measurement units. Can be.

In addition, the main server unit 300 may calculate target coordinates through coordinates of the intersection where a plurality of measurement units 200 intersect at the center of one target 100, and the other target coordinates installed at each It can be calculated by repeating the same method.

Through such a configuration, the calibration of the large-area panel measurement system according to the present invention is performed by using the measurement unit 200 and the main server unit 300 on which the vision device 210 and the rotation device 220 are installed. By applying this, it is possible to accurately determine each panel in the plywood process of the panel-based process for producing a large area panel 50 and to check the panel state so that each panel can be accurately aligned.

3 is a module configuration diagram of a measurement unit according to an exemplary embodiment of the present invention.

Referring to FIG. 3, the measurement unit 200 according to the present invention includes a transmission / reception processing module 250 for processing wired / wireless communication with a vision device and a rotating device, and a vision for recognizing a target having a specific form or identifier. A device module 251, and a rotating device module 253 which is fixedly mounted to the vision device and rotates such that the fixed mounted vision device is aimed at the target.

Next, the internal configuration of the measurement unit according to an embodiment of the present invention is as follows.

The transmission and reception processing module 250 is provided with information for performing wired and wireless communication with a vision device or a rotating device.

The vision device module 251 performs a function of managing a vision device that recognizes a target having a specific shape or identifier by transmitting image information photographed using a CCD camera and a zoom lens to a measurement unit.

The rotating device module 253 performs a function of managing the rotating device so that the vision device fixedly mounted to the rotating device may aim at the target having the specific shape or identifier.

4 is a module configuration diagram of an integrated management unit according to an exemplary embodiment of the present invention.

Referring to FIG. 4, the main server unit 300 calculates a target recognition and a target center from the image information and the angle information obtained from the measurement unit and transmitted to the integrated management unit 300 through the transmission / reception module. The image processing module 301 and the coordinate and length calculation module 302 for calculating the coordinate information of the target and the length of the scale bar using the angle information of the target of the target unit 100, and the calibration using the length information. A calibration module 303 to perform, and an integrated control module 304 for integrated control management of the measurement unit 200.

Next, the internal configuration of the main server unit according to an embodiment of the present invention is as follows.

The image processing module 301 recognizes the presence or absence of a target using the image information transmitted from the measurement unit, and performs an image processing function for finding an accurate center of the target.

The coordinate and length calculation module 302 calculates the coordinates of each target using the angle information of the targets transmitted from the measurement unit, and calculates the length of the scale bar of the target unit using the target coordinates.

The calibration module 303 performs calibration using the calculated length information and outputs coordinate information and posture information between measurement units.

The integrated control module 304 performs a function of controlling and managing the measurement units 200.

5 is a flowchart illustrating a calibration method using a system according to an exemplary embodiment of the present invention.

Referring to FIG. 5, the calibration method of the large-area panel measuring system according to the present invention first searches for a target having a specific shape or an identifier using a plurality of measuring units.

The step S401 is a step for measuring the coordinates of the target of the target unit in order to obtain the actual length information, the vision device to rotate and recognize the target in each measurement unit with a scale bar (Scale Bar) at an arbitrary position The angle information is obtained by rotating and observing.

The above step is a step for measuring the coordinates of the target of the target unit to obtain the actual length information, by rotating the camera using the approximate position where the target unit is located, select one of the two targets of the target unit and Is automatically recognized by the vision system, and the difference between the center of the image and the center of the target is calculated and passed to the rotation module. The rotation module rotates based on the information so that the target is positioned at the center of the image. Reading its rotation angle can take angle information.

Next, the main server unit calculates the position of the target using the angle information (S403).

In the step S403, the position of the target is calculated using the angle information of the target located at an arbitrary position through the main server unit.

At this time, the main server unit targets the target unit at an arbitrary number of positions as necessary for calibration, and repeatedly acquires the angle information and calculates the target position to target the target at a position equal to or greater than an unknown number for calculating the calibration. Place the unit.

Next, the main server unit calculates the length of the scale bar by using the calculated target position information (S405).

In step S405, the length of the scale bar is calculated by detecting the positions of two targets positioned at the arbitrary positions.

Next, the main server unit performs the step of applying the angle information of the target and the length information of the scale bar to the calibration algorithm (S407).

At this time, the calibration algorithm of step S407 uses the Levenberg-Marquardt Method for Nonlinear Equation using a well-known nonlinear equation.

The Levenberg-Marquardt method, which is a known algorithm known to those skilled in the art, is one of optimization algorithms, which models an objective function for optimization and obtains an unknown, that is, a solution by iterative computation that minimizes the objective function. One of the algorithms. In other words, the Levenberg-Marquardt method is an optimization algorithm that models an objective function for optimization and obtains an unknown value, or solution, by iterative computation that minimizes the objective function. It is a general method to solve the nonlinear least square problem to find the coefficient value that forms the function when the corresponding result value is known.

Algorithms for optimization include Newton Method, Newton-Gauss Method, Levenberg-Marquardt Method, and Newton Gauss Newton Gauss) and the Levenberg-Marquardt Method are mainly used. In an embodiment of the present invention, the Levenberg-Marquardt Method is used. Since the general method for solving the Levenberg-Marquardt Method is a well-known technique, the description thereof is omitted in the detailed description herein.

In other words, the Levenberg-Marquardt Method is a nonlinear regression called Levenberg-Marquardt Compromise to quickly and accurately find the optimal value of an important variable starting from an initial estimate of the variable's value. Use the modified version of.

The Levenberg-Marquardt Method used in the present invention was calculated by using the following equation by updating the algorithm with a repeated algorithm as a whole by dividing the positional relationship and the posture relationship between the measurement units.

[Equation]

[Reference]

If described with reference to the equation and the coordinate value and attitude information of the measurement unit 2,

The basic Levenberg-Marquardt method is the same as the above equation and the algorithm is updated by dividing the algorithm into two parts, Sub-iteration 1 and 2, as shown in the above reference.

That is, the basic Levenberg-Marquardt algorithm corresponds to one iterative sub-iteration in the reference diagram. However, the algorithm applied in the present invention composes one Iteration as a whole and applies Sub-Iteration which performs only position offset optimization separately, and applies Sub-Iteration which performs optimization of angle offset separately based on the result. To optimize.

The method of applying the algorithm without modifying or adjusting the Levenberg-Marquardt Method itself is not the basic method, but the Levenberg-Marquardt Method once applied to the position offset and the Levenberg-Marquardt Method once applied to the angle offset. It is a process.

The convergence of positional information and posture information convergence is divided separately and the repeated parts are separated to converge more accurately and quickly.

In other words, if the numerical unit of location information and posture information is greatly different and converged together, the value of the small value is affected by the long time of the repetition time due to the small part, or by the change of the large value. The convergence value fluctuates too much and tends not to converge. In order to eliminate this effect, the Levenberg-Marquardt method is used .

The calibration algorithm configures a sub iteration process in the entire iteration process, so that the first sub iteration 1 is repeated updating only α ₂ , and X ₂ and Y ₂ are fixed. In this state, the calibration algorithm is executed, and the second sub-iteration ₂ is fixed, α ₂ is fixed, and only the X ₂ and Y ₂ are updated, and the calibration algorithm is executed.

Next, a step of calculating coordinate values and attitude information from the reference measurement unit to the subordinate unit is performed (S409).

Step S409 measures the relationship between the reference measurement unit 0 (Origin; measurement unit 200) and measurement unit 1 (X ₁ , Y ₁ ) α ₁ (measurement unit 200-1) using the angle information and the length information. The relationship between the measurement unit 0 and the measurement unit 2 (X ₂ , Y ₂ ) α ₂ (measurement unit 200-2) is measured using the angle information and the length information, and the measurement unit 0 and the measurement unit 3 ( The relationship with X ₃ , Y ₃ ) α ₃ (measurement unit 200-3) is measured using the angle information and the length information.

The measuring unit may include a vision device for acquiring image information for recognizing a target having a specific shape or an identifier, and a rotation device for rotating the vision device to be aimed at the target.

The main server unit may recognize the target having the specific form or identifier by receiving image information from the measurement unit and confirming the presence or absence of the target having the specific form or identifier.

On the other hand, the embodiments of the present invention disclosed in the specification and drawings are merely presented specific examples to aid understanding, and are not intended to limit the scope of the present invention. It is apparent to those skilled in the art that other modifications based on the technical idea of the present invention can be carried out in addition to the embodiments disclosed herein.

According to the present invention, a calibration method of a large area panel measurement system for precisely manufacturing a large area panel recognizes target position information with a measurement unit and a main server unit, and the length of a scale bar according to the recognized target coordinates. By calculating the information, it is possible to improve the measurement accuracy by calculating the exact position and attitude relationship between each measurement unit.

Claims (4)

In the calibration method of a large area panel measurement system, (a) searching for a target having a specific shape or identifier formed by using the plurality of measurement units to obtain angle information (S401); (b) calculating a position of the target using the angle information of the target (S403); (c) calculating a length of a scale bar using the calculated position of the target (S405); (d) applying to a calibration algorithm through angle information of the target and length information of the scale bar (S407); And (e) calculating coordinate values and attitude information from the reference measurement unit to the dependent measurement unit (S409); Calibration method using a large area panel measurement system comprising a. The method of claim 1, In step (e), The relationship between the measurement unit 0 (Origin) and the measurement unit 1 (X ₁ , Y ₁ ) α _{1 is} measured using the angle information and the length information described above, and the measurement unit 0 and the measurement unit 2 (X ₂ , Y _2). ) α to the _second relationship between the measurement by using the angle information and the length information, and measured by using the measuring unit 0, and the measurement unit _{3 (X 3, Y 3)} α 3 and the angle information and the length information, wherein the relation A calibration method of a large area panel measurement system, characterized in that. The method of claim 1, The calibration algorithm of step (d), Each sub iteration is configured in the entire iteration process, and the first sub iteration is repeated updating only α ₂ , and the calibration algorithm is executed while X ₂ and Y ₂ are fixed. Calibration method using a large area panel measurement system. The method of claim 1, The calibration algorithm of step (d), Each sub iteration is configured in the entire iteration process, and the second sub iteration is characterized in that α ₂ is fixed, only X ₂ and Y ₂ are updated, and the calibration algorithm is executed in the repeated state. Calibration method using a large area panel measurement system.

Images