Method for correcting distortion of optical engine of DLP printer

By calculating the distortion coefficient of the optical machine and performing correction processing, the printing accuracy problem caused by optical machine distortion in DLP 3D printing technology is solved, and high-precision and fast printing effect are achieved.

WO2024239477A1PCT designated stage expired Publication Date: 2024-11-28NANJING CHAMLION LASER TECH CO LTD
View PDF 8 Cites 0 Cited by

Patent Information

Application Number
PCT/CN2023/116197
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2023-05-24
Filing Date
2023-08-31
Publication Date
2024-11-28

AI Technical Summary

Technical Problem

In DLP 3D printing technology, due to the nonlinear distortion of the optical machine, the projected image and the original image have pillow-shaped or barrel distortion, resulting in the problem of mismatch between the printing model and the original shape. The existing technology is difficult to effectively correct the distortion of the optical machine. Impacts printing accuracy.

Method used

Through the model generation server, surveying and mapping server and correction server, use high-precision imager to identify the solid slice pattern, calculate the distortion coefficient of the optical machine, use the correction formula group to perform the distortion correction of the optical machine, and calibrate it through the Internet. Finally, Correction of the printing model is used using corrective distortion coefficients.

Benefits of technology

It effectively avoids the origin coordinate error caused by resin shrinkage, improves the accuracy of the printing model, reduces the time-consuming of the correction machine, and ensures high accuracy and rapid completion of the printing results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN2023116197_28112024_PF_FP_ABST
    Figure CN2023116197_28112024_PF_FP_ABST
Patent Text Reader

Abstract

The present invention belongs to the field of 3D printing. Disclosed is a method for correcting distortion of an optical engine of a DLP printer. The method comprises: acquiring measured coordinates of standard coordinates corresponding to a printed actual model, calibrating a series of parameters such as an intrinsic parameter and an extrinsic parameter of an optical engine device, and distortion coefficients of an optical engine, and finally forming a distortion model by means of the parameters acquired in the second step, and substituting into a formula a standard corresponding to each pixel in an original image, so that a processed original image can be acquired. The technical problem of distortion of an optical engine of a 3D printer is solved. In the present invention, by regarding the optical engine as an inverse process of a camera, a result may be more precise; and if the measurement of the center of a circle is not affected and the number of dots of a printed model is increased, the result may be more precise.
Need to check novelty before this filing date? Find Prior Art

Description

A method for correcting optical and mechanical distortion of a DLP printer Technical Field

[0001] The present invention belongs to the field of 3D printing, and in particular relates to a method for correcting optical and mechanical distortion of a DLP printer. Background Art

[0002] DLP 3D printing technology projects digitally processed image signals onto a liquid photosensitive resin layer by layer, curing and shaping the resulting image. The use of optical lenses inevitably introduces distortion on the projection plane, leading to discrepancies between the projected image and the actual image. This can cause problems such as mismatches between DLP-printed models and their originals, and mismatches between printed dental membranes and their scaffolds.

[0003] In the existing technology, the homography matrix can only deal with the errors caused by the spatial position relationship between the projector and the projection plane. The nonlinear distortion of the optical machine itself cannot be eliminated by the homography matrix, and the accuracy of the optical machine projection is difficult to guarantee.

[0004] Summary of the Invention

[0005] The purpose of the present invention is to provide a method for correcting optical-mechanical distortion of a DLP printer, which solves the technical problem of optical-mechanical distortion of a 3D printer.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] A method for correcting optical-mechanical distortion of a DLP printer comprises the following steps:

[0008] Step 1: The model generation server obtains the original model file from the client server via the Internet, generates the original model based on the original model file, sets the resolution parameters and print size parameters of the slicing software, slices the original model using the slicing software, and generates a slice pattern of the original model;

[0009] The model generation server designs the printing dot matrix according to the slice pattern, and sets the relative coordinate system of the printing dot matrix, and generates the coordinates of each point in the printing dot matrix in the relative coordinate system, that is, the relative coordinates;

[0010] The model generation server sends the slice pattern, printing dot matrix, resolution parameters and printing format size parameters to the DLP printing device via the Internet;

[0011] Step 2: The DLP printing device adjusts its own printing format settings according to the printing format size parameters, and calibrates the projection clarity according to the resolution parameters. After the calibration is correct, the slice pattern is printed to obtain a physical slice pattern;

[0012] Step 3: The mapping server constructs a world coordinate system and uses a high-precision imager to identify the entity slice pattern and its corresponding relative coordinate system, aligning the origin of the relative coordinate system with the origin of the world coordinate system.

[0013] The measured coordinates of each point on the printed dot matrix in the solid slice pattern are identified by a high-precision imager, and the measured coordinates are plotted in the world coordinate system;

[0014] Step 4: The correction server retrieves the relative coordinates and measured coordinates of all points in the printed dot matrix from the model generation server and the surveying and mapping server via the Internet;

[0015] The optical distortion coefficient of the DLP printing device is calculated according to the following correction formula group:

[0016] Among them, (X c , Y c , Z c ) represents the coordinates of the point in the camera coordinate system, (X, Y, Z) represents the coordinates of the point in the world coordinate system, (u, v) represents the pixel coordinates, are the optical-mechanical internal parameters, k1, k2, k3, k4, k5, and k6 are radial distortion parameters, p1 and p2 are tangential distortion parameters, s1, s2, s3, and s4 are thin prism distortion parameters, (x′, y′) is the undistorted pixel coordinate, i.e., the relative coordinate, (x″, y″) is the measured coordinate, and f x and f y Represents the offset of the pixel in the x and y directions, C x and C y represents the pixel position of the intersection of the optical axis and the image, f represents the focal length, R represents the rotation matrix, and t represents the translation matrix; the distortion coefficient includes tangential distortion parameters, thin prism distortion parameters, and radial distortion parameters;

[0017] In the calculation, (x′, y′) takes the value of (u, v);

[0018] Step 5: Calibrate the calculated result in step 4 according to the error value between the relative coordinates and the measured coordinates. When the preset calibration value is reached, the calibration is successful and the distortion coefficient is recorded.

[0019] Step 6: Select any one of the slice patterns as a calibration image, and obtain the distortion coefficient of the calibration image as the correction distortion coefficient through the method of steps 1 to 5;

[0020] Step 7: Based on the correction distortion coefficients obtained in step 6, use the correction formula group to recalculate and print the original model to obtain the corrected model.

[0021] Preferably, all dots in the printing dot matrix are circular.

[0022] Preferably, when executing step 4 to calculate (x″, y″), the calculation result is a floating point number, and the calculation result is filled into the four adjacent pixel grids in the form of sub-pixels, that is, the image information of the input optical machine corresponding to the slice pattern is obtained.

[0023] Preferably, the model generation server, surveying and mapping server, client server, DLP printing device and correction server all communicate with each other via the Internet.

[0024] Preferably, when executing step 2, the DLP printing device is calibrated for clarity, with the condition that black stripes between dense white stripes in the image can be seen as meeting the standard.

[0025] The DLP printer optical distortion correction method described in the present invention solves the technical problem of optical distortion of 3D printers. The present invention regards the optical machine as the inverse process of the camera, and uses the result of camera calibration to distort the image. After the principle of inverse distortion projection of the optical machine, in the final processing of the original image, the floating-point result is filled with the four adjacent pixels using the sub-pixel idea, avoiding the distortion of some pixels caused by the conversion of floating points to integers. The number of printed model layers does not need to be too many, and the coordinate information of the printed model can be obtained at a short printing time, which reduces the time consumption of correcting the optical machine. The printed model can be processed as soon as possible, which can effectively avoid the origin coordinate error caused by resin shrinkage. The increase in the number of printed model points can make the result more accurate, and the increase in the number does not affect the measurement of the center of the circle. The increase in the number of printed model points can make the result more accurate, and the increase in the number does not affect the measurement of the center of the circle. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] FIG1 is a flow chart of the present invention;

[0027] FIG2 is a schematic diagram of a signal design printing accuracy model of the present invention;

[0028] FIG3 is a schematic diagram of the error before signal correction of the present invention;

[0029] FIG4 is a schematic diagram of the error after correction of the present invention. DETAILED DESCRIPTION

[0030] A method for correcting optical-mechanical distortion of a DLP printer, as shown in FIG1 to FIG4 , specifically includes the following steps:

[0031] Step 1: The model generation server obtains the original model file from the client server through the Internet, generates the original model based on the original model file, sets the resolution parameters and print format size parameters of the slicing software, slices the original model through the slicing software, and generates a slice pattern of the original model.

[0032] The model generation server designs a printing dot matrix according to the slice pattern, and sets a relative coordinate system for the printing dot matrix, generating the coordinates of each point in the printing dot matrix in the relative coordinate system, namely, relative coordinates.

[0033] All the dots in the printing dot matrix are circular.

[0034] In this example, the printing dot matrix is ​​designed according to user needs, such as a 33*14 dot matrix, where each dot is a circle. The relative coordinates of each dot in the dot matrix are designed. The dot matrix should cover the entire surface of the optical projection without multiple sampling. Since the model needs to be printed using a printer, the model's shaping needs to be considered when designing the model.

[0035] The model generation server sends the slicing pattern, printing dot matrix, resolution parameters and printing format size parameters to the DLP printing device via the Internet.

[0036] Step 2: The DLP printing device adjusts its own printing format settings according to the printing format size parameters, and calibrates the projection clarity according to the resolution parameters. After the calibration is correct, the slice pattern is printed to obtain a physical slice pattern;

[0037] When calibrating the clarity of a DLP printing device, the calibration is performed based on the condition that black stripes can be seen between dense white stripes in the image.

[0038] In this embodiment, when the DLP device is debugged, for example, when printing a 4K image, the slicing software sets the resolution to 3840*2160 and the size to 249.6mm*140.4mm. The printing format of the DLP device is also adjusted to 249.6mm*140.4mm. The projection clarity detects the image and adjusts the light machine projection clarity. If black stripes between dense white stripes can be seen in the image, the clarity meets the standard.

[0039] Use a DLP 3D printer to print the model. The printed model should be cleaned as soon as possible and its dots can be identified using an imager without curing.

[0040] Step 3: The mapping server constructs a world coordinate system and uses a high-precision imager to identify the entity slice pattern and its corresponding relative coordinate system, aligning the origin of the relative coordinate system with the origin of the world coordinate system.

[0041] The measured coordinates of each point on the printed dot matrix in the solid slice pattern are identified by a high-precision imager, and the measured coordinates are plotted in the world coordinate system;

[0042] A high-precision imager is used to identify each origin in the model, and then a coordinate system corresponding to the standard coordinates is established to output the measured physical coordinates of each origin. Because the method uses multi-point fitting, when identifying circles, each circle cannot be identified with bias.

[0043] Step 4: The correction server retrieves the relative coordinates and measured coordinates of all points in the printed dot matrix from the model generation server and the surveying and mapping server via the Internet;

[0044] The optical distortion coefficient of the DLP printing device is calculated according to the following correction formula group:

[0045] Among them, (X c , Y c , Z c ) represents the coordinates of the point in the camera coordinate system, (X, Y, Z)

[0046] Represents the coordinates of the point in the world coordinate system, (u, v) represents the pixel coordinates, are the internal parameters of the optical machine, which are the internal parameters of the optical machine and are specified by the optical machine manufacturer. k1, k2, k3, k4, k5 and k6 all represent radial distortion parameters, p1 and p2 represent tangential distortion parameters, s1, s2, s3 and s4 represent thin prism distortion parameters, (x′, y′) are the undistorted pixel coordinates, i.e. relative coordinates, (x″, y″) are the measured coordinates, and f x and f y Represents the offset of the pixel in the x and y directions, C x and C y represents the pixel position of the intersection of the optical axis and the image, f represents the focal length, R represents the rotation matrix, and t represents the translation matrix; the distortion coefficient includes tangential distortion parameters, thin prism distortion parameters, and radial distortion parameters;

[0047] When executing step 4 to calculate (x″, y″), the calculation result is a floating point number, and the calculation result is filled into the four adjacent pixel grids in the form of sub-pixels, that is, the image information of the input optical machine corresponding to the slice pattern is obtained.

[0048] In the calculation, (x′, y′) takes the value of (u, v);

[0049] In this embodiment, the calculation formula of (u, v) is derived by the following formula: x′=X c / Z c ; (Derived Formula 3) y′=Y c / Z c ; (derive formula 4) x′ 2 +y′ 2 =r 2 ; (derive formula 5)

[0050] Among them, derivation formula 1 is used to convert each pixel point in the slice pattern into the corresponding relative coordinates; derivation formula 2 is used to convert the relative coordinates into pixel coordinates that are convenient for image processing, that is, (u, v). In actual calculation, (u, v) is equal to (x′, y′).

[0051] In the derived formula 1, T is the translation matrix, which is the same as the translation matrix t in the correction formula group.

[0052] During calculation, two multivariate equations can be established for one point. Given 462 known points, a system of 924 multivariate equations can be obtained. The least squares method can be used to fit and solve the internal and external distortion coefficients of the optical machine.

[0053] Step 5: Calibrate the calculated result in step 4 according to the error value between the relative coordinates and the measured coordinates. When the preset calibration value is reached, the calibration is successful and the distortion coefficient is recorded.

[0054] Step 6: Select any one of the slice patterns as a calibration image, and obtain the distortion coefficient of the calibration image as the correction distortion coefficient through the method of steps 1 to 5;

[0055] Step 7: Based on the correction distortion coefficients obtained in step 6, use the correction formula group to recalculate and print the original model to obtain the corrected model.

[0056] The model generation server, surveying and mapping server, client server, DLP printing device and correction server all communicate with each other through the Internet.

[0057] In this embodiment, measured values ​​can be obtained by any method. The higher the precision of the obtained value, the better. The precision error of the measured value cannot be higher than the precision error between the final correction and the standard value. The number of dots in the printed model can be increased. As the number of dots increases, the correction accuracy increases. After printing the precision dot model, considering the shrinkage properties of the resin material, it is necessary to quickly clear it and then use the device to capture the measured value.

[0058] Use a printer to print the processed model image data, then use an imager to measure the actual value and compare it with the standard to observe the accuracy error. If the error is less than 0.05mm, the calibration is successful.

[0059] The present invention inputs measured coordinates and returns pixel coordinates on the corresponding image, with very small errors compared to the original image's corresponding pixel coordinates, averaging within 0.1 pixels. The present invention inputs standard coordinates into the function, and the output value is a non-standard image, that is, an inversely distorted image, which is then projected to the standard coordinates after distortion through the camera lens.

[0060] The DLP printer optical distortion correction method described in the present invention solves the technical problem of optical distortion of 3D printers. The present invention regards the optical machine as the inverse process of the camera, and uses the result of camera calibration to distort the image. After the principle of inverse distortion projection of the optical machine, in the final processing of the original image, the floating-point result is filled with the four adjacent pixels using the sub-pixel idea, avoiding the distortion of some pixels caused by the conversion of floating points to integers. The number of printed model layers does not need to be too many, and the coordinate information of the printed model can be obtained at a short printing time, which reduces the time consumption of correcting the optical machine. The printed model can be processed as soon as possible, which can effectively avoid the origin coordinate error caused by resin shrinkage. The increase in the number of printed model points can make the result more accurate, and the increase in the number does not affect the measurement of the center of the circle. The increase in the number of printed model points can make the result more accurate, and the increase in the number does not affect the measurement of the center of the circle.

Claims

1. A method for correcting optical-mechanical distortion of a DLP printer, characterized in that: The specific steps include: Step 1: The model generation server obtains the original model file from the client server through the Internet, generates the original model through the original model file, sets the resolution parameters and printing format size parameters of the slicing software, slices the original model through the slicing software, and generates the slicing pattern of the original model; The model generation server designs the printing dot matrix according to the slice pattern, and sets the relative coordinate system of the printing dot matrix, and generates the coordinates of each point in the printing dot matrix in the relative coordinate system, that is, the relative coordinates; The model generation server sends the slice pattern, printing dot matrix, resolution parameters and printing format size parameters to the DLP printing device via the Internet; Step 2: The DLP printing device adjusts its own printing format settings according to the printing format size parameters, and at the same time calibrates the projection clarity according to the resolution parameters. After the calibration is correct, the slice pattern is printed to obtain a physical slice pattern; Step 3: The mapping server constructs a world coordinate system, and uses a high-precision imager to identify the entity slice pattern and its corresponding relative coordinate system, and makes the origin of the relative coordinate system coincide with the origin of the world coordinate system; The measured coordinates of each point on the printed dot matrix in the solid slice pattern are identified by a high-precision imager, and the measured coordinates are plotted in the world coordinate system; Step 4: The correction server retrieves the relative coordinates and measured coordinates of all points in the printing dot matrix from the model generation server and the surveying and mapping server through the Internet; According to the following correction formula group, the optical machine distortion coefficient in the DLP printing device is calculated: Among them, (X c , Y c , Z c ) represents the coordinates of the point in the camera coordinate system, (X, Y, Z) represents the coordinates of the point in the world coordinate system, (u, v) represents the pixel coordinates, are the optomechanical internal parameters, k1, k2, k3, k4, k5 and k6 are radial distortion parameters, p1 and p2 are tangential distortion parameters, s1, s2, s3 and s4 are thin prism distortion parameters, (x′, y′) are the undistorted pixel coordinates, i.e., relative coordinates, (x″, y″) are the measured coordinates, and f x and f y Represents the offset of the pixel in the x and y directions, C x and C y represents the pixel position of the intersection of the optical axis and the image, f represents the focal length, R represents the rotation matrix, and t represents the translation matrix; the distortion coefficient includes tangential distortion parameters, thin prism distortion parameters, and radial distortion parameters; In calculation, (x′, y′) takes the value of (u, v); When executing step 4 to calculate (x″, y″), the calculation result is a floating point number, and the calculation result is filled into four adjacent pixel grids in the form of sub-pixels, that is, the image information of the input optical machine corresponding to the slice pattern is obtained; Step 5: Calibrate the calculated result in step 4 according to the error value between the relative coordinates and the measured coordinates. When the preset calibration value is reached, the calibration is successful and the distortion coefficient is recorded. Step 6: Select any one of the slice patterns as a calibration image, and obtain the distortion coefficient of the calibration image as the correction distortion coefficient through the method of steps 1 to 5; Step 7: According to the correction distortion coefficient obtained in step 6, the correction formula group is used to recalculate and print the original model to obtain the corrected model.

2. A method for correcting optical-mechanical distortion of a DLP printer as claimed in claim 1, characterized in that: All dots in the printing dot matrix are circular.

3. A method for correcting optical-mechanical distortion of a DLP printer as claimed in claim 1, characterized in that: The model generation server, mapping server, client server, DLP printing device and correction server all communicate with each other through the Internet.

4. A method for correcting optical-mechanical distortion of a DLP printer as claimed in claim 1, characterized in that: When executing step 2, the DLP printing device is calibrated when performing clarity calibration, with the condition that black stripes between dense white stripes in the image can be seen as a qualified condition.

Citation Information

Patent Citations

  • Method for correcting projection distortion of photo-curing 3D printer

    CN107945130A

  • DLP three-dimensional printer and calibrating method of projector of DLP three-dimensional printer

    CN108262953A

  • DLP printing precision improving method based on intelligent optical distortion correction

    CN110751609A

  • Intelligent correction control method for DLP 3D printing

    CN114851566A

  • Projection distortion and uniformity correction equipment and method for photocuring multi-picture 3D printer

    CN114986878A