Correction method of slice image for 3D printing

Abstract

The invention provides a method for correcting a slice image for 3D printing. The method comprises the following steps: acquiring an original slice image; acquiring first light field distribution of at least one first sub-pixel light spot of the 3D printer; obtaining a to-be-printed slice image corresponding to the original slice image according to the first light field distribution; training a neural network model according to the original slice image and the to-be-printed slice image to obtain a trained corrected neural network model; and inputting the current slice image to be printed into the corrected neural network model to obtain a corrected slice image. According to the method, the influence of light spot overlapping can be reduced by utilizing the learning ability of the neural network, fine features of the slice image can be generated, edge overexposure of large features can be eliminated, and the slice image of a complex graph can be corrected, so that the precision of a mask pattern is improved, and the precision of 3D printing is improved.

Description

Technical Field

[0001] This application mainly relates to 3D printing technology, and particularly to a method for correcting sliced images for 3D printing. Background Art

[0002] In patent CN105690753B, by reducing the pixel spot size of the original beam sub-image (such as reducing by 1 / m) and combining with an offset structure to offset and expose in sequence, a printed image is generated by combination. This technology uses a Gaussian beam as the image information carrier. The Gaussian beam has a spot radius, and the light intensity distribution conforms to the Gaussian distribution. However, in practical applications, the light intensity distribution range of the Gaussian spot causes spot overlap between pixels of the beam sub-image, and as the number of offsets and exposures increases, the overlapping area increases. This makes it difficult to generate fine features in the image, and the edges of larger features are prone to overexposure, resulting in distortion of the printed image and reducing the printing accuracy.

[0003] Therefore, there is an urgent need for a method that can reduce the influence of spot overlap, can generate fine features of the sliced image, can eliminate the edge overexposure of larger features, and correct the sliced image of complex graphics, thereby improving the accuracy of the mask pattern and the accuracy of 3D printing. Summary of the Invention

[0004] To solve the above technical problems, this application provides a method for correcting sliced images, which can generate fine features of the sliced image, can eliminate the edge overexposure of larger features, and improve the accuracy of 3D printing.

[0005] To solve the above technical problems, this application provides a method for correcting sliced images for 3D printing, including: obtaining an original sliced image; obtaining the first light field distribution of at least one first sub-pixel spot of a 3D printer; obtaining a sliced image to be printed corresponding to the original sliced image according to the first light field distribution; training a neural network model based on the original sliced image and the sliced image to be printed to obtain a trained corrected neural network model; and inputting the current sliced image to be printed into the corrected neural network model to obtain a corrected sliced image.

[0006] In some embodiments, the step of obtaining the first light field distribution of at least one first sub-pixel spot of a 3D printer includes: measuring the characteristic data of the first sub-pixel spot, where the characteristic data includes a number of light intensity values; fitting the characteristic data to obtain a mathematical model of the first sub-pixel spot; and obtaining the first light field distribution of at least one first sub-pixel spot according to the ratio of the mathematical model to the distance between adjacent first sub-pixel spots.

[0007] In some embodiments, the mathematical model includes a Gaussian function and / or a polynomial.

[0008] In some embodiments, the step of obtaining the first light field distribution according to the mathematical model and the ratio of the distances between two adjacent first sub-pixel light spots includes: calculating the full width at half maximum of the Gaussian function; and obtaining the first light field distribution according to the ratio of the full width at half maximum and the distance and the Gaussian function.

[0009] In some embodiments, the step of obtaining the to-be-printed slice image corresponding to the original slice image according to the first light field distribution includes: obtaining the second light field distribution of the first pixel light spot according to the first light field distribution, where the first pixel light spot includes a plurality of first sub-pixel light spots; calculating the second pixel light spot corresponding to the pixel of the original slice image through the second light field distribution, and the second pixel light spot corresponds to at least one pixel of the to-be-printed slice image; splicing the second pixel light spots to obtain the to-be-printed slice image.

[0010] In some embodiments, the step of obtaining the second light field distribution of the first pixel light spot according to the first light field distribution includes: for each first sub-pixel light spot, superimposing the light intensities of the first light field distributions of the plurality of first sub-pixel light spots at the first sub-pixel light spot to obtain the second light field distribution.

[0011] In some embodiments, the step of calculating the second pixel light spot corresponding to the pixel of the original slice image through the second light field distribution includes: if the pixel in the original slice image is bright, the light field distribution of the second pixel light spot is the second light field distribution.

[0012] In some embodiments, the step of splicing the second pixel light spots to obtain the to-be-printed slice image includes: for each second pixel light spot, if adjacent second pixel light spots have an overlap, superimposing and calculating the light intensity of the overlap of the adjacent second pixel light spots.

[0013] In some embodiments, resampling the to-be-printed slice image is performed to make the size of the to-be-printed slice image the same as the size of the original slice image.

[0014] In some embodiments, the original slice image is the label image of the neural network model, and the to-be-printed slice image is the input image of the neural network model.

[0015] This application can generate a to-be-printed slice image corresponding to the original slice image, and use the original slice image and the to-be-printed slice image as a training set to train the neural network model. The trained neural network model can learn the mapping relationship between the to-be-printed slice image and the original slice image. By inputting the current to-be-printed slice image, the corrected neural network model can output the corrected slice image. Therefore, this application can utilize the learning ability of the neural network to reduce the influence of light spot overlap, can generate fine features of the slice image, can eliminate overexposure at the edges of larger features, and correct the slice image of complex graphics, thereby improving the accuracy of the mask pattern and the accuracy of 3D printing. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The accompanying drawings are included to provide a further understanding of the present application, and they are incorporated in and constitute a part of this application. The drawings illustrate embodiments of the present application and, together with the description, serve to explain the principles of the present application. In the drawings:

[0017] Figure 1 is the original slice image in a 3D model provided by an embodiment of the present application;

[0018] Figure 2 is Figure 1 the slice image to be printed after the distortion of the original slice image shown;

[0019] Figure 3 is an exemplary flowchart of a method for correcting a slice image for 3D printing provided by an embodiment of the present application;

[0020] Figure 4 is a schematic diagram of some pixel points of the original slice image formed by a 3D printer provided by an embodiment of the present application;

[0021] Figure 5 is a schematic diagram of the first sub-pixel light spot formed by a 3D printer provided by an embodiment of the present application;

[0022] Figure 6 is a partial first sub-pixel light spot measured by a microscope camera provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0023] To more clearly illustrate the technical solutions of the embodiments of the present application, the accompanying drawings required for the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some examples or embodiments of the present application. For those of ordinary skill in the art, without creative efforts, the present application can also be applied to other similar scenarios based on these drawings. Unless obvious from the language context or otherwise stated, the same reference numerals in the drawings represent the same structure or operation.

[0024] As shown in the present application, unless the context clearly indicates an exceptional situation, words such as "a", "an", "one", and / or "the" are not specifically singular and may also include the plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0025] Unless otherwise specifically stated, the relative arrangements of components and steps, numerical expressions, and numerical values set forth in these embodiments do not limit the scope of the present application. At the same time, it should be understood that, for the sake of convenience of description, the dimensions of the various parts shown in the drawings are not drawn in actual proportional relationships. Technologies, methods, and devices known to those of ordinary skill in the relevant art may not be discussed in detail, but where appropriate, the said technologies, methods, and devices should be regarded as part of the specification. In all the examples shown and discussed here, any specific values should be construed as merely exemplary and not as a limitation. Therefore, other examples of the exemplary embodiments may have different values. It should be noted that like reference numerals and letters denote like items in the following figures, and thus, once an item is defined in one figure, it will not be further discussed in subsequent figures.

[0026] In the description of the present application, it should be understood that the orientation or positional relationships indicated by orientation words such as "front, back, up, down, left, right", "lateral, vertical, perpendicular, horizontal", and "top, bottom", etc., are usually based on the orientation or positional relationships shown in the drawings, and are only for the convenience of describing the present application and simplifying the description. Without contrary description, these orientation words do not indicate and imply that the device or element referred to must have a specific orientation or be constructed and operated in a specific orientation, and thus should not be construed as limiting the protection scope of the present application; the orientation words "inside, outside" refer to the inside and outside relative to the contour of each component itself.

[0027] For the convenience of description, spatial relative terms such as "above...", "over...", "on the upper surface of...", "above", etc., can be used here to describe the spatial positional relationships of one device or feature and other devices or features as shown in the figures. It should be understood that the spatial relative terms are intended to encompass different orientations in use or operation in addition to the orientation depicted in the figures. For example, if the device in the figure is inverted, the device described as "above other devices or structures" or "over other devices or structures" will then be positioned "below other devices or structures" or "under other devices or structures". Thus, the exemplary term "above..." can include both the orientation of "above..." and "below...". The device can also be positioned in other different ways (rotated 90 degrees or in other orientations), and corresponding interpretations should be made for the spatial relative descriptions used here.

[0028] In addition, it should be noted that the use of words such as "first" and "second" to define components is only for the convenience of distinguishing the corresponding components. If not otherwise stated, the above words have no special meaning and cannot be understood as limiting the scope of protection of this application. In addition, although the terms used in this application are selected from well-known and commonly used terms, some terms mentioned in the specification of this application may be selected by the applicant at his or her discretion, and their detailed meanings are explained in the relevant parts of the description of this article. In addition, it is required to understand this application not only by the actual terms used, but also by the meaning implied by each term.

[0029] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously. At the same time, other operations may be added to these processes, or one or more operations may be removed from these processes.

[0030] During the 3D printing process, the 3D printer needs to divide the 3D model to be printed into a series of two-dimensional cross sections, which are usually called slice images or slice layers. The 3D printer converts the slice images into transparent and opaque areas on the mask, solidifies the resin layer by layer according to the mask, and finally prints the complete target model.

[0031] The patent disclosure (CN105690753B) of the applicant of the present application can be used to explain the relevant contents of the present application. The patent disclosure records that different beam sub-images are obtained by reducing the pixel spot size of the original beam sub-image (for example, reducing it by 1 / m), and different beam sub-images (also called spot) are sequentially offset and exposed through an offset structure to obtain an image to be printed.

[0032] In the above patent document, a Gaussian beam (also called a Gaussian spot) is used as a carrier of image information. A Gaussian beam is a light wave mode in the field of optics, characterized by a transverse electric field and an irradiance distribution that approximately satisfies a Gaussian function. A Gaussian beam has a spot radius, which refers to the time at which the light intensity on the spot is reduced to 1 / e of the maximum light intensity. 2 The radius of the ring is , and the light intensity distribution of the Gaussian spot conforms to the Gaussian distribution.

[0033] The beamlet images in the above patent documents are formed by a number of Gaussian spots. As described in the previous text, a Gaussian spot is not a theoretical "point", and its light intensity distribution is within a certain range. Therefore, in practical applications, there will be an overlap of spots between the pixels of the beamlet images formed by a number of Gaussian spots. Moreover, when different beamlet images are offset and exposed multiple times by the offset mechanism, the overlapping area may be larger, and the overlapping area will gradually increase with the number of offsets. This will result in an inability to generate fine features in the formed image, and for larger features, there will be a problem of overexposed edges. If the original sliced image contains rich fine features and larger features, the printed image generated by the above technology may be distorted, resulting in a reduction in printing accuracy.

[0034] Figure 1 There is provided an original sliced image in a 3D model, and from this original sliced image, the contour features of a cross-section of the 3D model can be obtained. As Figure 1 shown, this original sliced image contains some fine features, such as the various acute angles and various thin lines included in this figure; it also contains some larger features, such as the various rectangles with larger areas in this figure.

[0035] Figure 2 There is given Figure 1 the distorted sliced image to be printed of the original sliced image shown above after being processed by the above patent documents. By carefully observing Figure 1 and Figure 2 , it can be found that Figure 2 the contour in Figure 2 is very blurred, especially at the corners of each rectangle. It can be seen from this that the target 3D model printed according to the sliced image to be printed in Figure 2 will also be distorted, and the printing accuracy is limited.

[0036] In order to solve the above technical problems, the present application provides a method for correcting a sliced image for 3D printing. As Figure 3 shown, this correction method includes the following steps:

[0037] S100. Obtain the original sliced image;

[0038] S200. Obtain the first light field distribution of at least one first sub-pixel spot of the 3D printer;

[0039] S300. According to the first light field distribution, obtain the sliced image to be printed corresponding to the original sliced image;

[0040] S400. Train a neural network model according to the original sliced image and the sliced image to be printed to obtain a trained corrected neural network model;

[0041] S500. Input the current sliced image to be printed into the corrected neural network model to obtain a corrected sliced image.

[0042] The above steps S100 - S500 will be described in detail below with reference to the accompanying drawings.

[0043] In S100, the original slice images can be obtained by inputting different 3D model files (3D models of the workpiece to be printed) into slicing software, and multiple original slice images of each 3D model file are automatically generated and saved. Slicing software includes Curahe, Slic3r, etc. The specific process executed by the slicing software can refer to the above - mentioned open - source software, which will not be elaborated here.

[0044] Figure 4 is a schematic diagram of some pixel points of the original slice image formed by the 3D printer. As Figure 4 shown, each small square represents a light spot of the original slice image, and a light spot also represents a pixel point of the original slice image. Figure 4 The shaded area marked in represents bright information. Although the shading is not marked on all light spots in this figure, it can be understood that each light spot in the original slice image contains "bright" or "dark" information.

[0045] Figure 5 is a schematic diagram of the first sub - pixel light spot formed by the 3D printer. The first sub - pixel light spot is a light spot formed after reducing the light spot of the original slice image by a certain proportion. This reduction ratio can be 1 / 2, 1 / 3, 1 / 4, etc. Reducing the light spot of the original slice image can be achieved through a focusing lens. As Figure 5 shown, this first sub - pixel light spot is formed after reducing the original slice image by 1 / 2, and the number of its pixel points is quadrupled. Figure 5 The shaded area in and Figure 4 The shaded area in correspond, representing bright information. For example, Figure 4 a1 in and Figure 5 the a1 area in correspond. Although the shading is not marked on all the first sub - pixel light spots in Figure 5 , it can be understood that each first sub - pixel light spot contains "bright" or "dark" information.

[0046] In step S200, the first light field distribution refers to the distribution of the intensity, phase, and polarization state of the first sub-pixel light spot in space. The first light field distribution can describe the light intensity variation of the first sub-pixel light spot at different spatial positions and the focusing characteristics of the first sub-pixel light spot. The first light field distributions of each first sub-pixel light spot formed by the 3D printer may not be exactly the same. In some embodiments of the present application, the first light field distribution of each first sub-pixel light spot can be obtained. In some other embodiments of the present application, the first light field distribution of any one first sub-pixel light spot can also be obtained and used to represent the first light field distributions of other first sub-pixel light spots. In some embodiments, regions can also be divided, and the first light field distribution of any one first sub-pixel light spot is obtained in each region respectively, and the first light field distributions of the first sub-pixel light spots in each region are represented by this first light field distribution respectively.

[0047] The purpose of step S300 is to obtain the slice image to be printed corresponding to the original slice image and construct a data set for training the neural network model in step S400. The slice image to be printed needs to be obtained according to the first light field distribution of the first sub-pixel light spot. In some embodiments, the slice image to be printed is calculated according to the first light field distribution of any one first sub-pixel light spot. In some other embodiments, the slice image to be printed is calculated according to the first light field distributions of each first sub-pixel light spot. In some embodiments, the slice image to be printed is calculated according to the first light field distributions of the first sub-pixel light spots in different regions.

[0048] In some embodiments, the slice image to be printed is the input data of the neural network model, and the original slice image is the label data of the neural network model. Under the training of a large amount of data, the neural network model can obtain the mapping relationship between the slice image to be printed and the original slice image. Since the original slice image has more obvious features, such as fine small features and larger features, during the training process, the neural network model can continuously adjust the mapping relationship, so that under the adjustment of the mapping relationship, the slice image to be printed reduces the influence of the overlap of the first sub-pixel light spots, generates the small features of the slice image, and eliminates the overexposure of the edges of the larger features. After the training is completed, through step S500, the corrected slice image is obtained. The 3D printer can print through the corrected slice image, thereby improving the accuracy of the 3D printer.

[0049] In some embodiments, S200 further includes the following steps:

[0050] S210: Measure the characteristic data of the first sub-pixel light spot, and the characteristic data includes a number of light intensity values;

[0051] S220: Fit the characteristic data to obtain a mathematical model of the first sub-pixel light spot; and

[0052] S230. Obtain the first light field distribution of at least one first sub-pixel spot according to the ratio of the mathematical model to the distance of the adjacent first sub-pixel spot.

[0053] Figure 6 These are some of the first sub-pixel spots measured by the microscopic camera. The first sub-pixel spots in the figure are obtained by scaling the microscopic camera according to a certain ratio. As Figure 6 shown, the first sub-pixel spot S is not a "point", and its light intensity distribution is within a region. Since the size of the first sub-pixel spot is very small and cannot be observed by the naked eye, it is necessary to image and measure it through a microscopic camera. Among them, the characteristic data includes a number of light intensity values, and the number of light intensity values refers to the light intensity values at different positions within a first sub-pixel spot. The mathematical model of the first sub-pixel spot can be obtained by fitting a number of light intensity values.

[0054] In some embodiments, the mathematical model includes a Gaussian function and / or a polynomial. Since the Gaussian spot generally conforms to the Gaussian distribution, the Gaussian function can be used to fit the first light field distribution of the first sub-pixel spot. The polynomial can also be used to fit the first light field distribution of the first sub-pixel spot, and the specific number of terms of the polynomial is not limited here. In this embodiment, the Gaussian function and the polynomial can be fitted by the least squares regression algorithm to obtain the coefficients to be fitted.

[0055] In some embodiments, the Gaussian function and the polynomial are used to fit the first light field distribution of the first sub-pixel spot. The introduction of the polynomial is to improve the fitting accuracy of the first light field distribution. In this embodiment, the Gaussian function and the polynomial can also be fitted by the least squares regression algorithm to obtain the coefficients to be fitted. The least squares regression algorithm can be implemented by tools such as Matlab or Python, and the specific implementation process of the least squares regression algorithm will not be elaborated here.

[0056] When using the Gaussian function to fit the first light field distribution of the first sub-pixel spot, the expression of the Gaussian function is as follows:

[0057]

[0058] where I(r) is the light intensity at a distance r from the center of the spot; I0 represents the peak light intensity at the center of the spot; w is the characteristic radius of the spot, that is, the radius at which the light intensity drops to 1 / e 2 of the peak light intensity; r is the distance from the center of the spot.

[0059] This application does not limit the fitting method in step S220, and various numerical fitting methods in the art can be adopted. In some embodiments, after obtaining the light intensity values of multiple first sub-pixel light spots, the Gaussian function can be fitted by the least squares regression algorithm in numerical analysis to obtain the coefficients to be fitted, such as I0 and w.

[0060] In some embodiments of this application, the first light field distribution of each first sub-pixel light spot can be obtained. In this embodiment, several light intensity values are measured for each first sub-pixel light spot, and each first sub-pixel light spot is fitted to obtain a mathematical model of the corresponding first light field distribution.

[0061] In some other embodiments of this application, the first light field distribution of any one first sub-pixel light spot can also be obtained, and the first light field distributions of other first sub-pixel light spots are represented by this first light field distribution. In these embodiments, only the light intensity values of several first sub-pixel light spots of one first sub-pixel light spot need to be collected and fitted to obtain a mathematical model of the corresponding first light field distribution. The advantage of fitting to obtain a mathematical model is that it can reduce the amount of calculation in the fitting process and the amount of data generated during the fitting process. When calculating the to-be-printed slice image subsequently, the amount of calculation can also be significantly reduced, saving the time required for calculation.

[0062] In some embodiments, step S230 further includes the following steps:

[0063] S231. Calculate the full width at half maximum of the Gaussian function; and

[0064] S232. Obtain the first light field distribution according to the ratio of the full width at half maximum and the distance and the Gaussian function.

[0065] Among them, the full width at half maximum (FWHM) of the Gaussian function is the distance between two points when the light intensity of the Gaussian light spot reaches half of the maximum value. The full width at half maximum reflects the width of the Gaussian light spot and can be used to assist in representing the shape of the light spot. There is a proportional relationship between the full width at half maximum and the characteristic radius (w). For example, the full width at half maximum is 2.354 times the characteristic radius. The full width at half maximum of the first sub-pixel light spot can be obtained by multiplying the characteristic radius by the proportional relationship, or can be obtained by calculating the light intensity when the light spot reaches half of the maximum value and substituting it into formula (1) for calculation.

[0066] Continue to refer to Figure 6 the content shown. There is a certain distance D between adjacent first sub-pixel light spots, and this distance D is determined by the physical characteristics of the 3D printer. For different 3D printers, the distance between adjacent first sub-pixel light spots is different, and this distance D can be obtained by reading the physical parameters of the 3D printer. For example, the distance D can be the distance between each lens among the focusing lenses. It should be noted thatFigure 6 is an image magnified by a microscopic camera, which magnifies the distance between the first sub-pixel light spots by several times. In fact, there will be partial overlap between adjacent first sub-pixel light spots, and the actual distance is not Figure 6 as large as that in []. Since there is a scaling factor during measurement and the pixels of the 3D printer device also have an absolute size, it is necessary to eliminate the differences in the scaling factor and the absolute size of the 3D printer pixels through normalization. The normalization method is to divide the full width at half maximum of the first sub-pixel light spot by the distance between two adjacent first sub-pixel light spots after calculating it. Assuming that the full width at half maximum of the first sub-pixel light spot is F and the distance between two adjacent first sub-pixel light spots is D, the normalized coefficient f can be obtained, and the expression of the coefficient f is shown in formula (2):

[0067]

[0068] After obtaining the normalized coefficient f, multiply it by the Gaussian function to obtain the normalized Gaussian function, as shown in formula (3):

[0069]

[0070] In some embodiments, step S300 includes the following steps:

[0071] S310. Obtain the second light field distribution of the first pixel light spot according to the first light field distribution, where the first pixel light spot includes a plurality of first sub-pixel light spots;

[0072] S320. Calculate the second pixel light spot corresponding to the pixel of the original slice image through the second light field distribution, where the second pixel light spot corresponds to at least one pixel of the slice image to be printed;

[0073] S330. Stitch the second pixel light spots to obtain the slice image to be printed.

[0074] The purpose of step S300 is to obtain the slice image to be printed corresponding to the original slice image through simulation calculation, and this slice image to be printed corresponds to the complete image after being processed by the method of patent document (CN105690753B). Among them, in step S310, the second light field distribution of the first pixel light spot is calculated, and the first pixel light spot includes a plurality of first sub-pixel light spots. That is, calculate the second light field distribution under the influence of the mutual overlap of multiple first sub-pixel light spots within the first pixel light spot. In some embodiments, the size of the first pixel light spot needs to be greater than or equal to the size of one light spot of the slice image to be printed, but the size of the first pixel light spot cannot be less than the size of one light spot of the slice image to be printed. If the size of the first pixel light spot is less than the size of one light spot of the slice image to be printed, the influence caused by the mutual overlap of several light spots cannot be simulated and calculated.

[0075] In some embodiments, step S310 includes the following steps: For each first sub-pixel light spot, the light intensities of the multiple first light field distributions of the multiple first sub-pixel light spots at the first sub-pixel light spot are superimposed to obtain a second light field distribution. Assume that the first pixel light spot includes n first sub-pixel light spots, and the first light field distribution of one first sub-pixel light spot is used to represent the first light field distributions of the n first sub-pixel light spots, ignoring the differences between the first sub-pixel light spots. Then the superposition formula of the first pixel light spot is as follows:

[0076]

[0077] wherein, I2 represents the second light field distribution of the first pixel light spot; i refers to the i-th first sub-pixel light spot being currently calculated; j refers to the j-th first sub-pixel light spot; r i,j refers to the distance of the j-th first sub-pixel light spot relative to the i-th first sub-pixel light spot; refers to the light intensity of the j-th first sub-pixel light spot at the i-th first sub-pixel light spot. Taking the example that the first pixel light spot includes 3 first sub-pixel light spots (n = 3), the calculation process of formula (4) is explained. First, calculate the light intensity of the first sub-pixel light spot, the light intensity of the second sub-pixel light spot at the first sub-pixel light spot, and the light intensity of the third sub-pixel light spot at the first sub-pixel light spot, and superimpose the results of the three. Secondly, calculate the light intensity of the second sub-pixel light spot, the light intensity of the first sub-pixel light spot at the second sub-pixel light spot, and the light intensity of the third sub-pixel light spot at the second sub-pixel light spot, and superimpose the results of the three. Then, calculate the light intensity of the third sub-pixel light spot, the light intensity of the first sub-pixel light spot at the third sub-pixel light spot, and the light intensity of the second sub-pixel light spot at the third sub-pixel light spot, and superimpose the results of the three. Finally, superimpose the results calculated above to obtain the second light field distribution of the first pixel light spot.

[0078] In some embodiments, the first light field distribution of each first sub-pixel light spot can be calculated, and then the second light field distribution can be calculated.

[0079] In some embodiments, the first light field distribution of the first sub-pixel light spots in each region can be calculated by partitioning, and then the second light field distribution can be calculated by region.

[0080] It can be seen from formula (4) that if the first light field distributions of each first sub-pixel light spot are different, or the first light field distributions of the first sub-pixel light spots are calculated by partitioning, the calculation process is very complicated and a large amount of intermediate data will be generated. To simplify the calculation process, in a preferred embodiment, the first light field distribution of any one first sub-pixel light spot in the 3D printer is used to represent the first light field distributions of all the first sub-pixel light spots, and the second light field distribution of the first pixel light spot is calculated using this first light field distribution.

[0081] In some embodiments, at least one pixel of the slice image to be printed is calculated through step S320. When the size of the first pixel spot is equal to the size of one pixel spot of the slice image to be printed, the first pixel spot exactly corresponds to one pixel of the slice image to be printed. When the size of the first pixel spot is greater than the size of one pixel spot of the slice image to be printed, the central spot of the first pixel spot corresponds to one pixel of the slice image to be printed, and the non-central spots of the first pixel spot correspond to multiple other pixels of the slice image to be printed. The central spot refers to the spot located at the middle position of the first pixel spot. For example, assuming that the size of the first pixel spot corresponds to nine pixels of the slice image to be printed, then the pixel spot located in the third row and the third column in the first pixel spot is the central spot of the first pixel spot.

[0082] In some embodiments, step S320 further includes the following steps: If the pixel in the original slice image is bright, the light field distribution of the second pixel spot is the second light field distribution. Continue to refer to Figure 2 the slice image to be printed shown in Figure 2 where the white area represents the area to be printed, and white represents "bright". The 3D printer needs to generate a Gaussian beam to cure the photosensitive resin in the white area; Figure 2 where the black area represents the area that does not need to be printed, and the 3D printer does not need to generate a beam, that is, it is dark and there is no need to cure the photosensitive resin. And Figure 2 the slice image to be printed in Figure 1 is obtained after processing the original slice image in the above patent document. Therefore, when generating the slice image to be printed, it can be calculated according to the brightness and darkness of the known original slice image.

[0083] In some embodiments, when calculating the second pixel spot, if the second pixel spot corresponds to a bright point in the original slice image, it is necessary to calculate according to formula (4). At this time, the light field distribution of the second pixel spot is the second light field distribution; if the second pixel spot corresponds to a dark point in the original slice image, no calculation is required and the second pixel spot is dark. In some embodiments, if the central spot of the second pixel spot corresponds to a bright point in the original slice image, calculate according to formula (4), and the light field distribution of the second pixel spot is the second light field distribution; if the central spot of the second pixel spot corresponds to a dark point in the original slice image, no calculation is required and the second pixel spot is dark. If the multiple pixels of the original slice image corresponding to the non-central area spots of the second pixel spot are also dark, then the non-central area of the second pixel spot is dark. If the multiple pixels of the original slice image corresponding to the non-central area spots of the second pixel spot are bright, then the light field distribution of the non-central area spots of the second pixel spot is the second light field distribution.

[0084] In some embodiments, step S330 further includes the following steps: If adjacent second pixel light spots overlap, calculate the superimposed light intensity of the overlapping adjacent second pixel light spots. As described above, the size of the second pixel light spot can be larger than the size of one pixel of the slice image to be printed. In this case, the calculation result can be closer to the slice image to be printed formed by the offset combination of multiple beam sub-images. Because the larger the size of the second pixel light spot, the more overlapping areas there will be, and thus it is closer to the area where multiple beam sub-images are offset and overlapped. In the actual process, the second pixel light spot is calculated based on one pixel of the original slice image, and each pixel corresponds to one second pixel light spot. Therefore, multiple second pixels need to be stitched together to obtain the slice image to be printed. During the stitching process, the pixels of the slice image to be printed corresponding to the second pixel light spots or the pixels of the slice image to be printed corresponding to the centers of the second pixel light spots can be stitched one by one in the order from left to right and from top to bottom. During the stitching process, if adjacent second pixel light spots overlap, the superimposed light intensity of the overlapping adjacent second pixel light spots can be calculated to generate the final slice image to be printed. Taking the stitching of two second pixel light spots as an example, assume that one second pixel light spot corresponds to 3 pixels of the slice image to be printed. Place the center of the first second pixel light spot on the first pixel of the slice image to be printed, and then place the center of the second second pixel light spot on the second pixel of the slice image to be printed. At this time, the first second pixel light spot and the second second pixel light spot overlap on the first pixel and the second pixel of the slice image to be sliced, and the light intensities of the first second pixel light spot and the second pixel light spot are superimposed at the overlapping part. It is equivalent to that for the first pixel of the slice image to be sliced, the superimposed light intensity is the final light intensity of the first pixel, and the same is true for the second pixel of the slice image to be sliced.

[0085] In some embodiments, the correction method further includes the following steps: Resample the slice image to be printed so that the size of the slice image to be printed is the same as the size of the original slice image. Because in the training of the neural network model, the data types and sizes of the input data and the label data are the same, it is necessary to process the slice image to be printed through the resampling step. In some embodiments, the printed slice image can also be processed through a binarization step. The slice image to be printed formed by stitching the second pixel light spots is not a black and white image with only "0" and "1". The purpose of binarization is to convert the slice image to be printed into a black and white graphic. The slice image to be printed is equivalent to the original slice image after being processed by the prior art. Therefore, it is necessary to make the size of the slice image to be printed the same as the size of the original slice image through resampling.

[0086] In some embodiments, the original slice image is the label image of the neural network model, and the slice image to be printed is the input image of the neural network model. So that the neural network can learn the mapping relationship between the slice image to be printed and the original slice image, thereby achieving the purpose of correcting the slice image. The present application does not limit the specific model of the neural network model. In some embodiments, the neural network model can be SRCNN (Super-Resolution CNN), EDSR (Enhanced DeepSR), SCNet (Shift-Conv Layer Network), FSRCNN (Fast Super-Resolution CNN), etc.

[0087] The present application can generate a slice image to be printed corresponding to the original slice image, and use the original slice image and the slice image to be printed as a training set to train the neural network model. The trained neural network model can learn the mapping relationship between the slice image to be printed and the original slice image. By inputting the current slice image to be printed, the corrected neural network model can output a corrected slice image. Therefore, the present application can utilize the learning ability of the neural network model to reduce the influence of spot overlap, can generate fine features of the slice image, can eliminate overexposure at the edges of larger features, and correct the slice image of complex graphics, thereby improving the accuracy of the mask pattern and the accuracy of 3D printing.

[0088] The present application uses specific terms to describe the embodiments of the present application. Such as "one embodiment", "an embodiment", and / or "some embodiments" mean a certain feature, structure, or characteristic related to at least one embodiment of the present application. Therefore, it should be emphasized and noted that the "one embodiment" or "an embodiment" or "an alternative embodiment" mentioned twice or more at different positions in this specification does not necessarily refer to the same embodiment. In addition, certain features, structures, or characteristics in one or more embodiments of the present application can be appropriately combined.

[0089] Some aspects of the present application may be executed entirely by hardware, entirely by software (including firmware, resident software, microcode, etc.), or by a combination of hardware and software. The above-mentioned hardware or software may be referred to as "data block", "module", "engine", "unit", "component" or "system". The processor may be one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DAPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), processors, controllers, microcontrollers, microprocessors, or combinations thereof. In addition, aspects of the present application may be embodied as a computer product located in one or more computer-readable media, the product including computer-readable program code. For example, the computer-readable media may include, but is not limited to, magnetic storage devices (e.g., hard disks, floppy disks, magnetic tapes...), optical disks (e.g., compact disk CD, digital versatile disk DVD...), smart cards, and flash memory devices (e.g., cards, sticks, key drives...).

[0090] The computer-readable medium may contain a propagated data signal having computer program code embodied therein, for example, on a baseband or as part of a carrier wave. The propagated signal may take many forms, including electromagnetic, optical, or the like, or suitable combinations thereof. The computer-readable medium may be any computer-readable medium other than a computer-readable storage medium, which can communicate, propagate, or transport a program for use by being connected to an instruction execution system, apparatus, or device. The program code located on the computer-readable medium may be propagated through any appropriate medium, including radio, cable, fiber optic cable, radio frequency signal, or similar media, or any combination of the above media.

[0091] Similarly, it should be noted that, in order to simplify the presentation of the disclosure of the present application and thus help the understanding of one or more embodiments of the invention, in the foregoing description of the embodiments of the present application, sometimes multiple features are combined into one embodiment, drawing, or description thereof. However, this method of disclosure does not mean that the features required by the subject matter of the present application are more than the features mentioned. In fact, the features of the embodiments are less than all the features of the single embodiment disclosed above.

[0092] In some embodiments, numbers are used to describe components and the quantity of attributes. It should be understood that such numbers used in the description of embodiments are, in some examples, modified by the modifiers "about", "approximately" or "substantially". Unless otherwise specified, "about", "approximately" or "substantially" indicate that the said numbers are allowed a variation of ±20%. Accordingly, in some embodiments, the numerical parameters used in this application are all approximate values, and such approximate values may change according to the characteristics required by individual embodiments. In some embodiments, numerical parameters should consider the specified significant digits and adopt the method of retaining general digits. Although the numerical ranges and parameters used in some embodiments of this application to confirm the breadth of their scope are approximate values, in specific embodiments, such numerical settings are as precise as possible within the feasible range.

Claims

1. A method for correcting sliced images for 3D printing, characterized in that, Including: Obtaining an original sliced image; Obtaining a first light field distribution of at least one first sub-pixel light spot of a 3D printer; Obtaining a to-be-printed sliced image corresponding to the original sliced image according to the first light field distribution; Training a neural network model based on the original sliced image and the to-be-printed sliced image to obtain a trained corrected neural network model; And Inputting the current to-be-printed sliced image into the corrected neural network model to obtain a corrected sliced image.

2. The correction method according to claim 1, characterized in that, The step of obtaining the first light field distribution of at least one first sub-pixel light spot of the 3D printer includes: Measuring characteristic data of the first sub-pixel light spot, where the characteristic data includes a number of light intensity values; Fitting the characteristic data to obtain a mathematical model of the first sub-pixel light spot; and Obtaining the first light field distribution of at least one of the first sub-pixel light spots according to a ratio of the mathematical model to a distance between adjacent first sub-pixel light spots.

3. The correction method according to claim 2, wherein The mathematical model includes a Gaussian function and / or a polynomial.

4. The correction method according to claim 3, characterized in that, The step of obtaining the first light field distribution according to the mathematical model and a ratio of distances between two adjacent first sub-pixel light spots includes: Calculating the full width at half maximum of the Gaussian function; and Obtaining the first light field distribution according to the ratio of the full width at half maximum and the distance to the Gaussian function.

5. The correction method according to claim 4, characterized in that, The step of obtaining the to-be-printed sliced image corresponding to the original sliced image according to the first light field distribution includes: Obtaining a second light field distribution of a first pixel light spot according to the first light field distribution, where the first pixel light spot includes a plurality of the first sub-pixel light spots; Calculating a second pixel light spot corresponding to a pixel of the original sliced image through the second light field distribution, where the second pixel light spot corresponds to at least one pixel of the to-be-printed sliced image; Stitching the second pixel light spots to obtain the to-be-printed sliced image.

6. The correction method according to claim 5, wherein The step of obtaining the second light field distribution of the first pixel light spot according to the first light field distribution includes: For each first sub-pixel light spot, superimposing light intensities of a plurality of first light field distributions of the plurality of first sub-pixel light spots at the first sub-pixel light spot to obtain the second light field distribution.

7. The correction method according to claim 6, wherein The step of calculating the second pixel light spot corresponding to the pixel of the original sliced image through the second light field distribution includes: If the pixel in the original sliced image is bright, the light field distribution of the second pixel light spot is the second light field distribution.

8. The correction method according to claim 7, wherein, The step of stitching the second pixel light spots to obtain the to-be-printed sliced image includes: For each second pixel light spot, if adjacent second pixel light spots have an overlap, superimposing and calculating light intensities of the overlap of the adjacent second pixel light spots.

9. The correction method according to claim 8, characterized in that, Resampling the to-be-printed sliced image to make the size of the to-be-printed sliced image the same as the size of the original sliced image.

10. The correction method according to claim 1, characterized in that, The original sliced image is a label image of the neural network model, and the to-be-printed sliced image is an input image of the neural network model.

Images