A Defect Detection and Repair Method for Insufficient Resin Filling during DLP 3D Printing

Abstract

A defect detection and repair method for insufficient resin filling in the DLP 3D printing process relates to the fields of intelligent control and image processing, and can be adapted to different resin materials and equipment. In the present invention, an optical camera is used to collect images of the printing scene, the collected images are cropped based on the position information of the exposed area, and aligned with the sliced images; further, through a contrast enhancement method, the K-Means method is used to perform clustering segmentation on the collected printing area images, and combined with an image post-processing method, the captured defect area is mapped to the actual slice; next, aiming at the position information of the detected defect area, by changing the gray levels of different pixels at the defect position, a global optimization repair algorithm for slices is designed to achieve the repair of the defect problem caused by insufficient filling. The present invention can improve the qualification rate of DLP 3D printing, reduce the waste of resin materials and labor time costs, and is adaptable to any model with portability and printability.

Description

Technical Field

[0001] The present invention relates to intelligent control and image processing technologies, and particularly to detecting pictures of the printing process collected by an optical camera, analyzing whether there are defective areas in the exposure images of single-layer slices through comparison, and if there are defects, designing a defect repair algorithm to repair the defective areas. Background Art

[0002] 3D printing technology refers to a device that uses 3D printing technology to produce a real three-dimensional model. Its basic principle is to use special consumables (such as glue, resin, or powder) to form each layer of powder into a shape by depositing a binder according to a three-dimensional solid model pre-designed by a computer, and finally print out a 3D entity. Currently, a variety of different 3D printing forming processes have been developed, such as stereolithography (SLA), laminated object manufacturing (LOM), fused deposition modeling (FDM), selective laser sintering (SLS), three-dimensional printing (3DP), surface exposure printing, etc. The surface exposure 3D printing technology based on digital light processing (DLP) uses photosensitive resin as the curing material and uses a DLP light machine or a projector as the light source, which can expose layer by layer, and has relatively rapidly improved the printing speed and printing accuracy compared with other printing methods. However, the unstable quality of printed products has always been a factor hindering the development of DLP 3D printing. Among them, the depression caused by the insufficient filling area is one of the reasons for the failure of DLP printing. The reason is that after each layer is printed during the printing process, the resin needs to be fully replenished to consume part of it for printing the next layer. However, due to the loss of resin during the printing process or unreasonable printing parameter settings, the resin filling is insufficient. Although there are some methods that use color-changing materials or advanced sensors to sense the DLP printing state, most of these methods stop printing after discovering possible problems during the printing process, which also causes a waste of part of the material and time costs, and even affects the further application of DLP 3D printing technology. Summary of the Invention

[0003] An embodiment of the present invention will provide a method for defect detection and repair of insufficient resin filling during the DLP 3D printing process. It mainly extracts the features of the current printing slice image in real time, uses the brightness difference for clustering segmentation to judge whether there is a filling defect, and at the same time designs a globally optimized target repair algorithm. According to the relative position of the defective area, by changing the gray value and exposure time of different areas of the next layer slice image, the defect is repaired. This method can save materials and improve the printing efficiency and product qualification rate.

[0004] To solve the above technical problems, the embodiments of the present invention adopt the following technical solutions:

[0005] A method for defect detection and repair of insufficient resin filling during the DLP 3D printing process includes the following steps:

[0006] Step 100: Project a color picture using a DLP light source, collect the projected image through an optical camera with a fixed position, and process it to align the projected picture I(x, y) with the captured image, obtaining a new picture I′(x′, y′).

[0007] Step 200: Use an optical camera to collect the current layer slice image during the printing process in real time, enhance the contrast of this image, then perform clustering segmentation and processing on it using the brightness difference, subtract it from the previously collected image of the previous layer, and compare it with the difference of the actual slice image to determine whether there is a defective area. If not, continue printing the next layer. If it is determined that there is a defect, map it to the next slice according to the relationship between the projected picture and the captured picture.

[0008] Step 300: Based on the position of the defect in the slice image, on the premise of ensuring the repair of the defective area, with the goal of ensuring the printing quality of the overall slice layer and minimizing the increase in printing time as much as possible, according to the next slice image to be printed, optimize the gray values of different pixels at the corresponding defect position, thereby generating a new exposure image and exposure time to achieve the repair of the defective area.

[0009] Step 400: Replace the repaired slice and exposure time with the next slice to be printed and the printing parameters respectively, and send them to the printer. At the same time, continue to monitor whether there are defect problems in the subsequent slices until the final printing is completed.

[0010] Among them, Step 100 of using a DLP light source to project a color picture, collecting the projected image through an optical camera with a fixed position, and processing it to align the projected picture I(x, y) with the captured image, obtaining a new picture I′(x′, y′) includes:

[0011] Project a color picture with four sides of different colors and pixel values all being 1 using a DLP light source.

[0012] Collect the image projected by the DLP light source through an optical camera with a fixed position, identify the slope of different color sides to calculate the position corresponding relationship between the collected image and the actual picture, and record the rotation angle of the image as θ.

[0013] According to formula (1), rotate the original image I(x, y) to obtain a new image I′(x′, y′) corresponding to the position of the projected image:

[0014]

[0015] An optical camera is used to collect the current layer slice image in the printing process in real time, and the contrast of this image is enhanced. Then, clustering segmentation and processing are performed on it using the brightness difference. The difference is taken between it and the image collected in the previous layer, and compared with the difference of the actual slice image to determine whether there is a defective area. If not, the next layer is printed continuously. If it is determined that there is a defect, it will be mapped to the next slice according to the relationship between the projected picture and the captured picture. Step 200 includes:

[0016] An optical camera is used to collect the current layer slice image in the printing process in real time, and it is cropped according to the exposure area to reduce unnecessary interference;

[0017] The obtained printing image is binarized, and the gray value corresponding to each pixel point is linearly stretched to achieve the effect of contrast enhancement. Then, K-Means is used for clustering segmentation according to the gray difference of the collected image;

[0018] The segmented image is subjected to M-order median filtering to reduce the interference of individual reflective points on the resin liquid surface to the detection. Then, a K*J all-1 rectangular structure element is used to scan each pixel in the image, and the AND operation is performed between each pixel in the structure element and the pixel it covers. If all are 1, then the pixel is 1, otherwise it is 0;

[0019] Then, the same-sized structure element is used to scan each pixel in the image, and the AND operation is performed between each pixel in the structure element and the pixel it covers. If all are 0, then the pixel is 0, otherwise it is 1. Through this processing, the interference of noise points can be eliminated and the shape boundary can be smoothed;

[0020] The currently processed collected image is subtracted from the image collected in the previous layer that has also been processed as above, and compared with the difference of its corresponding slice image. If the matching degree of the comparison image is greater than 95%, it is considered that there is no defect and printing continues, and the next layer printing image is collected and step 200 is entered for continued detection;

[0021] If the matching degree of the difference comparison image is less than 95%, the defective area is marked on the next slice image to be printed that has been rotated by formula (1), and step 300 is entered.

[0022] According to the position of the defect in the slice image, on the premise of ensuring the repair of the defective area, with the goal of ensuring the overall printing quality of the slice layer and increasing the printing time as little as possible, according to the next slice image to be printed, the gray values of different pixel points corresponding to the defective position are optimized, so as to generate a new exposure image and exposure time, and realize the repair of the defective area. Step 300 includes:

[0023] According to the sliced image with defect areas marked obtained in step 200, at this time, let S1 be the non-defect area and S2 be the defect area;

[0024] Use a power meter to measure the mapping relationship T(I′(x′,y′)) between different grayscales and light intensities in the exposure plane;

[0025] On the premise of ensuring the repair of the defect area, with the goal of ensuring the overall printing quality of the sliced layer and minimizing the increase in printing time as much as possible, design the repair function shown in formula (2) according to the relative position of the defect area in the next sliced image to be printed:

[0026]

[0027] Among them, MIN in formula (2) represents taking the minimum value of the power of the defect area. At the same time, in defect repair, the image belonging to the non-defect area S1 only performs a smoothing operation with the defect area S2. sm(I′(x′,y′)) is a parameter used to maintain the smooth transition of the grayscale of the defect area and the non-defect area image, as shown in formula (3):

[0028]

[0029] At this time, the exposure time of the defect area can be calculated through the Jacobs working curing equation, as shown in formulas (4), (5), and (6):

[0030]

[0031]

[0032]

[0033] Among them, C d is the curing depth (μm), D p is the transmission depth (μm), E o is the cumulative light energy (mJ / cm 2 ) accumulated by the DLP light source at the curing interface at time t, E c is the critical exposure amount for resin curing under ultraviolet light (mJ / cm 2 ), D p and E c can both be measured through experiments;

[0034] The I′(x′,y′) obtained at this time is the image after defect repair, and t(x′,y′) is the exposure time of each pixel point, which is input into step 400.

[0035] Replace the next slice to be printed and the printing parameters with the repaired slice and exposure time respectively, and input them into the printer. At the same time, continue to monitor whether there are defects in the subsequent slices until the final printing is completed. Step 400 includes:

[0036] Input the repaired slice I′(x′,y′) and exposure time t(x′,y′) obtained in step 300 into the DLP 3D printer, replace the original printing slice and control parameters, and print the current layer;

[0037] Determine whether it is the last layer of the slice. If not, continue the detection until the final printing is completed.

[0038] A method for detecting and repairing the defect of insufficient resin filling in the DLP 3D printing process according to an embodiment of the present invention has the following advantages:

[0039] 1) Improve the manufacturing qualification rate of DLP 3D printing;

[0040] 2) Reduce the waste of resin materials and labor time costs;

[0041] 3) Portability. The DLP 3D printing system using different projection devices and photosensitive resins can all adopt this method to detect and repair the insufficient resin filling. Therefore, the present invention has certain application value and significance. Description of the Drawings

[0042] Figure 1 It is a schematic diagram of the DLP 3D printing device according to an embodiment of the present invention.

[0043] Figure 2 It is a flowchart of a method for detecting and repairing the defect of insufficient resin filling in the DLP 3D printing process according to an embodiment of the present invention.

[0044] Figure 3 It is a physical diagram of the printing process. Among them Figure 3 (a) shows the printed model, Figure 3 (b) the 15th layer slice of the model, Figure 3 (c) the 16th layer slice of the model, Figure 3 (d) the result of subtracting the slices.

[0045] Figure 4 It shows a schematic diagram of slice repair. Detailed Embodiments

[0046] The embodiments of the present invention will be described in detail below with reference to the drawings. It should be noted that, without conflict, the embodiments and features in the embodiments of the present application can be combined with each other arbitrarily.

[0047] The present invention provides a method for defect detection and repair of insufficient resin filling during DLP 3D printing. By using an optical camera to collect image information of the printing surface in real time during the printing process, enhancing the contrast of the collected printing area image, using the K-Means method to perform clustering segmentation on the image, and cooperating with image post-processing methods, the captured defect area is mapped to the actual slice. At the same time, a global optimization repair algorithm for slices is designed to achieve the detection and repair of defect problems caused by insufficient resin filling.

[0048] Compared with traditional DLP 3D printers, the DLP 3D printer based on the present invention places an optical camera under the resin tank to collect slice image information during the printing process. Figure 1 It is a schematic diagram of the principle of the DLP 3D printing device applicable to this method, including a Z-axis lifting unit 1010 for controlling the up and down movement of the printing platform 1020. The resin material for curing is placed in the resin tank 1030, and the DLP light source 1040 and the optical camera 1050 are respectively located under the resin tank.

[0049] Figure 2 It is a flowchart of the method for defect detection and repair of insufficient resin filling during DLP 3D printing according to an embodiment of the present invention.

[0050] An embodiment of the present invention proposes a method for defect detection and repair of insufficient resin filling during DLP 3D printing, including:

[0051] Step 100: Use the DLP light source to project a color picture, collect the projected image through the optically fixed camera, and perform processing to align the projected picture I(x, y) with the captured image to obtain a new picture I′(x′, y′).

[0052] Step 200: Use the optical camera to collect the current layer slice image in real time during the printing process, enhance the contrast of this image, then perform clustering segmentation and processing on it using the brightness difference, subtract it from the image collected in the previous layer, and compare it with the difference of the actual slice image to determine whether there is a defect area. If not, continue printing the next layer. If it is determined that there is a defect, it will be mapped to the next slice according to the relationship between the projected picture and the captured picture.

[0053] Step 300: According to the position of the defect in the slice image, on the premise of ensuring the repair of the defect area, aiming to ensure the printing quality of the overall slice layer and increase the printing time as little as possible, according to the next slice image to be printed, optimize the gray values of different pixels at the corresponding defect position, so as to generate a new exposure image and exposure time to achieve the repair of the defect area.

[0054] Step 400: Replace the repaired slice and exposure time with the next slice to be printed and the printing parameters respectively, and transmit them to the printer. At the same time, continue to monitor whether there are defects in the subsequent slices until the final printing is completed.

[0055] Among them, processing step 100 includes:

[0056] Sub-step 110: Use a DLP light source to project a color picture with different colors on the four sides and pixel values all being 1.

[0057] Sub-step 120: Collect the image projected by the DLP light source through an optical camera with a fixed position, identify the slopes of different color edges, calculate the position correspondence relationship between the collected image and the actual picture for different edges, and record the rotation angle of the image as θ at the same time.

[0058] Sub-step 130: According to formula (1), rotate the original image I(x, y) to obtain a new image I′(x′, y′) corresponding to the position of the projected image:

[0059]

[0060] Furthermore, step 200 includes:

[0061] Sub-step 210: Use an optical camera to collect the current layer slice image during the printing process in real time, and crop it according to the exposure area to reduce unnecessary interference.

[0062] Sub-step 220: Perform binarization on the obtained printing image, linearly stretch the gray value corresponding to each pixel point to achieve the effect of contrast enhancement, and then use K-Means for clustering segmentation according to the gray difference of the collected image.

[0063] Sub-step 230: Perform M-order median filtering on the segmented image to reduce the interference of individual reflective points on the resin liquid surface to the detection. Then, use a K*J all-1 rectangular structuring element to scan each pixel in the image, and perform an AND operation on each pixel in the structuring element with the pixels it covers. If all are 1, then this pixel is 1, otherwise it is 0.

[0064] Optionally, the values of M, K, and J are determined according to the actual camera resolution and the recognition accuracy requirements. Here, M is taken as 5, and both K and J are taken as 3.

[0065] Sub-step 240: Then, use a structuring element of the same size to scan each pixel in the image, and perform an AND operation on each pixel in the structuring element with the pixels it covers. If all are 0, then this pixel is 0, otherwise it is 1. Through this processing, the interference of small noise points can be eliminated and the shape boundary can be smoothed.

[0066] Sub-step 250: Take the difference between the currently processed acquired image and the image acquired in the previous layer that has also undergone the above processing, and compare it with the difference of its corresponding slice image. Here Figure 3 (a) shows the printed model, Figure 3 (b) The 15th layer slice of the model, Figure 3 (c) The 16th layer slice of the model, Figure 3 (d) The result of subtracting the slices. If they are consistent, it is considered that there are no defects and printing continues, and the next layer of printed image is acquired and step 200 is entered for continued detection.

[0067] Sub-step 250: If the difference comparison images are inconsistent, mark the defective area on the next slice image to be printed that has been rotated by formula (1), and enter step 300.

[0068] Furthermore, step 300 includes:

[0069] Sub-step 310: According to the slice image with the defective area marked obtained in step 200, at this time, let S1 be the non-defective area and S2 be the defective area.

[0070] Sub-step 320: Use a optical power meter to measure the mapping relationship T(I′(x′,y′)) between different grayscales and light intensities in the exposure plane.

[0071] Sub-step 330: On the premise of ensuring the repair of the defective area, with the goal of ensuring the overall printing quality of the slice layer and increasing the printing time as little as possible, design the repair function shown in formula (2) according to the relative position of the defective area in the next slice image to be printed:

[0072]

[0073] Where MIN in formula (2) represents taking the minimum value of the power of the defective area. At the same time, in defect repair, the image belonging to the non-defective area S1 only performs a smoothing operation with the defective area S2. sm(I′(x′,y′)) is a parameter used to keep the gray levels of the defective area and the non-defective area images smooth, as shown in formula (3):

[0074]

[0075] Sub-step 340: At this time, the exposure time of the defective area can be calculated through Jacobs' working curing equation, as shown in formulas (4)(5)(6):

[0076]

[0077]

[0078]

[0079] Among them, C d is the curing depth (μm), D p is the transmission depth (μm), E o is the accumulated light energy (mJ / cm 2 ) of the DLP light source at the curing interface within time t, E c is the critical exposure dose (mJ / cm 2 ) for resin curing under ultraviolet light, D p and E c can both be measured through experiments;

[0080] Sub-step 350: At this time, the obtained I′(x′, y′) is the image after defect repair, and t(x′, y′) is the exposure time of each pixel point. Input them into step 400, where Figure 4 is a schematic diagram of slice repair. In the figure, the density of the hatching represents the level of the gray value. The darker the hatching, the lower the gray value.

[0081] The said step 400 includes:

[0082] Sub-step 410: Input the repaired slice I′(x′, y′) and the exposure time t(x′, y′) obtained in step 300 into the DLP 3D printer, replace the original printing slice and control parameters, and print the current layer;

[0083] Sub-step 420: Determine whether it is the last layer of the slice. If not, continue the detection until the final printing is completed.

[0084] A method for defect detection and repair of insufficient resin filling during DLP 3D printing according to an embodiment of the present invention has the following advantages:

[0085] 1) Improve the manufacturing qualification rate of DLP 3D printing;

[0086] 2) Reduce the waste of resin materials and labor time costs;

[0087] 3) Portability. The DLP 3D printing system using different projection devices and photosensitive resins can all adopt this method for detecting and repairing insufficient resin filling. Therefore, the present invention has certain application value and significance.

Claims

1. A defect detection and repair method for insufficient resin filling during the DLP 3D printing process, characterized in that, It includes the following steps: Step 100: Project a color picture using a DLP light source, collect the projected image through an optical camera with a fixed position, and perform processing to align the projected picture I(x, y) with the captured image to obtain a new picture I ′ (x ′ ,y ′ ); Step 200: Use an optical camera to collect the current layer slice image during the printing process in real time, enhance the contrast of this image, then perform clustering segmentation and processing on it using the brightness difference, subtract it from the image collected in the previous layer, and compare it with the difference of the actual slice image to determine whether there is a defective area. If not, continue to print the next layer. If it is determined that there is a defect, map it to the next slice according to the relationship between the projected picture and the captured picture; Step 300: According to the position of the defect in the slice image, based on the next slice image to be printed, generate a new exposure image and exposure time by optimizing the gray values of different pixels at the corresponding defect position, so as to repair the defective area; Step 400: Replace the repaired slice and exposure time with the next slice to be printed and the printing parameters respectively, and send them to the printer. At the same time, continue to monitor whether there are defect problems in the subsequent slices until the final printing is completed.

2. The method according to claim 1, wherein Step 100 includes the following sub-steps: Use a DLP light source to project a color picture with different colors on the four sides and pixel values all being 1; Collect the image projected by the DLP light source through an optical camera with a fixed position, identify the slope of different color edges to calculate the position correspondence relationship between different edges of the collected image and the actual picture, and record the rotation angle of the image as θ; According to formula (1), the original image I(x, y) is rotated to obtain a new image I corresponding to the position of the projection image ′ (x ′ , y ′ ):

3. The method according to claim 1, wherein Step 200 includes the following sub-steps: Perform binarization operation on the obtained printing image, linearly stretch the gray value corresponding to each pixel point, and then use K-Means for clustering segmentation according to the gray difference of the collected image; Perform M-order median filtering on the segmented image, and then use a full-1 rectangular structure element of K*J to scan each pixel in the image. Perform an AND operation on each pixel in the structure element and the pixels it covers. If all are 1, then this pixel is 1, otherwise it is 0; Then use a structure element of the same size to scan each pixel in the image again, perform an AND operation on each pixel in the structure element and the pixels it covers. If all are 0, then this pixel is 0, otherwise it is 1; Subtract the currently processed collected image from the image collected in the previous layer that has also been processed as above, and compare it with the difference of its corresponding slice image. If the matching degree of the comparison image is greater than 95%, it is considered that there is no defect and continue to print, and collect the next layer printing image to enter Step 200 for continued detection; If the matching degree of the difference comparison image is less than 95%, mark the defective area on the next slice image to be printed that has been rotated by formula (1), and enter Step 300.

4. The method according to claim 1, wherein Step 300 includes the following sub-steps: According to the slice image marked with the defective area obtained in Step 200, at this time, record S1 as the non-defective area and S2 as the defective area; Use a light power meter to measure the mapping relationship T(I′(x′,y′)) between different gray levels and light intensity in the exposure plane; Design the repair function shown in formula (2) according to the relative position of the defective area in the next slice image to be printed: Among them, MIN in formula (2) represents taking the minimum value of the power of the defective area. At the same time, in defect repair, the image belonging to the non-defective area S1 only performs a smoothing operation with the defective area S2. sm(I′(x′,y′)) is a parameter used to maintain the smooth transition of the gray levels of the images in the defective area and the non-defective area, as shown in formula (3): At this time, the exposure time of the defective area is obtained by calculating through Jacobs' working curing equation, as shown in formulas (4), (5), and (6): Among which C d is the curing depth (μm), D p is the transmission depth (μm), E o is the cumulative light energy (mJ / cm 2 ) accumulated by the DLP light source at the curing interface within time t, E c is the critical exposure amount for resin curing under ultraviolet light (mJ / cm 2 ), D p and E c are both measured through experiments; At this time, the obtained I′(x′,y′) is the image after defect repair, and t(x′,y′) is the exposure time of each pixel point, which is input into step 400.

5. The method according to claim 1, wherein, Step 400 includes the following sub-steps: Input the repaired slice I′(x′,y′) and the exposure time t(x′,y′) obtained in step 300 into the DLP 3D printer, replace the original printing slice and control parameters, and print the current layer; Judge whether it is the last layer of the slice. If not, continue the detection until the final printing is completed.

6. The method according to claim 1, wherein In order to ensure that the optical camera can collect the image information during the printing process in real time without missing any inspection, the selected frame rate of the optical camera needs to be greater than 30 FPS.

Images