A method for adaptively adjusting the height of a printing platform during DLP 3D printing

Abstract

A method for adaptively adjusting the printing platform's elevation during DLP 3D printing relates to the fields of target detection, target segmentation, and intelligent control technology. This method is adaptable to different resin materials and equipment. It includes the following steps: using an optical camera to capture images of markers in the printing scene; determining the marker's position and the resin and fluorinated oil liquid levels using a YOLOv5 network; determining whether there is insufficient liquid based on the liquid level information and controlling a water pump to replenish the insufficient liquid; and calculating the optimal elevation height of the printing platform for printing the next layer based on the liquid level information. This proposed method for adaptively adjusting the printing platform's elevation during DLP 3D printing can improve DLP 3D printing efficiency, reduce labor costs, and exhibit portability and printability.

Description

Technical Field

[0001] The present invention relates to target detection, target segmentation, intelligent control and simulation technologies, and specifically to technologies such as the detection and segmentation of markers, and the automatic control of resin and fluorinated oil liquid levels, thereby realizing a method for adaptively adjusting the height of the printing platform during the DLP 3D printing process. Background Art

[0002] 3D printers were first invented by American scientists in the mid-1980s. They utilize 3D printing technology to produce realistic three-dimensional objects. The basic principle is to use special consumables (glue, resin, or powder) to build a computer-predefined three-dimensional model. The process then deposits a binder to bond each layer of powder together, ultimately creating a solid 3D object.

[0003] Rapid prototyping technology, with its fast processing speed and low cost, is widely used in model creation during the product development phase. 3D printing, a type of rapid prototyping technology, first converts an object into 3D data. It then uses a bondable material, such as powdered metal or plastic, to cut and print the part layer by layer. Mold manufacturing and industrial design are used to construct models, and are now evolving into product manufacturing, forming "direct digital manufacturing." Currently, a variety of rapid prototyping processes have been developed, including stereolithography (SLA), laminated object manufacturing (LOM), fused deposition modeling (FDM), selected laser sintering (SLS), and three-dimensional printing (3DP). Stereolithography (SLA) involves scanning a laser spot with a photocurable resin point by point and line by line, creating a solid model with a layer of solidified resin. Compared to other rapid prototyping processes, SLA offers high precision, excellent surface quality, and simplified post-processing. Stereolithography is widely used, with an installed capacity exceeding 69% of the market.

[0004] For surface exposure rapid prototyping systems, not only is the cost low, but it can also achieve single-shot exposure and curing of the entire layer, significantly shortening production time and improving production efficiency. The height of the printing platform is determined by the amount of resin. During the printing process, the liquid levels of the resin and fluorinated oil change, affecting the optimal height of the printing platform for each layer. However, in traditional layered printing, the height of the printing platform for each layer is consistent and set to a large value to ensure that each layer can be fully filled with resin. However, this will cause unnecessary printing time and reduce printing efficiency. Summary of the Invention

[0005] An embodiment of the present invention provides a method for adaptively adjusting the printing platform elevation during DLP 3D printing, which is used to monitor and automatically replenish the resin and fluorinated oil liquid levels in a surface exposure rapid prototyping system, and determine the optimal printing platform elevation for each layer based on the resin and fluorinated oil liquid levels.

[0006] To implement the above system, the embodiment of the present invention adopts the following technical solutions:

[0007] A method for adaptively adjusting the height of a printing platform during DLP 3D printing comprises the following steps:

[0008] Step 100: Use a well-placed optical camera to obtain the resin tank position and record it. Then, collect image data of the markers present during the printing process and input it into the YOLOv5 network for training.

[0009] Step 200: Calculate the actual distance represented by each pixel based on the position of the resin tank;

[0010] Step 300: Input the image collected during the actual printing process into the trained network for target detection and segmentation, obtain the vertical coordinates r1 and r2 of the marker, and calculate the resin and fluorinated oil liquid levels based on the actual distance represented by each pixel;

[0011] Step 400: Determine whether the resin and fluorinated oil are missing based on their liquid level. If missing, pause printing and control the KX2701PW water pump to replenish the missing liquid. At the same time, return to step 300 to check the current liquid level until it is not missing. If not missing, proceed to step 500.

[0012] Step 500: Calculate and set the optimal lifting height of the printing platform when printing the next layer based on the obtained resin and fluorinated oil liquid levels.

[0013] The resin tank position is acquired and recorded using a pre-positioned optical camera, and image data of markers present during the printing process is collected and then input into a YOLOv5 network for training. Step 100 includes:

[0014] Use the optical camera that has been placed in a certain position to obtain the vertical coordinate y1 of the highest point and the vertical coordinate y2 of the lowest point of the resin tank;

[0015] Then, the image data during the printing process is collected and input into the YOLOv5 network for training;

[0016] The YOLOv5 network parameter settings are the same as the initial settings, with an initial learning rate of 0.01, 64 samples per batch, and a maximum number of iterations of 200.

[0017] According to the position of the resin tank, the actual distance represented by each pixel is calculated in step 200, including:

[0018] The actual distance k represented by each pixel is calculated using the formula k = M / (y2 - y1), where M is the actual height of the resin tank. Because the image's coordinate origin is in the upper left corner, y2 - y1 is the pixel height of the resin tank in the image.

[0019] The images collected during the actual printing process are input into the trained network for target detection and segmentation to obtain the vertical coordinates r1 and r2 of the markers. The resin and fluorinated oil liquid levels are calculated based on the actual distance represented by each pixel. Step 300 includes:

[0020] There are two markers in total, marker A floats above the resin liquid surface, and marker B floats above the fluorinated oil liquid surface, which are used to help determine the location of the target liquid surface.

[0021] An optical camera is mounted on the side of the resin tank to capture images of markers A and B in real time. The captured images are fed into a trained YOLOv5 network, which first performs object detection on the input image to determine the approximate location of the markers. The image is then cropped based on the detection frame to remove most background noise.

[0022] Perform target segmentation on the cropped image again, separate the marker from the image, and obtain the vertical coordinates r1 and r2 of the bottom edge of the marker in the original image. Marker A corresponds to r1, and marker B corresponds to r2;

[0023] According to the formulas m1 = (y2-r1)*k and m2 = (y2-r2)*k, where y2-r1 and y2-r2 represent the pixel heights of the two markers relative to the bottom of the resin tank, respectively, the liquid level heights m1 and m2 corresponding to the two markers are calculated using the mapping relationship K between pixels and actual distances. m1 corresponds to marker A, and m2 corresponds to marker B.

[0024] Determine whether the resin and fluorinated oil are missing based on their liquid levels. If missing, pause printing, calculate the amount of missing liquid to be replenished, control the KX2701PW water pump to replenish the missing liquid, return to step 300 to re-check the current liquid level. If not missing, proceed to step 500. Step 400 includes:

[0025] The user enters thresholds C1 and C2 based on the resin tank, corresponding to resin and fluorinated oil respectively. C2 is 1.2 times the height of the flow channel in the resin tank. C1 + C2 does not exceed 80% of the resin tank height.

[0026] To ensure that every consumption of liquid will cause the pump to replenish, and is set as the threshold for judging the absence of resin and fluorinated oil, and C1 and C2 are set as the targets for replenishment after the absence of resin and fluorinated oil;

[0027] like If the resin is missing, printing stops and the water pump is called to extract the resin into the resin tank. At the same time, the process jumps to step 300 to continue monitoring the liquid level until m1-m2>C1 and the resin is stopped.

[0028] like If the fluorinated oil is missing, printing stops, and the water pump is called to extract the fluorinated oil into the resin tank. At the same time, the process jumps to step 300 to continue monitoring the liquid level. When m2>C2, the extraction of fluorinated oil stops.

[0029] The water pump is indirectly controlled by the Arduino development board and a 12V relay. To turn on the water pump, the computer port sends a 1 character to the Arduino development board. After receiving the 1 character, the development board turns on the relay, thereby indirectly turning on the water pump. To turn off the water pump, the computer port sends a 0 character to the Arduino development board. After receiving the 0 character, the development board turns off the relay, thereby indirectly turning off the water pump.

[0030] If both the resin and the fluorinated oil are present, the process proceeds to step 500 .

[0031] The step 500 of calculating the optimal lifting height of the printing platform for printing the next layer according to the obtained resin and fluorinated oil liquid levels includes:

[0032] According to the formula Calculate the optimal lifting height, where T is the time required for resin filling, H S For optimal lifting height, v m Improve the speed of printing platform, L m is the filling distance of the resin, μ1 is the viscosity of the resin, ρ is the density of the resin, g is the acceleration of gravity, and h is the height of the liquid level of the resin, that is, h = m1-m2. The optimal lifting height of the printing platform is H when T is the smallest. S ;

[0033] Set the print platform lift height to H S , start printing the next layer.

[0034] A method for adaptively adjusting the height of a printing platform during DLP 3D printing according to an embodiment of the present invention has the following advantages:

[0035] 1) Improve the manufacturing qualification rate of DLP 3D printing;

[0036] 2) Reduce labor costs;

[0037] 3) Portability: This method can be used for automatic closed-loop control of fluorinated oil and resin and adaptive adjustment of the printing platform height for DLP 3D printing systems with different resins. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 is the image of the collected marker.

[0039] Figure 2 It is a schematic diagram of the DLP 3D printing device according to an embodiment of the present invention.

[0040] Figure 3 The present invention is a flowchart of a method for adaptively adjusting the height of a printing platform during DLP 3D printing. DETAILED DESCRIPTION

[0041] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be noted that, unless there is a conflict, the embodiments and features in the embodiments of the present application can be combined with each other in any manner.

[0042] The present invention provides a method for adaptively adjusting the height of the printing platform during DLP 3D printing. Since there are two markers that can float above the resin and fluorinated oil liquids respectively, the existing YOLOv5 target detection and target segmentation can be used to detect and segment the two markers. By using an optical camera to collect image information of the side of the printing area in real time during the printing process, Figure 1 This is an image captured by an optical camera, with the marker within the frame. YOLOv5 first detects the approximate position of the marker, crops it within the detection frame, and then performs object segmentation on the cropped image to determine the position of the marker's bottom surface within the original input image. This determines the height of the resin and fluorinated oil liquid levels. The liquid levels are then used to determine whether there is a shortage of liquid, and the water pump is controlled to replenish the insufficient liquid. The liquid levels are then used to determine the optimal lift height for the next printing platform. This method can reduce costs and improve printing efficiency and yield.

[0043] Compared to traditional DLP 3D printers, the DLP 3D printer based on this invention has two markers fixed near the side walls inside the resin tank. The markers can only move up and down. An optical camera is placed in front of it to collect image information of the marker positions during the printing process. A water pump device is also installed above the resin tank. Figure 2Schematic diagram of the principle of the DLP 3D printing equipment applicable to this method, including a Z-axis lifting unit 1010, a printing platform 1020, a marker and its fixing device 1030, a water pump 1040, a resin tank 1050, an optical camera 1060 and a DLP light source 1070. The Z-axis lifting unit is used to control the up and down movement of the printing platform. The resin and fluorinated oil are placed in the resin tank. The KX2701PW water pump is mounted on the resin tank. The water pump does not affect the lifting and printing of the printing platform. The optical camera is mounted in front of the resin tank, and the DLP light source is located below the resin tank.

[0044] Figure 3 This is a flow chart of a method for adaptively adjusting the height of a printing platform during a DLP 3D printing process according to an embodiment of the present invention.

[0045] The present invention provides a method for adaptively adjusting the height of a printing platform during DLP 3D printing, comprising:

[0046] Step 100: Use a well-placed optical camera to obtain the resin tank position and record it. Then, collect image data of the markers present during the printing process and input it into the YOLOv5 network for training.

[0047] Step 200: Calculate the actual distance represented by each pixel based on the position of the resin tank;

[0048] Step 300: Input the image collected during the actual printing process into the trained network for target detection and segmentation, obtain the vertical coordinates r1 and r2 of the marker, and calculate the resin and fluorinated oil liquid levels based on the actual distance represented by each pixel;

[0049] Step 400: Determine whether the resin and fluorinated oil are missing based on their liquid level. If missing, pause printing and control the KX2701PW water pump to replenish the missing liquid. At the same time, return to step 300 to check the current liquid level until it is not missing. If not missing, proceed to step 500.

[0050] Step 500: Calculate and set the optimal lifting height of the printing platform when printing the next layer based on the obtained resin and fluorinated oil liquid levels.

[0051] Wherein, step 100 includes:

[0052] Sub-step 110 : Using a well-positioned optical camera, obtain the vertical coordinates of the highest point and the lowest point of the resin tank.

[0053] Sub-step 120: Then collect image data during the printing process and input it into the YOLOv5 network for training.

[0054] In sub-step 130, the YOLOv5 network parameter settings are the same as the initial settings, with an initial learning rate of 0.01, 64 samples per batch, and a maximum number of iterations of 200.

[0055] Optionally, the choice of object detection and object segmentation network is not limited to YOLOv5. However, the number of image frames processed per second must be greater than 12. At the same time, the training accuracy of the YOLOv5 network (mAP@0.5) must be 0.984. The accuracy of other selected object detection and segmentation networks must be as high as possible.

[0056] Furthermore, step 200 includes:

[0057] Sub-step 210 : Calculate the actual distance k represented by each pixel according to the formula k=M / (y2-y1), where M is the actual height of the resin tank. Since the coordinate origin in the image is located in the upper left corner, it is y2-y1.

[0058] Furthermore, step 300 includes:

[0059] Sub-step 310: Figure 1 As shown, there are two markers in total, marker A floats above the resin liquid surface, and marker B floats above the fluorinated oil liquid surface, which are used to help determine the position of the target liquid surface.

[0060] The resin used in this example was LCD-5201 heat-resistant resin T1 (amber, purchase link: https: / / m.tb.cn / h.5q3Qn69?tk=f640WSGyx2LCZ3457) from Zisu Office Supplies Store in Minhang District, Shanghai, and the fluorinated oil was 1H,1H,2H,2H-perfluoroheptadecantriethylsilane from Beijing Huawei Ruike Chemical Co., Ltd. (purchase link: http: / / www.hwrkchemical.com / search.html?wd=HWG21226).

[0061] Sub-step 320: Mount an optical camera on the side of the resin tank to capture images of markers A and B in real time. The captured images are fed into a trained YOLOv5 network. The network first performs object detection on the input image to determine the approximate location of the markers. The image is then cropped based on the detection frame to remove most background noise.

[0062] Sub-step 330: perform target segmentation on the cropped image to separate the markers from the image, and obtain the vertical coordinates r1 and r2 of the bottom edges of the markers in the original image. Marker A corresponds to r1, and marker B corresponds to r2.

[0063] Sub-step 340, according to the formula m1 = (y2-r1)*k and m2 = (y2-r2)*k, where y2-r1 and y2-r2 respectively represent the pixel heights of the two markers based on the bottom of the resin tank, and the mapping relationship K between pixels and actual distances is used to calculate the liquid level heights m1 and m2 corresponding to the two markers, where m1 corresponds to marker A and m2 corresponds to marker B.

[0064] Furthermore, step 400 includes:

[0065] Sub-step 410: The user inputs thresholds C1 and C2 according to the resin tank, corresponding to resin and fluorinated oil, respectively, C2 being 1.2 times the height of the flow channel in the resin tank, and C1 + C2 not exceeding 80% of the resin tank height;

[0066] In sub-step 420, in order to ensure that every consumption of liquid will cause the water pump to replenish, and is set as the threshold for judging the absence of resin and fluorinated oil, and C1 and C2 are set as the targets for replenishment after the absence of resin and fluorinated oil;

[0067] Sub-step 430: If If the resin is missing, printing stops and the water pump is called to extract the resin into the resin tank. At the same time, the process jumps to step 300 to continue monitoring the liquid level until m1-m2>C1 and the resin is stopped.

[0068] Sub-step 440: If If the fluorinated oil is missing, printing stops, and the water pump is called to extract the fluorinated oil into the resin tank. At the same time, the process jumps to step 300 to continue monitoring the liquid level. When m2>C2, the extraction of fluorinated oil stops.

[0069] In sub-step 450, the water pump is indirectly controlled by the Arduino development board and a 12V relay. To turn on the water pump, the computer port sends a 1 character to the Arduino development board. The development board receives the 1 character, which turns on the relay, thereby indirectly turning on the water pump. To turn off the water pump, the computer port sends a 0 character to the Arduino development board. The development board receives the 0 character, which turns off the relay, thereby indirectly turning off the relay.

[0070] Optionally, the choice of development board is not limited to Arduino, but it must be able to communicate with the computer port and have at least two GND interfaces. The choice of water pump is not limited to KX2701PW, but it must ensure that the rated voltage is greater than or equal to DC 12V and the size is less than 32*32*70mm to ensure that it does not affect the printing process.

[0071] Sub-step 460: If not missing, jump to step 500.

[0072] The step 500 includes:

[0073] Sub-step 510: According to the formula Calculate the optimal lifting height, where T is the time required for resin filling, H S For optimal lifting height, v m Improve the speed of printing platform, L m is the filling distance of the resin, μ1 is the viscosity of the resin, is the inverse of the change in pressure in the x-direction. Since the gravity of the resin is the main driving force for resin filling, Replace with Where ρ is the density of the resin, h is the height of the resin liquid level, that is, h = m1-m2. The optimal lifting height of the printing platform is H when T is the smallest. S ;

[0074] Sub-step 520: Set the printing platform elevation height to H S , proceed to next layer printing.

[0075] A method for adaptively adjusting the height of a printing platform during DLP 3D printing according to an embodiment of the present invention has the following advantages:

[0076] 1) Improve the manufacturing qualification rate of DLP 3D printing;

[0077] 2) Reduce labor costs;

[0078] 3) Portability: This method can be used for automatic closed-loop control of fluorinated oil and resin and adaptive adjustment of the printing platform height for DLP 3D printing systems with different resins.

Claims

1. A method for adaptively adjusting the height of a printing platform during DLP 3D printing, characterized in that: The following steps are involved: Step 100: Use a well-placed optical camera to obtain the resin tank position and record it. Then, collect image data of the markers present during the printing process and input it into the YOLOv5 network for training. Step 200: Calculate the actual distance represented by each pixel based on the position of the resin tank; Step 300: Input the image collected during the actual printing process into the trained network for target detection and segmentation, obtain the vertical coordinates r1 and r2 of the marker, and calculate the resin and fluorinated oil liquid levels based on the actual distance represented by each pixel; Step 400: Determine whether the resin and fluorinated oil are missing based on their liquid levels. If missing, pause printing and control the KX2701PW water pump to replenish the missing liquid. Meanwhile, return to step 300 to check the current liquid level until it is not missing. If not, proceed to step 500. Step 500: Calculate and set the optimal lifting height of the printing platform for printing the next layer based on the obtained resin and fluorinated oil liquid levels; Step 500 includes the following sub-steps: According to the formula Calculate the optimal lifting height, where T is the time required for resin filling, H S For optimal lifting height, v m Improve the speed of printing platform, L m is the filling distance of the resin, μ1 is the viscosity of the resin, ρ is the density of the resin, g is the acceleration of gravity, and h is the liquid level of the resin, that is, h = m1-m2; the optimal lifting height of the printing platform is H when T is the smallest. S ; Set the print platform lift height to H S , start printing the next layer.

2. The method according to claim 1, wherein Step 100 includes the following sub-steps: Use the optical camera that has been placed in a certain position to obtain the vertical coordinates of the highest and lowest points of the resin tank; Then, the image data during the printing process is collected and input into the YOLOv5 network for training; The YOLOv5 network parameter settings are the same as the initial settings, with an initial learning rate of 0.01, 64 samples per batch, and a maximum number of iterations of 200.

3. The method according to claim 1, wherein Step 200 includes the following sub-steps: The actual distance k represented by each pixel is calculated using the formula k = M / (y2 - y1), where M is the actual height of the resin tank. When the image's coordinate origin is in the upper left corner, y2 - y1 is the pixel height of the resin tank in the image.

4. The method according to claim 1, wherein Step 300 includes the following sub-steps: There are two markers in total, marker A floats above the resin liquid level, and marker B floats above the fluorinated oil liquid level, which are used to help determine the location of the target liquid level; An optical camera is mounted on the side of the resin tank to capture images of markers A and B in real time. The captured images are fed into a trained YOLOv5 network, which first performs object detection on the input image to determine the approximate location of the markers. The image is then cropped based on the detection frame to remove some background noise. Perform target segmentation on the cropped image again, separate the marker from the image, and obtain the vertical coordinates r1 and r2 of the bottom edge of the marker in the original image. Marker A corresponds to r1, and marker B corresponds to r2; According to the formulas m1 = (y2-r1)*k and m2 = (y2-r2)*k, where y2-r1 and y2-r2 represent the pixel heights of the two markers relative to the bottom of the resin tank, respectively, the liquid level heights m1 and m2 corresponding to the two markers are calculated using the mapping relationship K between pixels and actual distances. m1 corresponds to marker A, and m2 corresponds to marker B.

5. The method according to claim 1, wherein Step 400 includes the following sub-steps: The user enters thresholds C1 and C2 based on the resin tank, corresponding to resin and fluorinated oil respectively. C2 is 1.2 times the height of the flow channel in the resin tank. C1 + C2 does not exceed 80% of the resin tank height. To ensure that every consumption of liquid will cause the pump to replenish, and is set as the threshold for judging the absence of resin and fluorinated oil, and C1 and C2 are set as the targets for replenishment after the absence of resin and fluorinated oil; like If the resin is missing, printing stops and the water pump is called to extract the resin into the resin tank. At the same time, the process jumps to step 300 to continue monitoring the liquid level until m1-m2>C1 and the resin is stopped. like If the fluorinated oil is missing, printing stops, and the water pump is called to extract the fluorinated oil into the resin tank. At the same time, the process jumps to step 300 to continue monitoring the liquid level. When m2>C2, the extraction of fluorinated oil stops. The water pump is indirectly controlled by the Arduino development board and a 12V relay. If the water pump needs to be turned on, the computer port sends a 1 character to the Arduino development board. After receiving the 1 character, the development board turns on the relay, thereby indirectly turning on the water pump. If the water pump needs to be turned off, the computer port sends a 0 character to the Arduino development board. After receiving the 0 character, the development board turns off the relay, thereby indirectly turning off the water pump. If both the resin and the fluorinated oil are present, the process proceeds to step 500 .

Images