Three-dimensional copying method based on structured light three-dimensional reconstruction

Abstract

The invention discloses a three-dimensional copying method and system based on MEMS galvanometer type structured light three-dimensional reconstruction, and the system breaks through the limitation that only plane images can be obtained through traditional two-dimensional photographing, and successfully achieves the high-precision three-dimensional digitization and materialization restoration of an object. The method comprises the following steps: firstly, developing a novel MEMS galvanometer type structured light three-dimensional reconstruction technology, modulating line laser by using an MEMS galvanometer to project a stripe pattern at a high speed, collecting a stripe sequence on the surface of a measured object, and resolving to generate three-dimensional point cloud data after processing such as system calibration; carrying out multi-view splicing, gridding processing and rendering processing on the 3D printing model file to obtain a model file adaptive to 3D printing; and finally, the model file is imported into a 3D printer, and accurate copying of the target entity is completed. According to the method, the integrated process from three-dimensional digital scanning to entity copying is realized, the scanning modeling efficiency is high, the copying precision is good, and the method can adapt to the three-dimensional record restoration requirements of various daily entities, so that the method has good practical value and popularization prospect.

Description

Technical Field

[0001] This invention relates to the field of structured light 3D reconstruction and additive manufacturing technology, and in particular to a stereoscopic replica solid printing system and method based on MEMS galvanometer fringe projection. Background Technology

[0002] The technical problem to be solved by the present invention is a three-dimensional replica entity solution based on MEMS galvanometer fringe projection, which can be used for three-dimensional digital scanning, modeling and 3D printing of objects in the real world, to replace the traditional two-dimensional photographic planar recording method.

[0003] Previously, the main technologies for physical 3D replication were as follows, but all of them had obvious drawbacks: Method a) After constructing a high-precision geometric model using professional 3D modeling software such as SolidWorks, the model needs to undergo topology optimization and format standardization before being imported into a 3D printing machine for layer-by-layer deposition to form the solid object. The core limitation of this technique is that it must be based on a pre-set digital model and cannot directly obtain and copy the 3D structural information of the physical object.

[0004] Method b) first uses a 3D reconstruction system to scan the real object, and then sends the obtained data to a 3D printer for printing. This method involves fragmented steps and does not yet form a truly integrated closed loop for 3D replication. In existing physical replication solutions, scanning, modeling, and 3D printing are mostly independent processes, with different data formats for each step, requiring manual format conversion and model correction. This process is not only cumbersome but also highly susceptible to human error, thus affecting replication accuracy.

[0005] Method c) uses a DLP-based structured light 3D reconstruction system for 3D scanning. This method is slow in terms of fringe projection, has low pixel resolution, and a small depth of field, which is detrimental to the real-time performance and accuracy of 3D scanning. Furthermore, its large size hinders system miniaturization and widespread adoption in consumer applications. More importantly, the core of DLP relies on the DMD chip developed by Texas Instruments, which is detrimental to the independent control of high-end domestically produced instruments. Additionally, the galvanometers used in current mainstream scanning schemes lack an angle compensation mechanism, making them prone to fringe projection shift during high-frequency oscillation, affecting the stability and accuracy of 3D reconstruction. Summary of the Invention

[0006] To overcome one or more of the shortcomings of previous methods, this invention provides a stereoscopic replica printing method based on MEMS galvanometer structured light 3D reconstruction, aiming to realize an integrated process from 3D scanning and modeling to 3D printing, and improve the accuracy, efficiency and ease of use of the replica.

[0007] The core concept of this invention lies in using a MEMS galvanometer to modulate a line laser to project standard sinusoidal fringes onto an object. Angle detection and pulse feedback are used to compensate for galvanometer oscillation in real time, ensuring the stability and accuracy of the projected fringes. This invention converts the acquired fringe image sequence into 3D point cloud data through system calibration, phase recovery, and phase masking. Then, through point cloud denoising, stitching, rendering, and meshing, a regular 3D point cloud model is obtained. Its core lies in the end-to-end automated output of the entire process—from fringe projection, image acquisition, 3D reconstruction to model optimization—through a fully integrated control system. This process requires no manual intervention; data flows seamlessly and is precisely synchronized in time at each stage. It can directly construct model files suitable for 3D printing and promptly transmit them to the 3D printer for accurate physical replication, truly forming an integrated forming and replication process from stereoscopic photography to physical object creation.

[0008] Compared with the prior art, the beneficial effects of the present invention include: (1) The three-dimensional reconstruction effect is good. The present invention uses a MEMS galvanometer structure to modulate the line laser, which has the characteristics of high spatial resolution and can accurately collect the stripe image of the solid surface. Combined with the three-frequency 12-step phase shift method, the accuracy and efficiency of the three-dimensional point cloud data acquisition are greatly improved.

[0009] (2) The system is small in size and easy to promote and popularize. Compared with the traditional scanning structure, the MEMS galvanometer structure is much smaller in size and does not require complicated supporting equipment. It is more suitable for the application needs of consumer scenarios than the traditional structure and has good prospects for promotion.

[0010] (3) Integrated end-to-end process. This invention establishes an integrated end-to-end process from projection stripes, image acquisition, 3D reconstruction to model optimization and 3D printing replication. The various links in this system are interconnected, breaking the drawbacks of the fragmented nature of traditional technologies. The entire process does not require manual intervention in format conversion, parameter adjustment, and model correction, greatly reducing the difficulty of operation and avoiding the introduction of human error.

[0011] (4) Breaking through the limitations of two-dimensional photography technology. Traditional two-dimensional photography can only record planar images, while this invention breaks through this limitation and can accurately capture various angles and details of physical objects, and record and restore them completely in the form of three-dimensional digital models and physical objects, thus meeting the user's need for complete preservation of physical objects.

[0012] (5) Strong generalization ability: It can be applied to most objects in the real world, with little restriction on the material, color and size of the object, and can be applied to many application scenarios such as industrial design, cultural relic protection, and medical modeling. Attached Figure Description

[0013] Figure 1 This is a schematic diagram of the principle of 3D copying based on MEMS galvanometer structured light 3D reconstruction.

[0014] Figure 2 Structure diagram of MEMS galvanometer-based structured light 3D reconstruction system Figure 3 This is a flowchart of the 3D photocopying process. Detailed Implementation

[0015] The technical solution of the present invention will be further described in detail below with reference to specific embodiments.

[0016] Example 1: Stereoscopic Replica System Based on MEMS Galvanometer Fringe Projection The system in this embodiment is an integrated physical replication system for converting 2D photography into 3D stereoscopic modeling, specifically including: (1) MEMS galvanometer-type stripe projection scanning unit. It includes an FPGA control subunit, a MEMS galvanometer unit, a line laser emission subunit, an industrial-grade high-speed camera, an angle detection subunit, and a pulse feedback subunit. The FPGA control subunit outputs a drive signal to control the MEMS galvanometer unit to swing at a fixed frequency, modulating the line laser emission subunit to project a periodic stripe pattern; the angle detection subunit collects the galvanometer swing angle in real time, and the pulse feedback subunit sends the angle deviation signal back to the FPGA to realize dynamic compensation of the galvanometer swing, ensuring the stability of stripe projection and providing a foundation for stereoscopic imaging to acquire high-quality object surface stripe image sequences.

[0017] (2) 3D Reconstruction and Model Optimization Unit. Using supporting software developed in C++, the stripe image sequence acquired by the camera is systematically calibrated. Phase recovery is then achieved using a three-frequency 12-step phase-shifting method, and the absolute phase information of the object surface is calculated using a phase masking method to generate 3D point cloud data. After outlier removal, meshing, rendering, and hole filling, the obtained data yields a 3D model file in STL format.

[0018] (3) 3D Printing Execution Unit. This unit supports the direct import of STL format model files. The device has a built-in intelligent parameter matching system that can automatically configure printing parameters, including layer thickness accuracy and infill density, based on the geometric parameters of the model and the characteristics of the material. Furthermore, the automated parameter matching module minimizes human intervention errors, achieving high-precision conversion of the 3D model into a physical entity, completing the entire chain from data acquisition to physical replication.

[0019] Example 2: A method for stereoscopic replicating solid objects based on MEMS galvanometer fringe projection The method in this embodiment is an integrated solid replication method for converting two-dimensional photography into three-dimensional stereoscopic modeling, specifically including the following steps: (1) Stripe projection and stereoscopic photography acquisition: Start the FPGA control subunit, drive the MEMS galvanometer unit to swing and modulate the line laser to project a periodic stripe pattern, the angle detection subunit collects the galvanometer angle in real time, and the pulse feedback subunit dynamically compensates for the swing offset; synchronously control the high-speed camera to take stereoscopic pictures of the target object after the projected stripes, collect the stripe image sequence and transmit it to the three-dimensional reconstruction and model optimization unit in real time to complete the two-dimensional image data acquisition.

[0020] (2) 3D Reconstruction and Stereo Model Forming: First, the captured images are systematically calibrated, and the phase information is calculated and a phase map is generated using the three-frequency 12-step phase shift method. Then, the absolute phase is obtained by combining phase mask processing, thus obtaining 3D point cloud data. After denoising, stitching, and topology optimization, the obtained point cloud data is finally processed to generate an STL format file adapted for 3D printing.

[0021] (3) 3D Printed Entity Replication: The STL model file generated by the 3D reconstruction is directly imported into the docked 3D printing equipment through the data interface. The built-in parameter matching module of the equipment automatically adapts the printing parameters. Among them, the layer thickness parameter can be dynamically adjusted according to the model details, and the infill density is automatically adjusted according to the structural strength requirements and material utilization. The entire parameter configuration process does not require manual intervention, and printing can begin after starting the printing program.

[0022] Implementation effect The system and method of this embodiment were used to conduct experiments on the conversion of stereoscopic photography to 3D solid modeling and replication. The system completed stereoscopic photography and 3D replication of the solid object. The 3D reconstruction speed was fast, the model topology optimization processing efficiency was excellent, and the 3D printed product had a high similarity to the original object in appearance. The entire process from stereoscopic photography to solid modeling was automatically connected smoothly and in a short time. It achieved efficient and accurate replication of 2D photography to 3D stereoscopic modeling, and the overall replication effect was significantly better than the existing technology.

Claims

1. A stereoscopic copying system based on structured light three-dimensional reconstruction, characterized in that, include: (1) A MEMS galvanometer-type stripe projection scanning unit includes an FPGA control subunit, a MEMS galvanometer unit, a line laser emission subunit, an image acquisition subunit, an angle detection subunit, and a pulse feedback subunit; the FPGA control subunit is used to drive the MEMS galvanometer unit to modulate the line laser emission subunit to project a standard sinusoidal stripe pattern, and to compensate the galvanometer swing angle in real time through the angle detection subunit and the pulse feedback subunit, and to synchronously control the image acquisition subunit to acquire the stripe image sequence of the surface of the object under test; (2) The three-dimensional reconstruction and model optimization unit is used to perform system calibration, phase recovery and phase masking on the image sequence to generate three-dimensional point cloud data, and to construct a three-dimensional model file that conforms to the 3D printing standard through algorithms such as outlier removal, meshing, rendering and hole filling, so as to complete the forming transformation from two-dimensional image to three-dimensional model; (3) 3D printing execution unit. This unit supports importing optimized STL files and has a built-in module for automatic parameter adjustment. It can automatically match parameters such as printing layer thickness, infill density and printing rate according to the shape characteristics of the model and the physical properties of the material. This largely avoids the deviation caused by manually setting parameters, ensures the accuracy of the conversion of digital models into three-dimensional entities, and ultimately realizes the stereoscopic photographic replication of real objects.

2. A stereoscopic copying method based on structured light three-dimensional reconstruction, characterized in that, Includes the following steps: Step 1: Stripe Projection and Image Acquisition: By controlling the MEMS galvanometer to modulate the line laser through FPGA, a standard sinusoidal stripe pattern is projected onto the target object at high speed. At the same time, the galvanometer swing angle is compensated in real time, and the stripe image sequence on the object surface is acquired synchronously to complete the two-dimensional photographic image acquisition, providing a data foundation for the two-dimensional to three-dimensional conversion. Step 2: 3D Reconstruction and Model Optimization: The image sequence is subjected to system calibration, phase recovery, and phase masking to generate 3D point cloud data; the point cloud data is denoised and stitched to generate a full-view 3D point cloud; then, a standard 3D model file adapted for 3D printing is constructed through point cloud post-processing algorithms such as meshing, rendering, hole filling, and model closure, realizing the stereoscopic transformation from 2D image to 3D model. Step 3: 3D Printing Replication: Import the adapted 3D model file directly into the corresponding 3D printer. The printer's built-in parameter matching module can automatically adapt the printing parameters. This step can realize the transformation from 2D to 3D without manual adjustment, and finally realize the stereoscopic imaging of the object.

3. The system of claim 1, wherein, The image acquisition subunit is an industrial-grade high-speed camera, designed to match the high-speed scanning of the MEMS galvanometer, ensuring high-speed image acquisition and real-time 3D reconstruction.

4. The system of claim 1, wherein, The accompanying software can optimize the point cloud data model and connect the MEMS galvanometer-type stripe projection scanning unit and the 3D printing execution unit to ensure the smoothness and continuity of the entire stereoscopic imaging process.

5. The method according to claim 2, characterized in that, The phase recovery method uses a three-frequency 12-step phase shift method. This method calculates the absolute phase information of the object surface by using fringe images with different phase shifts, which can ensure the accuracy of phase calculation during the two-dimensional to three-dimensional conversion process.

6. The method according to claim 2, characterized in that, The system calibration includes camera intrinsic parameter calibration, projector intrinsic parameter calibration, and extrinsic parameter calibration between the camera and the projector, which helps to improve the accuracy of 3D point cloud data and ensure the overall replication accuracy of 2D to 3D photographic stereoscopic forming.

Images