A cross light bar analysis method based on a double-side projection three-dimensional measuring device

Abstract

The application provides a cross light strip analysis algorithm based on a double-side projection three-dimensional measuring device, and relates to the technical field of three-dimensional measurement. The method comprises the following steps: constructing a double-side projection three-dimensional measuring device, wherein the double-side projection three-dimensional measuring device comprises a carrier table, a to-be-measured object arranged on the carrier table, a camera arranged above the to-be-measured object, and two DLP light machines arranged on the two sides of the carrier table respectively; simultaneously projecting a multi-line structured light stripe to the surface of the to-be-measured object by using the two DLP light machines, and acquiring a cross light strip image collected by the camera at this time; controlling the two DLP light machines to simultaneously switch the projection pictures one by one, so that the multi-line structured light stripe projected to the surface of the to-be-measured object is shifted, the corresponding cross light strip image is collected by the camera, and a cross light strip image group is generated; analyzing each cross light strip image in the cross light strip image group, and extracting corresponding light strip data, so as to determine which light plane of which DLP light machine each light strip belongs to. The application is based on the principle of three-dimensional reconstruction of a line structured light, and the measurement accuracy can reach the micron level, so that the high accuracy and analysis accuracy of measurement are ensured.

Description

Technical Field

[0001] This invention relates to the field of three-dimensional measurement technology, and specifically to a method for resolving cross light stripes based on a two-sided projection three-dimensional measurement device. Background Technology

[0002] With technological advancements, the complexity and dimensional accuracy of part surface morphology are constantly increasing. Currently, various methods exist for the 3D inspection of complex, micro-sized parts. Among them, non-contact measurement, primarily based on optical measurement, offers advantages such as high measurement speed and accuracy, and has been increasingly adopted in industry in recent years. However, achieving rapid, high-resolution, and blind-zone-free surface morphology reconstruction remains a challenging problem for objects with complex surface morphology.

[0003] Some scholars have adopted a multi-axis system approach, but this method is costly in terms of both time and hardware, and the correction of multi-axis system errors is also complex, making it unsuitable for the rapid and low-cost measurement needs of industry. Others have proposed using multi-sensor combination measurements, which are more efficient at addressing blind spots in complex geometrically shaped micro-components compared to multi-axis systems. Grating projection is a high-precision, rapid measurement method that achieves blind-spot-free, high-precision measurement of complex geometrically shaped micro-components based on double-sided grating projection. However, some scholars are not satisfied with the alternating projection method using double-sided projectors because of its low efficiency. Simultaneous projection of fringes from both sides has been used, but the phase information obtained from the resolved double-sided projection gratings is currently not very accurate. Line structured light offers high measurement accuracy and performs well when dealing with complex structures and reflective properties. Furthermore, compared to grating projection, information from line structured light is easier to extract and resolve; only the center line of the light stripes needs to be extracted. However, line structured light typically requires a high-precision displacement stage for push-broom measurements, resulting in high cost and low efficiency.

[0004] In summary, the field of 3D measurement currently lacks a 3D measurement device and a cross-strip analysis algorithm that can achieve fast, high-resolution, and blind-zone-free 3D measurement. Summary of the Invention

[0005] This invention aims to at least partially solve one of the technical problems in the prior art. Therefore, the purpose of this invention is to provide a cross-strip analysis method based on a bilateral projection three-dimensional measurement device. Based on the principle of line structured light three-dimensional reconstruction, the measurement accuracy can reach the micrometer level, ensuring high measurement precision and analytical accuracy.

[0006] Technical solution: In a first aspect, the present invention provides a method for resolving intersecting light stripes based on a two-sided projection three-dimensional measurement device, comprising:

[0007] A dual-projection three-dimensional measurement device is constructed, comprising a stage, an object to be measured set on the stage, a camera set above the object to be measured, and two DLP optical engines respectively set on both sides of the stage.

[0008] Two DLP optical engines are used to simultaneously project multi-line structured light stripes onto the surface of the object under test, and the cross-stripes image captured by the camera at this time is obtained.

[0009] The two DLP optical engines are controlled to switch the projected images one by one at the same time, so that the multi-line structured light stripes projected onto the surface of the object under test are shifted, and the camera acquires the corresponding cross light stripe images to generate a cross light stripe image group.

[0010] Each cross-strip image in the cross-strip image group is analyzed to extract the corresponding light stripe data in order to determine which output plane of which DLP optical engine each light stripe belongs to.

[0011] Furthermore, before simultaneously projecting multi-line structured light stripes onto the surface of the object under test using two DLP optical engines and acquiring the cross-stripes image captured by the camera, the method further includes calibrating the multiple light-emitting planes of the two DLP optical engines with the camera respectively.

[0012] Furthermore, the step of calibrating the multiple light-emitting planes of the two DLP optical engines with the camera includes: calculating the plane equation of each light-emitting plane involved in the measurement in the camera coordinate system, as follows:

[0013] Formula (1)

[0014] Where A, B, C, and D are the equation parameters of the light-emitting plane, and (x, y, z) are the coordinates of each point cloud data belonging to the light-emitting plane in the camera coordinate system.

[0015] Furthermore, after analyzing each cross light bar image in the cross light bar image group and extracting the corresponding light bar data to determine which plane of which DLP optical engine each light bar belongs to, the method further includes: performing three-dimensional reconstruction of the surface morphology of the object under test.

[0016] Furthermore, the three-dimensional reconstruction of the surface morphology of the object under test includes:

[0017] Based on the calibration results of the output plane of the DLP optical engine and the camera, the extracted light stripe data is substituted into the plane equation of the corresponding output plane to obtain the three-dimensional point cloud data of the surface of the object under test. The three-dimensional reconstruction of the surface morphology of the object under test is completed based on all the three-dimensional point cloud data.

[0018] Furthermore, the step of parsing each intersecting light bar image in the intersecting light bar image group and extracting the corresponding light bar data includes performing light bar feature enhancement processing on the intersecting light bar images:

[0019] Regions of interest (ROI) are defined based on the grayscale distribution of the intersecting light stripe image;

[0020] Adaptive thresholding is used to perform ROI region image enhancement and image binarization.

[0021] Furthermore, the process of enhancing the light bar feature of the intersecting light bar image further includes detecting light bar features:

[0022] Determine whether the shape of the enhanced light stripe is approximately a straight line or a curve;

[0023] If the enhanced light stripe is approximately straight, the first-order Ransac algorithm is used to detect the approximate outline position of each light stripe in the cross light stripe image. If the enhanced light stripe is approximately curved, the second-order Ransac algorithm is used to detect the approximate outline position of each light stripe in the cross light stripe image.

[0024] Based on the approximate outline position of the detected light stripes, a polynomial fitting is performed on each light stripe to obtain the characteristic equation for each light stripe:

[0025] y=b0+k1x+k2x 2 +k3x 3 +⋯ Formula (2)

[0026] Where (x,y) represents the pixel position of the light stripe, and b0,k1,k2,k3 are the feature parameters of each light stripe.

[0027] Furthermore, the detection of light stripe features is followed by light stripe data recognition and extraction:

[0028] After obtaining the characteristic parameters of each light stripe, the light stripes are classified and sorted using the characteristic parameters to determine which output plane of which DLP optical engine each light stripe belongs to.

[0029] In the original cross-bar image captured by the camera, the approximate outline position of each corresponding light bar is locked;

[0030] Obtain the pixel position and gray level of each light bar in the original intersecting light bar image, and use the gray level centroid method to extract the center sub-pixel of the light bar, thus obtaining the center pixel (u,v) of each light bar.

[0031] Beneficial effects: Based on the ideas of this invention, the problem of blind spots when measuring complex structural objects can be effectively solved, ensuring the integrity of the three-dimensional topography reconstruction of the object surface.

[0032] Based on the algorithm proposed in this invention, it is possible to quickly analyze the intersecting light stripes formed by simultaneous projection from both sides, thereby improving the efficiency of bilateral projection measurement and meeting the speed requirements of modern manufacturing for the measurement field.

[0033] The dual-projection three-dimensional measurement device of the present invention uses DLP optical engine to quickly project and switch projected images, realizing rapid displacement measurement of light stripes. It can achieve high sampling rate measurement without the need for displacement equipment, and has the advantage of low cost.

[0034] This measurement system is based on the principle of line structured light 3D reconstruction, and its measurement accuracy can reach the micrometer level, ensuring high measurement precision and accurate analysis. Attached Figure Description

[0035] Figure 1 A flowchart illustrating a cross-strip analysis method based on a dual-projection three-dimensional measurement device provided by the present invention;

[0036] Figure 2 This is a schematic diagram of the structure of the dual-projection three-dimensional measuring device provided by the present invention;

[0037] Figure 3 This invention provides a schematic diagram of partial light stripe displacement in a projected image of a cross light stripe analysis method based on a dual-projection three-dimensional measurement device.

[0038] Figure 4 A flowchart of the cross-light stripe analysis method based on a dual-projection three-dimensional measurement device provided by the present invention;

[0039] Figure 5 An example diagram of a cross light stripe formed by a cross light stripe analysis method based on a two-sided projection three-dimensional measurement device provided by the present invention.

[0040] In the diagram: 1. Camera; 2. Double telecentric lens; 3. DLP optical engine on the left; 4. DLP optical engine on the right; 5. Multi-line structured light projected on the left; 6. Multi-line structured light projected on the right; 7. Object under test; 8. Stage. Detailed Implementation

[0041] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0042] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0043] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0044] like Figure 1 As shown, the first embodiment of the present invention provides a method for resolving intersecting light stripes based on a two-sided projection three-dimensional measurement device, comprising:

[0045] S1. Construct a dual-sided projection three-dimensional measurement device, which includes a stage 8, a test object 7 set on the stage 8, a camera 1 set above the test object 7, and two DLP optical engines set on both sides of the stage respectively.

[0046] S2. Simultaneously project multi-line structured light stripes onto the surface of the object under test using two DLP optical engines, and acquire the cross-stripes image captured by the camera at this time;

[0047] S3. Control the two DLP optical engines to switch the projected images one by one at the same time, so that the multi-line structured light stripes projected onto the surface of the object to be measured are shifted, and the camera acquires the corresponding cross light stripe images to generate a cross light stripe image group.

[0048] S4. Analyze each cross light bar image in the cross light bar image group, extract the corresponding light bar data, and determine which output plane of which DLP optical engine each light bar belongs to.

[0049] Figure 2 This is a schematic diagram of the structure of the dual-projection three-dimensional measurement device of the present invention. This dual-projection three-dimensional measurement device can be used to measure the three-dimensional morphology of objects with complex surface undulations, including but not limited to... Figure 2 The form of the object being measured.

[0050] Construction of a Dual-Side Projection 3D Measurement Device: Based on the requirements of field of view and accuracy, select appropriate camera 1, dual telecentric lenses 2, left DLP optical engine 3, and right DLP optical engine 4. The left and right DLP optical engines 3 and 4 respectively project multi-line structured light 5 and 6 projected from the left and right sides, respectively. The measurement device adopts a single-camera centered, dual-side projection setup. Based on the working distance of the lenses and DLP optical engines, depth of field, and other practical considerations, the relative positions of each component are rationally designed. The two DLP optical engines are installed facing each other at a 150-degree angle. While meeting measurement requirements, the system is kept as compact as possible. Simultaneously, the overall system needs integrated design to achieve system integration. The entire system is controlled by a host computer, which controls the left and right DLP optical engines 3 and 4 to simultaneously project multi-line structured light stripes and simultaneously switch the projected images to shift the light stripes, such as... Figure 3 As shown, the light stripes shift by one pixel each time. Line I represents the pixel position of a group of multi-line structured light stripes in the projected image before shifting; line II represents the pixel position of a group of multi-line structured light stripes in the projected image after shifting. Camera 1 acquires images of the intersecting light stripes deformed by the surface modulation of the object under test 7 in real time. The acquired images are finally processed and analyzed by the host computer. After the intersecting light stripes are resolved, the three-dimensional reconstruction of the object surface is achieved.

[0051] System Calibration: Due to the insensitivity of the dual telecentric lens 2 to depth changes along the optical axis (z-direction), this calibration does not consider the transformation relationship along the z-axis, simplifying the imaging model and using the simplified dual telecentric imaging model to obtain the camera intrinsic parameters. A multi-light plane calibration method assisted by a displacement stage and a fast prediction calibration method for multi-light planes based on DLP co-rotation axis are used to calibrate the multi-light planes involved in the measurement process. During multi-light plane calibration, a z-axis displacement stage is used (the z-axis displacement stage can be installed above the stage 8). The z-axis displacement stage provides z-axis information that the lens cannot perceive, thus obtaining the complete coordinate information (x, y, z) of each point on the light plane in the camera coordinate system. The least squares plane fitting is then used to obtain the light plane equation, completing the multi-light plane calibration. The plane equation in the camera coordinate system for each light plane involved in the measurement is obtained, where A, B, C, and D are the parameters of the light plane equation, and (x, y, z) are the coordinates of each point cloud data belonging to that light plane in the camera coordinate system.

[0052] Formula (1)

[0053] Cross-stripe image acquisition: The object to be measured 7 is placed within the measurable area on the stage 8. Two DLP optical engines simultaneously project multi-line structured light stripes onto the surface of the object, forming cross-stripes. Modulated by the height of the object, these cross-stripes create distorted cross-stripe images under the camera. Subsequently, as the two DLPs simultaneously switch projected images, the light stripes shift, forming new cross-stripe images under the camera. Therefore, by controlling the camera with software to acquire images one by one, a series of cross-stripe images are obtained.

[0054] Cross-light bar analysis: Since multiple images of cross-light bars are acquired, it is necessary to analyze the cross-light bars in each image, that is, to determine which light plane of which projector each light bar belongs to. This invention proposes a cross-light bar analysis algorithm based on cross-light bar feature information, solving the problem of difficult cross-light bar analysis.

[0055] 3D reconstruction of the object under test: Based on the system calibration results, the extracted light stripe data is substituted into the corresponding light plane equation to obtain 3D point cloud data. Finally, the 3D reconstruction of the surface morphology of the object under test is completed based on all the point cloud data, and subsequent evaluation is carried out according to actual needs.

[0056] The proposed cross-strip resolving algorithm, employing a measurement method that simultaneously projects multi-line structured light from both sides, can solve the common blind zone problem in 3D measurement. The projection device rapidly switches between projected images to shift the multi-line structured light, improving the high sampling rate of the measurement system without the need for displacement equipment. The cross-strip resolving algorithm ensures the measurement efficiency of bilateral projection, overcoming the difficulty of separating and resolving the structured light on both sides in traditional bilateral projection measurements. Ultimately, the measurement system achieves rapid, high-sampling-rate, blind-zone-free 3D measurement of complex, minute parts.

[0057] Furthermore, before simultaneously projecting multi-line structured light stripes onto the surface of the object under test 7 using two DLP optical engines to obtain the cross-stripes image captured by the camera, the method further includes calibrating the multiple light-emitting planes of the two DLP optical engines with the camera respectively.

[0058] Furthermore, the step of calibrating the multiple light-emitting planes of the two DLP optical engines with the camera includes: calculating the plane equation of each light-emitting plane involved in the measurement in the camera coordinate system, as follows:

[0059] Formula (1)

[0060] Where A, B, C, and D are the equation parameters of the light-emitting plane, and (x, y, z) are the coordinates of each point cloud data belonging to the light-emitting plane in the camera coordinate system.

[0061] Furthermore, after analyzing each cross light bar image in the cross light bar image group and extracting the corresponding light bar data to determine which plane of which DLP optical engine each light bar belongs to, the method further includes: performing three-dimensional reconstruction of the surface morphology of the object under test.

[0062] Furthermore, the three-dimensional reconstruction of the surface morphology of the object under test includes:

[0063] Based on the calibration results of the output plane of the DLP optical engine and the camera, the extracted light stripe data is substituted into the plane equation of the corresponding output plane to obtain the three-dimensional point cloud data of the surface of the object under test. The three-dimensional reconstruction of the surface morphology of the object under test is completed based on all the three-dimensional point cloud data.

[0064] To meet the requirements of measurement speed and blind zone resolution, this invention combines line structured light with grating projection measurement techniques to build a dual-sided projection 3D measurement device. Based on the line structured light measurement principle, the device projects multi-line structured light stripes from two projectors. Mimicking the phase-change characteristics of grating projection, the two projectors switch projected images to shift the multi-line structured light stripes, ensuring that the light stripe at every pixel position of the projector is projected and participates in the measurement, thus achieving high-resolution measurement of the measured area. To ensure measurement efficiency in addressing blind zones, the two projectors simultaneously project multi-line structured light, which is projected onto the surface of the measured object at a certain angle. Reconstruction requires first resolving the intersecting light stripes, i.e., determining which projection plane each light stripe belongs to. However, current multi-line structure recognition algorithms mainly focus on parallel projected multi-line structures. The proposed method can identify dense stripes, but research on resolving intersecting light stripes is limited. Furthermore, during measurement, the light stripes may deform or misalign due to surface modulation, further complicating the differentiation and identification of intersecting light stripes. Therefore, this invention designs a dedicated cross-strip parsing algorithm to achieve fast parsing of cross-stripes.

[0065] Specifically, the step of parsing each intersecting light bar image in the intersecting light bar image group and extracting the corresponding light bar data includes performing light bar feature enhancement processing on the intersecting light bar images:

[0066] Regions of interest (ROI) are defined based on the grayscale distribution of the intersecting light stripe image;

[0067] Adaptive thresholding is used to perform ROI region image enhancement and image binarization.

[0068] Furthermore, the process of enhancing the light bar feature of the intersecting light bar image further includes detecting light bar features:

[0069] Determine whether the shape of the enhanced light stripe is approximately a straight line or a curve;

[0070] If the enhanced light stripe is approximately straight, the first-order Ransac algorithm is used to detect the approximate outline position of each light stripe in the cross light stripe image. If the enhanced light stripe is approximately curved, the second-order Ransac algorithm is used to detect the approximate outline position of each light stripe in the cross light stripe image.

[0071] Based on the approximate outline position of the detected light stripes, a polynomial fitting is performed on each light stripe to obtain the characteristic equation for each light stripe:

[0072] y=b0+k1x+k2x 2 +k3x 3 +⋯ Formula (2)

[0073] Where (x,y) represents the pixel position of the light stripe, and b0,k1,k2,k3 are the feature parameters of each light stripe.

[0074] Furthermore, the detection of light stripe features is followed by light stripe data recognition and extraction:

[0075] After obtaining the characteristic parameters of each light stripe, the light stripes are classified and sorted using the characteristic parameters to determine which output plane of which DLP optical engine each light stripe belongs to.

[0076] In the original cross-bar image captured by the camera, the approximate outline position of each corresponding light bar is locked;

[0077] Obtain the pixel position and gray level of each light bar in the original intersecting light bar image, and use the gray level centroid method to extract the center sub-pixel of the light bar, thus obtaining the center pixel (u,v) of each light bar.

[0078] The cross-striping analysis algorithm is the core of this invention, and the specific process is as follows: Figure 4 As shown, the specific process and method are as follows:

[0079] Input an image of the intersecting light stripes. The deformation of the intersecting light stripes varies depending on the object being measured, such as... Figure 5 As shown, the present invention divides the measurement situation into three categories: measuring planar objects, measuring curved objects, and measuring complex surface objects. The cross light stripes corresponding to these three situations are respectively straight cross light stripes, curved cross light stripes, and irregular line cross light stripes.

[0080] Enhancing light bar features. Images with intersecting light bars typically have complex grayscale distributions, requiring light bar enhancement processing. First, the Region of Interest (ROI) is divided based on the image's grayscale distribution. Then, adaptive thresholding is used to complete image enhancement and binarization. Because the shape characteristics of the light bars vary, different structuring elements are selected to extend and thicken them. For example, straight light bars can be enhanced using linear structuring elements, while irregular light bars can be enhanced using cross-shaped or disk-shaped structuring elements. Light bar enhancement processing ensures the continuity of the light bars, avoiding discontinuities that could affect the extraction of subsequent light bars, and the thickened light bar features are more prominent.

[0081] Detecting light stripe features. Determine whether the shape of the enhanced light stripe is approximately a straight line or a curve. Use the corresponding first-order or second-order Ransac algorithm to detect the approximate position of each light stripe in the image and display the detection effect in the image. Based on the detected light stripe, polynomial fitting can be performed on each light stripe to obtain the characteristic equation of each light stripe. The pixel position (u,v) of the light stripe at this time can be regarded as (x,y), then we have formula (2).

[0082] y=b0+k1x+k2x 2 +k3x 3 +⋯ Formula (2)

[0083] Light stripe recognition and extraction. After obtaining the feature parameters b0, k1, k2, k3 for each light stripe, the light stripes are classified and sorted using these parameters. This allows us to determine which optical plane of which DLP optical engine each light stripe belongs to. Next, the location region of each detected light stripe is located in the original image. The pixel position and grayscale of each light stripe in the original image are obtained. The grayscale centroid method is used to extract the sub-pixel center of each light stripe, thus obtaining the center pixel (u, v) of each light stripe.

[0084] In summary, this invention designs and constructs a dual-projection 3D measurement device capable of performing 3D measurements on objects with complex surface undulations. It also proposes an algorithm for light stripe analysis based on intersecting light stripe feature information. Ultimately, this achieves rapid, blind-zone-free, and high-resolution measurement of the object under test.

[0085] Based on the concept of this invention, the problem of blind spots when measuring complex structural objects can be effectively solved, ensuring the integrity of the three-dimensional topography reconstruction of the object's surface.

[0086] Based on the algorithm proposed in this invention, it is possible to quickly analyze the intersecting light stripes formed by simultaneous projection from both sides, thereby improving the efficiency of bilateral projection measurement and meeting the speed requirements of modern manufacturing for the measurement field.

[0087] The dual-projection three-dimensional measurement device of the present invention uses DLP optical engine to quickly project and switch projected images, realizing rapid displacement measurement of light stripes. It can achieve high sampling rate measurement without the need for displacement equipment, and has the advantage of low cost.

[0088] This measurement system is based on the principle of line structured light 3D reconstruction, and its measurement accuracy can reach the micrometer level, ensuring high measurement precision and accurate analysis.

[0089] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media containing computer-usable program code. These computer program instructions can also be stored in a computer-readable storage medium capable of directing a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means implemented in a process. Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0090] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

Claims

1. A cross light bar analysis method based on a two-side projection three-dimensional measuring device, characterized in that, The application relates to a bilateral projection three-dimensional measuring device, which adopts DLP light machines to quickly project and switch projection pictures, and comprises a material loading platform, a measured object arranged on the material loading platform, a camera arranged above the measured object and two DLP light machines arranged on the two sides of the material loading platform. Two DLP light machines are used to simultaneously project multi-line structured light stripes onto the surface of the measured object, and cross light stripe images collected by the camera at the moment are acquired. The two DLP light machines are controlled to simultaneously switch projection pictures one by one, so that the multi-line structured light stripes projected onto the surface of the measured object are shifted, the corresponding cross light stripe images are collected by the camera, and a cross light stripe image group is generated. Each cross light stripe image in the cross light stripe image group is analyzed, and corresponding light stripe data is extracted, so that it is determined which light plane of which DLP light machine each light stripe belongs to. The analysis of each cross light stripe image in the cross light stripe image group and the extraction of corresponding light stripe data comprises light stripe feature enhancement processing of the cross light stripe image. An ROI region is divided according to the gray distribution of the cross light stripe image. An adaptive threshold is used to complete ROI region image enhancement and image binarization. After the light stripe feature enhancement processing of the cross light stripe image, light stripe feature detection is further included. It is judged whether the shape of the enhanced light stripe is a straight line or a curve. If the shape of the enhanced light stripe is a straight line, a first-order Ransac algorithm is used to detect the contour position of each light stripe on the cross light stripe image, and if the shape of the enhanced light stripe is a curve, a second-order Ransac algorithm is used to detect the contour position of each light stripe on the cross light stripe image. According to the detected light stripe contour position, each light stripe is polynomially fitted, and the characteristic equation of each light stripe is acquired. Where (x, y) is the pixel position of the light stripe, and b0, k1, k2 and k3 are the characteristic parameters of each light stripe. y=b0+k1x+k2x 2 +k3x 3 +⋯ Formula (2) After the light stripe feature detection, light stripe data recognition and extraction are further included. After the characteristic parameters of each light stripe are obtained, the characteristic parameters are used to complete light stripe classification and sorting, so that it is determined which light plane of which DLP light machine each light stripe belongs to. The corresponding contour position of each light stripe is locked in the original cross light stripe image collected by the camera. The pixel position and gray scale of each light stripe in the original cross light stripe image are acquired, the gray scale barycenter method is used to realize light stripe center sub-pixel extraction, and the center pixel (u, v) of each light stripe is obtained. Before the two DLP light machines are used to simultaneously project multi-line structured light stripes onto the surface of the measured object, the multiple light planes of the two DLP light machines are calibrated with the camera.

2. The cross light bar analysis method based on a dual-side projection three-dimensional measuring device according to claim 1, characterized in that, The calibration of the multiple light planes of the two DLP light machines with the camera comprises calculation of the plane equation of each light plane participating in measurement in the camera coordinate system, as follows:

3. The cross fringes resolving method based on a dual-side projection three-dimensional measuring device according to claim 2, wherein, Where A, B, C and D are equation parameters of the light plane, and (x, y, z) is the coordinate of each point cloud data belonging to the light plane in the camera coordinate system. Equation (1) ​ 4. The cross fringes resolving method based on a dual-side projection three-dimensional measuring device according to claim 3, wherein, The analyzing each cross light bar image in the cross light bar image group and extracting corresponding light bar data to determine which plane of which DLP light engine each light bar belongs to further comprises: performing three-dimensional reconstruction of the surface topography of the object to be measured.

5. The cross fringes resolving method based on a dual-side projection three-dimensional measuring device according to claim 4, wherein, The three-dimensional reconstruction of the surface topography of the object to be measured comprises: Based on the calibration results of the light emitting planes of the DLP light engine and the camera, the extracted light bar data is substituted into the plane equation of the corresponding light emitting plane to obtain three-dimensional point cloud data of the measured object surface, and the three-dimensional reconstruction of the surface topography of the measured object is completed according to all the three-dimensional point cloud data.

Images