A phase solving method based on sequence pixel extreme value positioning

Abstract

The application discloses a phase solving method based on sequence pixel extreme value positioning, relates to the fields of computer vision and optical measurement, and comprises the following steps: designing M line shift or phase shift fringe patterns and N complementary gray code patterns, projecting the patterns to an object to be measured in sequence, collecting images of the object to be measured under different fringe patterns, and obtaining M line shift fringe images or phase shift fringe images and N complementary gray code images; performing phase calculation on the collected M line shift fringe images or phase shift fringe images, and obtaining a wrapped phase image; decoding the collected first N-1 gray codes to obtain K1 orders, and decoding all the N gray codes to obtain K2 orders; combining the wrapped phase image and the multiple gray code projection images to perform phase unwrapping, and obtaining a phase unwrapping image; and through the innovative extreme value positioning and adaptive exposure processing mechanism, the accuracy, the robustness and the reliability of the phase solving are effectively improved, and the problem that the measurement effect of the prior art is poor in a complex environment is solved.

Description

Technical Field

[0001] This invention relates to the fields of computer vision and optical measurement technology, and in particular to a phase solution method based on the localization of extrema of sequential pixels. Background Technology

[0002] Structured light 3D measurement systems are an important non-contact optical measurement technology. They project coded patterns onto the surface of an object, capture deformation fringes using a camera, and then reconstruct the object's 3D shape through decoding and phase calculation. Phase calculation, the core of this technology, typically includes two steps: phase wrapping and phase unwrapping. The accuracy and robustness of the phase wrapping directly determine the final reconstruction result. Existing technologies mainly employ Gray code phase-shifting and Gray code line-shifting methods: The former uses sinusoidal phase-shifting codes to solve the phase wrapping, designs an original sine wave pattern according to a specific phase shift method, generates several phase-shifted fringe code patterns, and uses these patterns to project and acquire a sequence of phase-shifted fringe image sequences. The phase wrapping phase is then deciphered by processing the image sequences. The latter uses binarized line-shifting codes to solve the phase wrapping, designs an original fringe pattern according to a specific line-shift method, generates several line-shifted fringe code patterns, and uses these patterns to project and acquire a sequence of line-shifted fringe image sequences. The phase wrapping phase is then deciphered by processing the image sequences. Gray code, as a spatial coding scheme, provides order information for phase expansion by uniquely identifying pixel positions, and has advantages such as strong anti-interference and low bit error rate.

[0003] However, in practical applications, due to factors such as uneven reflection of objects, ambient light interference, noise, and lens defocus, especially on highly reflective surfaces or in complex environments, the fringe structure is difficult to distinguish, and the quality of the fringe image is easily degraded, leading to a decrease in the accuracy of phase solving using traditional methods.

[0004] Therefore, a phase calculation method based on the localization of extrema of sequential pixels is provided to solve the above problems. Summary of the Invention

[0005] To address the aforementioned challenges, this invention provides a phase calculation method based on sequence pixel extremum localization. Building upon the traditional Gray code phase shift method, a complementary Gray code pattern is added to eliminate the periodic blurring of the stripe pattern. Through interpolation, fitting, and iteration, the precise maxima, minima, and their corresponding stripe offsets of the discrete sequence are calculated. Then, based on overexposure and underexposure conditions, the final accurate phase is calculated. This method is crucial for improving the stability and accuracy of structured light 3D measurement systems in complex scenarios.

[0006] To achieve the above objectives, this invention provides a phase calculation method based on the localization of extrema of sequential pixels, comprising the following steps:

[0007] S1: Design M line-shifted stripe coding patterns or phase-shifted stripe patterns and N complementary Gray code patterns, and project the patterns onto the object to be tested in sequence. At the same time, collect images of the object to be tested under different stripe patterns to obtain M line-shifted stripe images or phase-shifted stripe images and N complementary Gray code images.

[0008] S2: Perform phase calculation on the acquired M-amplitude line-shifted fringe images or phase-shifted fringe images to obtain the wrapped phase map;

[0009] S3: Decode the first N-1 Gray codes to obtain the K1 level, and decode all N Gray codes to obtain the K2 level;

[0010] S4: Combine the wrapped phase map obtained in S2 with the multiple Gray code projection maps obtained in S3 to perform phase unrolling, and obtain the phase unrolled map.

[0011] Preferably, the design of the M-width line-shifting stripe coding pattern in S1 specifically includes:

[0012] Design the first original line-shifted fringe pattern; wherein the interval between two adjacent bright fringe patterns in the original fringe pattern is M / 2 pixels, and a bright fringe occupies M / 2 pixels in the column direction, where M is an even number not less than 4; with a step size of 1 pixel, the original line-shifted fringe pattern is moved sequentially in the same direction to generate M line-shifted fringe patterns; the expression for the j-th line-shifted fringe encoding pattern is:

[0013] ;

[0014] Where j = 0, 1, ..., M-1; x and y are the x and y coordinates of the pixel, respectively, with the starting point of the coordinates being... M represents the number of steps the line is moved, and % represents the modulo operation.

[0015] Preferably, the design of the M-amplitude phase-shifting stripe coding pattern in S1 specifically includes:

[0016] When the number of phase shift steps is M, the expression for the j-th phase shift fringe coding pattern is:

[0017] ;

[0018] Where j = 0, 1, ..., M-1; x and y are the x and y coordinates of the pixel, respectively, with the starting point of the coordinates being... M is the number of phase shift steps, A is the background light intensity, and B is the modulation intensity. For phase value, The phase shift value, This is the phase shift frequency.

[0019] Preferably, S2 specifically includes:

[0020] S21: For each pixel, extract its gray value sequence in M ​​images and identify the maximum and minimum points in the sequence;

[0021] S22: Determine the exposure status of the pixel based on the gray values ​​of the maximum and minimum points;

[0022] S23: Based on different exposure states, select the corresponding extreme points and their neighboring pixels, and use a single-peak one-dimensional Gaussian model to calculate the sub-pixel level extreme positions.

[0023] S24: Calculate the wrapping phase value of a pixel based on its sub-pixel-level extreme position.

[0024] Preferably, S21 specifically includes:

[0025] The acquired M-image line-shifted fringe or phase-shifted fringe images are sorted according to their encoding order, with j as the abscissa. or Using the vertical axis as the ordinate, plot the pixel values ​​of M images as a discrete dot plot. Compare the M pixel values ​​and find the maximum value. or Corresponding minimum value or Corresponding .

[0026] Preferably, the method for determining the exposure status of pixels includes:

[0027] when or When the value is 255, it is considered overexposed.

[0028] when or When the value is less than or equal to 10, it is considered underexposed.

[0029] when or Less than 255 or A value greater than 10 indicates normal exposure.

[0030] The preferred one-dimensional Gaussian model is expressed as:

[0031] ;

[0032] in, Peak height This is the peak position. The standard deviation is denoted as .

[0033] Preferably, the wrapping phase value of a pixel is represented as:

[0034] ;

[0035] in, , Subpixel level obtained by single-peak one-dimensional Gaussian fitting The minimum and maximum values ​​of .

[0036] Preferably, the phase expansion in S4 is represented as:

[0037] ;

[0038] in, , It is a level.

[0039] Therefore, this invention employs a phase-solving method based on sequence pixel extreme value localization, primarily applicable to scenarios such as industrial inspection, robot vision, and cultural heritage digitization. It achieves sub-pixel-level extreme value localization through Gaussian fitting, overcoming computational errors caused by discrete sampling and significantly improving the accuracy of the wrapped phase solution. The introduction of complementary Gray codes provides redundant order information, enabling more accurate determination of the true phase order during phase unfolding, reducing the risk of transition errors, and enhancing the reliability and continuity of the entire 3D reconstruction result. Through innovative extreme value localization and adaptive exposure processing mechanisms, it effectively improves the accuracy, robustness, and reliability of the phase solution, addressing the pain point of poor measurement performance in complex environments in existing technologies.

[0040] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0041] Figure 1 This is a flowchart illustrating a phase-solving method based on the localization of extreme values ​​of sequential pixels in this invention.

[0042] Figure 2 This is the first line-shifted stripe coding pattern (i.e., the original stripe pattern) designed in this embodiment of the invention;

[0043] Figure 3 This is a set of line-shifted fringe patterns generated by shifting the original fringe pattern in an embodiment of the present invention;

[0044] Figure 4 This is a set of line-shifted fringe patterns generated by the phase-shifting method in an embodiment of the present invention;

[0045] Figure 5 This is a coding diagram of complementary Gray codes in an embodiment of the present invention;

[0046] Figure 6 This is a flowchart of the unpacking phase in an embodiment of the present invention;

[0047] Figure 7 This is a discrete dot plot of pixel values ​​in overexposed states in this embodiment of the invention;

[0048] Figure 8 This is a discrete dot plot of pixel values ​​under normal exposure conditions in an embodiment of the present invention;

[0049] Figure 9 This is a discrete dot plot of pixel values ​​in an underexposed state in this embodiment of the invention;

[0050] Figure 10 This is a curve obtained by Gaussian fitting under the overexposure state in an embodiment of the present invention;

[0051] Figure 11 This is a curve obtained by Gaussian fitting under normal exposure conditions in an embodiment of the present invention;

[0052] Figure 12 This is a curve obtained by Gaussian fitting for the underexposure state in an embodiment of the present invention. Detailed Implementation

[0053] The following detailed description of embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0054] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0055] The terms "comprising" or "including" as used in this invention mean that the element preceding the term encompasses the element listed after the term, and do not exclude the possibility of encompassing other elements as well. Terms such as "inner," "outer," "upper," and "lower" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. When the absolute position of the described object changes, the relative positional relationship may also change accordingly. In this invention, unless otherwise explicitly specified and limited, the term "attached" and similar terms should be interpreted broadly. For example, it can refer to a fixed connection, a detachable connection, or an integral part; it can refer to a direct connection or an indirect connection through an intermediate medium; it can refer to the internal communication of two elements or the interaction relationship between two elements. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0056] Example

[0057] A phase calculation method based on the localization of extrema of sequential pixels, such as... Figure 1 As shown, it includes the following steps:

[0058] S1: Design M line-shifted stripe coding patterns or phase-shifted stripe patterns and N complementary Gray code patterns, and project the patterns onto the object to be tested in sequence. At the same time, collect images of the object to be tested under different stripe patterns to obtain M line-shifted stripe images or phase-shifted stripe images and N complementary Gray code images.

[0059] In a specific embodiment, the projection resolution is 1280. 720 pixels, taking an 8-step line shift as an example, that is, M=8, the image is divided into 160 periods, each period contains 8 pixels.

[0060] The design of the M-width line shift stripe coding pattern in S1 specifically includes:

[0061] Design the first original line-shifted fringe pattern; wherein the interval between two adjacent bright fringe patterns in the original fringe pattern is M / 2 pixels, and a bright fringe occupies M / 2 pixels in the column direction, where M is an even number not less than 4; with a step size of 1 pixel, the original line-shifted fringe pattern is moved sequentially in the same direction to generate M line-shifted fringe patterns; the expression for the j-th line-shifted fringe encoding pattern is:

[0062] ;

[0063] Where j = 0, 1, ..., M-1; x and y are the x and y coordinates of the pixel, respectively, with the starting point of the coordinates being... M is the number of steps the line moves, M is an even number not less than 4, and % is the modulo operation.

[0064] Specifically, the line spacing of each coded pattern is 4 pixels (with a period of 8 pixels). The first coded pattern designed (i.e., the original stripe pattern) is as follows: Figure 2 As shown, in the diagram, the first four columns of pixels in each cycle are set to 1, and each bright stripe occupies 4 pixels in the column direction. Subsequent coded patterns are shifted one pixel to the right based on the previous coded pattern. In this embodiment, taking the synthesis of M=8 step line-shifted stripe patterns as an example, using the above formula, a total of M line-shifted stripe patterns are obtained, as shown... Figure 3 The image shows a set of line shifting stripe patterns generated.

[0065] The design of the M-amplitude phase-shifting stripe coding pattern in S1 specifically includes:

[0066] When the number of phase shift steps is M, the expression for the j-th phase shift fringe coding pattern is:

[0067] ;

[0068] Where j = 0, 1, ..., M-1; x and y are the x and y coordinates of the pixel, respectively, with the starting point of the coordinates being... M is the number of phase shift steps, A is the background light intensity, and B is the modulation intensity. For pixels phase value, The pixels of the j-th phase-shift fringe coding pattern The phase shift value, This is the phase shift frequency.

[0069] Specifically, taking the synthesis of M=8 step phase-shifted fringe patterns as an example, using the above formula, a total of M line-shifted fringe patterns are obtained, such as... Figure 4 The image shows a set of generated phase-shifted fringe patterns.

[0070] Using the complementary Gray code encoding scheme, we obtain the following: Figure 5 The N (N=9) Gray code projection diagrams are shown.

[0071] S2: Perform phase calculation on the acquired M-amplitude line-shifted fringe images or phase-shifted fringe images to obtain the wrapped phase map;

[0072] Phase calculation is performed on the M line-shifted or phase-shifted fringe patterns to obtain a folded phase map. In this embodiment, the goal of phase calculation is to extract the phase information of the surface of the object under test from the acquired M line-shifted or phase-shifted fringe patterns. This phase information reflects the phase change caused by the reflection of light on the object surface and is directly related to the shape of the object surface. Phase calculation mainly uses multiple fringe patterns with different offsets to calculate the phase value of each pixel. These phase values ​​form a phase map, where the phase value of each pixel represents the height or shape change of that point relative to a reference plane.

[0073] like Figure 6 As shown, the specific process of S2 includes:

[0074] With a projection resolution of 1280 720 pixels, eight-step line shift or phase shift (M=8), the image is divided into 160 periods, each containing 8 pixels. In this step, phase calculation is performed using the 8-step line shift or phase shift calculation formula to obtain a folded phase map. The same calculation process is performed for each pixel as follows:

[0075] S21: For each pixel, extract its gray value sequence in M ​​images and identify the maximum and minimum points in the sequence;

[0076] Specifically, the acquired M line-shifted fringe images or phase-shifted fringe images are arranged in encoding order, with j as the horizontal axis. or Using the vertical axis as the ordinate, plot the pixel values ​​of M images as a discrete dot plot. Compare the M pixel values ​​and find the maximum value. or Corresponding minimum value or Corresponding The value of j (j = 0, 1, ..., M-1). In practical applications, it can be plotted as follows: Figures 7-9 The diagram shows a discrete point plot.

[0077] S22: Determine the exposure status of the pixel based on the gray values ​​of the maximum and minimum points;

[0078] In practice, an image may be in three states: overexposed, normal, and underexposed. It is stipulated that upper and lower thresholds are set: when the maximum pixel value is 255, it is in an overexposed state; when the minimum pixel value is less than 10, it is in an underexposed state; and the rest are in a normal exposed state.

[0079] S23: Based on different exposure states, select the corresponding extreme points and their neighboring pixels, and use a single-peak one-dimensional Gaussian model to calculate the sub-pixel level extreme positions.

[0080] when or When the value is 225, it indicates an overexposure. At this point, only [the value is applied to] [the area under the overexposed state]. Process and search The two points on the left and right, for the three points mentioned above ( , ), ( , ), ( , Fitting with a unimodal one-dimensional Gaussian model, such as Figure 10 Find the fitted .

[0081] when or Less than 255, or A value greater than 10 indicates normal exposure. Process and search The two points on the left and right, for the three points mentioned above ( , ), ( , ), ( , Use a unimodal one-dimensional Gaussian model for fitting, and find the fitted result. ;right Process and search The two points on the left and right, for the three points mentioned above ( , ), ( , ), ( , Fitting with a unimodal one-dimensional Gaussian model, such as Figure 11 Find the fitted .

[0082] when or When the value is less than or equal to 10, it indicates an underexposure state. At this point, only [the value is applied to]... Process and search The two points on the left and right, for the three points mentioned above ( , ), ( , ), ( , Fitting with a unimodal one-dimensional Gaussian model, such as Figure 12 Find the fitted .

[0083] The fitted model is as follows:

[0084] ;

[0085] in, Peak height This is the peak position. The standard deviation is denoted as .

[0086] The solution process is as follows:

[0087] remember (a = 1,2,3), then:

[0088] ;

[0089] For pairwise differences, eliminate :

[0090] ;

[0091] Expand and organize:

[0092] ;

[0093] make , ,

[0094] This yields a linear relationship:

[0095] ;

[0096] By selecting two sets of differences, we obtain a system of linear equations (the unknowns are...). and ):

[0097] ;

[0098] remember:

[0099] ;

[0100] like ,but:

[0101] ;

[0102] Explicitly write out the solution (using row and column numbers):

[0103] make:

[0104] ;

[0105] but:

[0106] ;

[0107] ;

[0108] Calculated , This refers to the peak position obtained after fitting.

[0109] For the ( , ), ( , ), ( , The result obtained after Gaussian fitting of the three points ,right( , ), ( , ), ( , The result obtained after Gaussian fitting of the three points .

[0110] S24: Calculate the wrapping phase value of a pixel based on its sub-pixel-level extreme position.

[0111] The wrapped phase is obtained by using the following formula for phase calculation:

[0112] ;

[0113] in, , Subpixel level obtained by single-peak one-dimensional Gaussian fitting The minimum and maximum values ​​of .

[0114] S3: Decode the first N-1 Gray codes to obtain the K1 level, and decode all N Gray codes to obtain the K2 level;

[0115] Specifically, the first eight Gray codes collected can be decoded using traditional methods to obtain the order. All nine Gray codes can be decoded to obtain the order. .

[0116] S4: Combine the wrapped phase map obtained in S2 with the multiple Gray code projection maps obtained in S3 to perform phase unrolling, and obtain the phase unrolled map.

[0117] In this embodiment, the main purpose of phase unrolling is to eliminate the discontinuities caused by phase periodicity in the folded phase image, thereby restoring a true and continuous phase distribution. During phase measurement, the phase value has periodic characteristics (usually an integer multiple of M), so sudden phase jumps often appear in the folded phase image. If not handled properly, this will affect the accuracy of subsequent 3D reconstruction.

[0118] The folded phase map is combined with the multiple Gray code projection maps to perform phase unfolding, resulting in a phase unfolded map.

[0119] The phase expansion method for this point is as follows:

[0120] ;

[0121] in, , It is a level.

[0122] Therefore, this invention employs the aforementioned phase-finding method based on sequence pixel extremum localization, which falls under the category of time-coded structured light. Under the premise of keeping the object stationary, when a projector projects a series of equidistant offset structured light stripes in a specific sequence, the grayscale value of any pixel on the object's surface will change over time (i.e., with the sequence of the stripe pattern), forming a regularly changing sequence. This sequence is essentially a discrete sampling of grayscale values ​​changing with stripe offset. The precise maxima, minima, and their corresponding stripe offsets of the discrete sequence are calculated using interpolation, fitting, and iteration methods. Finally, based on overexposure and underexposure conditions, the final accurate phase is calculated.

[0123] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A phase calculation method based on the localization of extrema of sequential pixels, characterized in that, Includes the following steps: S1: Design M line-shifted stripe coding patterns or phase-shifted stripe patterns and N complementary Gray code patterns, and project the patterns onto the object to be tested in sequence. At the same time, collect images of the object to be tested under different stripe patterns to obtain M line-shifted stripe images or phase-shifted stripe images and N complementary Gray code images. S2: Perform phase calculation on the acquired M-amplitude line-shifted fringe images or phase-shifted fringe images to obtain the wrapped phase map; S2 specifically includes: S21: For each pixel, extract its gray value sequence in M ​​images and identify the maximum and minimum points in the sequence; S22: Determine the exposure status of the pixel based on the gray values ​​of the maximum and minimum points; S23: Based on different exposure states, select the corresponding extreme points and their neighboring pixels, and use a single-peak one-dimensional Gaussian model to calculate the sub-pixel level extreme positions. S24: Calculate the wrapping phase value of a pixel based on its sub-pixel level extreme position; S3: Decode the first N-1 Gray codes to obtain the K1 level, and decode all N Gray codes to obtain the K2 level; S4: Combine the wrapped phase map obtained in S2 with the multiple Gray code projection maps obtained in S3 to perform phase unwrapping and obtain the phase unwrapped map; The design of the M-width line shift stripe coding pattern in S1 specifically includes: Design the first original line-shifted fringe pattern; wherein the interval between two adjacent bright fringe patterns in the original fringe pattern is M / 2 pixels, and a bright fringe occupies M / 2 pixels in the column direction, where M is an even number not less than 4; with a step size of 1 pixel, the original line-shifted fringe pattern is moved sequentially in the same direction to generate M line-shifted fringe patterns; the expression for the j-th line-shifted fringe encoding pattern is: ； Where j = 0, 1, ..., M-1; x and y are the x and y coordinates of the pixel, respectively, with the starting point of the coordinates being... M represents the number of steps the line is moved, and % represents the modulo operation; The design of the M-amplitude phase-shifting stripe coding pattern in S1 specifically includes: When the number of phase shift steps is M, the expression for the j-th phase shift fringe coding pattern is: ； Where j = 0, 1, ..., M-1; x and y are the x and y coordinates of the pixel, respectively, with the starting point of the coordinates being... M is the number of phase shift steps, A is the background light intensity, and B is the modulation intensity. For pixels phase value, The pixels of the j-th phase-shift fringe coding pattern The phase shift value, The phase shift frequency; S21 specifically includes: The acquired M-image line-shifted fringe or phase-shifted fringe images are sorted according to their encoding order, with j as the abscissa. or Using the vertical axis as the ordinate, plot the pixel values ​​of M images as a discrete dot plot. Compare the M pixel values ​​and find the maximum value. or Corresponding minimum value or Corresponding .

2. The phase calculation method based on sequence pixel extreme value localization as described in claim 1, characterized in that, Methods for determining the exposure status of pixels include: when When the value is 255, it is considered overexposed. when At that time, it was in an underexposed state; when Less than 255, This is a normal exposure state.

3. The phase calculation method based on sequence pixel extreme value localization as described in claim 2, characterized in that, The one-dimensional Gaussian model with a single peak is represented as: ； in, Peak height This is the peak position. The standard deviation is denoted as .

4. The phase calculation method based on sequence pixel extreme value localization as described in claim 3, characterized in that, The wrapper phase value of a pixel is represented as: ； in, , Subpixel level obtained by single-peak one-dimensional Gaussian fitting The minimum and maximum values ​​of .

5. The phase calculation method based on sequence pixel extreme value localization as described in claim 4, characterized in that, The phase expansion in S4 is represented as follows: ； in, , It is a level.

Images