Speckle image noise suppression method based on digital image correlation method

Abstract

The invention discloses a speckle image noise suppression method based on a digital image correlation method, and the method comprises the steps: carrying out the noise suppression of a speckle image sequence collected by a camera, selecting a certain image in the sequence as a reference image, and carrying out the pixel-by-pixel digital image correlation of the reference image and other images; obtaining the sub-pixel positions of the whole pixel point of the reference image in other images, solving the gray value of the position by using a high-precision interpolation algorithm, carrying out average mapping transformation on the gray value to replace the gray value of the corresponding point of the reference image, and carrying out the steps on all the images in the sequence to complete primary noise suppression of the whole image sequence; and repeating the whole suppression process until a convergence condition is met. According to the method, errors caused by speckle image noise are effectively reduced, and the measurement precision is improved.

Description

Technical Field

[0001] The present invention relates to the field of optical measurement experimental solid mechanics and image measurement technology, and in particular to a speckle image noise suppression method based on a digital image correlation method. Background Art

[0002] Optical measurement methods are widely used for three-dimensional structural topography and deformation measurement due to their non-contact, non-destructive, full-field, and rapid performance. Digital image correlation methods and the fusion of line projection and digital image correlation utilize the surface texture of the object being measured or a prepared speckle pattern for matching. However, the images captured by the camera contain noise, which significantly affects measurement accuracy. Traditional noise suppression algorithms are mainly divided into frequency domain and spatial domain methods. Frequency domain-based algorithms require speckle and noise information to be distinguished spectrally. However, both noise and speckle are complex in images, making it difficult to achieve noise suppression in speckle images through frequency domain separation. Spatial domain-based algorithms lose some of the texture features of the speckle image, reducing its contrast and the measurement accuracy of digital image correlation methods. Image noise suppression algorithms combined with deep learning have also developed rapidly in recent years. However, deep learning-based denoising methods require pre-training and consume high computational resources. Therefore, there is still a need for noise suppression methods for speckle images to improve measurement accuracy. Summary of the Invention

[0003] The object of the present invention is to provide a speckle image noise suppression method based on a digital image correlation method to improve measurement accuracy.

[0004] The technical solution to achieve the purpose of the present invention is: a method for suppressing speckle image noise based on a digital image correlation method, comprising the following steps:

[0005] Step 1: Build the experimental system, including the following equipment: electronic computer, industrial camera and industrial lens, projector, speckle pattern, optical platform, and object to be tested;

[0006] When using a camera-projector system for measurement, the industrial camera and projector are fixed on an optical platform. Red and blue speckles are prepared on the surface of the object to be measured by water transfer printing and fixed on the optical platform. The angle and position of the industrial camera and projector are adjusted so that the object to be measured is located in the center of the camera's field of view. The camera and projector are focused on the object to be measured and connected to an electronic computer via wires.

[0007] When using a dual-camera system for measurement, two industrial cameras are fixed on an optical platform with an angle of about 35 degrees between the cameras. Black and white speckles are prepared on the surface of the object to be measured by water transfer printing and fixed to the optical platform. The angle and position of the industrial cameras are adjusted so that the object to be measured is located in the center of the camera's field of view. The two cameras are focused on the object to be measured and connected to an electronic computer via wires.

[0008] Step 2: When using a camera-projector system for measurement, a red pattern is projected by a projector, and the speckle pattern on the surface of the object to be measured is captured by an industrial camera to obtain a set of speckle image sequences;

[0009] When using a dual-camera system for measurement, two industrial cameras are used to capture the speckle pattern on the surface of the object to be measured, and two sets of speckle image sequences are obtained;

[0010] Step 3: Select an image from a set of speckle image sequences as a reference image, perform pixel-by-pixel digital image correlation with other images in the sequence, and obtain the sub-pixel position of the whole pixel of the reference image in other images;

[0011] Step 4: Use a high-precision interpolation algorithm to solve the grayscale values of the sub-pixel positions of other images in the sequence, and take the grayscale average value to map and transform to replace the grayscale value of the corresponding point in the reference image;

[0012] Step 5: Perform steps 3 to 4 on all images in the sequence to complete noise suppression for the entire image sequence.

[0013] In step 6, the image sequence obtained in step 5 is returned to step 3 to continue iterative optimization until the absolute difference between the average pixel displacements of two adjacent iterations meets the convergence condition or the number of iterations is reached, that is, the image noise suppression is completed and a set of low-noise image sequences is obtained.

[0014] Furthermore, in step 3, a certain image in the speckle image sequence is selected as the reference image, and pixel-by-pixel digital image correlation is performed with other images in the sequence to obtain the sub-pixel position of the whole pixel of the reference image in the other images. The specific method is as follows:

[0015] The base image is used as the reference image, and another image in the sequence is used as the deformed image. Assuming that the coordinates of a certain image point in the reference image are (u, v), a rectangular sub-region of size (2N+1)×(2M+1) is selected around the point as the center. N and M represent the pixel height and pixel width of the sub-region respectively. The deformed sub-regions of the same size are selected in the deformed image, and the similarity between the two sub-regions is measured by the correlation function, which is expressed as follows:

[0016]

[0017] Where, f(ui ,v j ) represents the point in the reference sub-region (u i ,v j ) grayscale value, g(u i ′,v′ j ) represents the point in the deformation sub-region (u i ′,v′ j ) grayscale value, f m and g m Denote the average grayscale values in the reference sub-region and the deformed sub-region, respectively. The subscripts i and j represent the position index relative to the sub-region center point. The center point (u′, v′) of the deformed sub-region with the smallest correlation coefficient is selected as the corresponding point of the reference image point (u, v).

[0018] Furthermore, in step 4, a high-precision interpolation algorithm is used to solve the grayscale values of the sub-pixel positions of other images in the sequence, and the grayscale average value is used to map and transform the grayscale value of the corresponding point in the reference image. The specific interpolation method is:

[0019] Assume that the coordinates of the point to be interpolated in the deformed image are (u′, v′), and with this point as the center, select a 4×4 neighborhood of integer pixels, and use i and j to number the 16 integer pixels, i = 0, 1, 2, 3, j = 0, 1, 2, 3, and then use bicubic interpolation to interpolate the grayscale. The expression is:

[0020]

[0021] Where a ij is the weight coefficient, which is determined by the cubic interpolation kernel function W(x):

[0022] a ij =W(u′-u′ i )×W(v′-v′ j ) (3)

[0023]

[0024] Taking a = 0.5 and applying equations (3) and (4) in parallel, we can get the weight coefficient a of each pixel in the neighborhood: ij , substituting the weight coefficient into formula (2) can obtain the grayscale value of the sub-pixel position (u′, v′) in the deformed image.

[0025] Furthermore, in step 6, the image sequence obtained in step 5 is returned to step 3 to continue iterative optimization until the absolute difference between the average pixel displacements of two adjacent iterations meets the convergence condition or the number of iterations is reached, that is, the image noise suppression is completed and a set of low-noise image sequences is obtained. The specific method is:

[0026] Return the image sequence obtained in step 5 to step 3 and continue iterative optimization until the absolute difference in the average pixel displacement of two adjacent iterations is less than 10 -4 If the number of pixel or iterations is greater than or equal to 7, image noise suppression is completed.

[0027] A speckle image noise suppression system based on a digital image correlation method implements the speckle image noise suppression method based on the digital image correlation method to achieve speckle image noise suppression based on the digital image correlation method, comprising an electronic computer, an industrial camera and industrial lens, a projector, a speckle pattern, an optical platform, an object to be measured, and five modules for respectively performing steps 2 to 6.

[0028] A computer device comprises a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the method for suppressing speckle image noise based on the digital image correlation method is implemented to achieve speckle image noise suppression based on the digital image correlation method.

[0029] A computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the method for suppressing speckle image noise based on a digital image correlation method is implemented to achieve speckle image noise suppression based on a digital image correlation method.

[0030] Compared with existing technologies, this invention offers the following significant advantages: 1) It uniquely applies digital image correlation methods to speckle image noise suppression, effectively reducing the errors caused by speckle image noise. 2) It improves measurement accuracy in specific application scenarios through algorithmic analysis and has achieved good results in experiments. 3) It requires simple equipment and has wide applicability. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 Schematic diagram of the experimental setup for measurement using a camera-projector system according to the present invention.

[0032] Figure 2 Schematic diagram of the experimental setup for measurement using a dual-camera system in the present invention.

[0033] Figure 3 Flow chart of the method of the present invention.

[0034] Figure 4 The relationship between the number of iterations of the method of the present invention and the mean absolute error and standard deviation of the horizontal displacement when simulating the measurement of the camera-projector system is shown.

[0035] Figure 5 This is the relationship between the number of iterations of the method of the present invention and the mean absolute error and standard deviation of the horizontal displacement when simulating dual-camera system measurement.

[0036] 1: Electronic computer;

[0037] 2: Industrial cameras;

[0038] 3: Industrial lens;

[0039] 4: Optical platform;

[0040] 5: Object to be measured;

[0041] 6: red and blue speckle;

[0042] 7: Projector;

[0043] 8: Black and white speckle. DETAILED DESCRIPTION

[0044] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0045] like Figure 1 and Figure 2 As shown, a speckle image noise suppression method based on a digital image correlation method includes the following steps:

[0046] Step 1: Build an experimental system, including the following equipment: electronic computer, industrial camera, industrial lens, projector, speckle pattern, optical platform, and object to be tested;

[0047] Figure 1 The camera-projector measurement system is constructed as follows: fix the industrial camera and projector on an optical platform, prepare red and blue speckles on the surface of the object to be measured by water transfer printing and fix them on the optical platform, adjust the angle and position of the industrial camera and projector so that the object to be measured is located in the middle of the camera's field of view, focus the camera and projector on the object to be measured, and connect them to an electronic computer through wires.

[0048] The specific construction process of the dual-camera system measurement system is as follows: two industrial cameras are fixed on an optical platform with an angle of about 35° between the cameras. Black and white speckle patterns are prepared on the surface of the object to be measured by water transfer printing and fixed on the optical platform. The angle and position of the industrial cameras are adjusted so that the object to be measured is located in the center of the camera's field of view. The two cameras are focused on the object to be measured and connected to an electronic computer via wires.

[0049] Step 2: When using a camera-projector system for measurement, a red pattern is projected by a projector, and the speckle pattern on the surface of the object to be measured is captured by an industrial camera to obtain a set of speckle image sequences;

[0050] When using a dual-camera system for measurement, two industrial cameras are used to capture the speckle pattern on the surface of the object to be measured, resulting in two sets of speckle image sequences. In the following steps, the two sets of speckle image sequences need to be operated independently. The speckle image sequences of the two cameras are processed completely independently, including independent selection of reference images, independent execution of digital image correlation matching, and independent convergence judgment. There is no data interaction or joint optimization between the two sets of sequences.

[0051] Step 3: Select an image in the speckle image sequence as the reference image, perform pixel-by-pixel digital image correlation with other images in the sequence, and obtain the sub-pixel position of the whole pixel of the reference image in other images.

[0052] The baseline image is used as the reference image, and another image in the sequence is used as the deformed image. Assume that the coordinates of a certain image point in the reference image are (u, v). With this point as the center, a rectangular sub-region of size (2N+1)×(2M+1) is selected around it. N and M represent the pixel height and pixel width of the sub-region respectively. In the deformed image, a deformed sub-region of the same size is selected, and the similarity between the two sub-regions is measured by the correlation function. The expression is:

[0053]

[0054] Where, f(u i ,v j ) represents the point in the reference sub-region (u i ,v j ) grayscale value, g(u i ′,v′ j ) represents the point in the deformation sub-region (u i ′,v′ j ) grayscale value, f m and g m Denote the average grayscale values in the reference sub-region and the deformed sub-region, respectively. The subscripts i and j represent the position index relative to the sub-region center point. The center point (u′, v′) of the deformed sub-region with the smallest correlation coefficient is selected as the corresponding point of the reference image point (u, v).

[0055] In some embodiments, a square sub-area of size 21×21 is selected, and formula (1) is converted to

[0056]

[0057] Step 4: Use a high-precision interpolation algorithm to solve the grayscale values of sub-pixel positions in other images in the sequence, and take the grayscale average value to map and transform instead of the grayscale value of the corresponding point in the reference image.

[0058] Assume that the coordinates of the point to be interpolated in the deformed image are (u′, v′), and with this point as the center, select a 4×4 neighborhood of integer pixels, and use i and j to number the 16 integer pixels, i = 0, 1, 2, 3, j = 0, 1, 2, 3, and then use bicubic interpolation to interpolate the grayscale. The expression is:

[0059]

[0060] Where a ij is the weight coefficient, which is determined by the cubic interpolation kernel function W(x):

[0061] a ij =W(u′-u′ i )×W(v′-v′ j ) (3)

[0062]

[0063] Taking a = 0.5 and applying equations (3) and (4) in parallel, we can get the weight coefficient a of each pixel in the neighborhood: ij , substituting the weight coefficient into formula (2) can obtain the grayscale value of the sub-pixel position (u′, v′) in the deformed image.

[0064] Step 5: Perform steps 3 to 4 on all images in the sequence to complete noise suppression for the entire image sequence.

[0065] Step 6: Return the image sequence obtained in step 5 to step 3 and continue iterative optimization until the absolute difference in the average pixel displacement between two adjacent iterations is less than 10 -4 If the number of pixel or iterations is greater than or equal to 7, the image noise suppression is completed and a set of low-noise image sequences is obtained.

[0066] Assume that the average pixel displacement calculated by correlation between the reference image after the Kth and K-1th iterations and the other digital images in the sequence is s K and s K－1 , when the absolute difference between two adjacent average pixel displacements |s K -s K-1 Less than 10 - 4 pixel, the method converges.

[0067] Example 1

[0068] In order to verify the effectiveness of the solution of the present invention when using a camera-projector system for measurement, the following numerical simulation experiments are performed.

[0069] First, a 500×500 pixel speckle image with a speckle diameter of 5 pixels was generated as the original image. Gaussian random noise with a mean of 0 and a standard deviation of 10 was added to the original image to create a reference image. The original image was then horizontally shifted by 1 to 10 pixels to generate 10 sets of shifted images. Independent random noise with a mean of 0 and a standard deviation of 10 was added to each shifted image to create the deformed image.

[0070] Select 5 and 10 deformed images and the reference image to form an image sequence, and use the above steps 2 to 6 to perform noise suppression, and the results are as follows Figure 4 As shown. Among them, Figure 4 (a) is the result of suppressing 5 deformed images; Figure 4 (b) shows the suppression results of 10 selected deformed images; the black line represents the mean absolute error of the horizontal displacement calculated by the digital image correlation method; the blue line represents the standard deviation of the horizontal displacement calculated by the digital image correlation method. It can be clearly seen that the patent of this invention can effectively reduce the error caused by speckle image noise and improve measurement accuracy.

[0071] Example 2

[0072] In order to verify the effectiveness of the solution of the present invention when measuring with a dual-camera system, the following numerical simulation experiments are performed.

[0073] First, a 500×500 pixel speckle image with a speckle diameter of 5 pixels was generated as the original image. Gaussian random noise with a mean of 0 and standard deviations of 20 and 30 was added to the original image to obtain the reference images for cameras 1 and 2. The original image was horizontally shifted by 1 to 10 pixels, generating 10 sets of shifted images. Independent random noise with a mean of 0 and standard deviations of 20 and 30 was added to each shifted image to obtain the deformed images for cameras 1 and 2.

[0074] For camera 1, the noise suppression is performed using steps 2 to 6 above, and the result is as follows: Figure 5 (a) As shown; for camera 2, the noise suppression is performed using steps 2 to 6 above, and the result is as follows Figure 5 (b) The black line represents the mean absolute error of the horizontal displacement calculated by the digital image correlation method; the blue line represents the standard deviation of the horizontal displacement calculated by the digital image correlation method. It is clear that the present invention can effectively reduce the error caused by speckle image noise and improve measurement accuracy.

[0075] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0076] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A method for suppressing speckle image noise based on digital image correlation method, characterized in that: The following steps are involved: Step 1: Build the experimental system, including the following equipment: electronic computer, industrial camera and industrial lens, projector, speckle pattern, optical platform, and object to be tested; When using a camera-projector system for measurement, the industrial camera and projector are fixed on an optical platform. Red and blue speckles are prepared on the surface of the object to be measured by water transfer printing and fixed on the optical platform. The angle and position of the industrial camera and projector are adjusted so that the object to be measured is located in the center of the camera's field of view. The camera and projector are focused on the object to be measured and connected to an electronic computer via wires. When using a dual-camera system for measurement, two industrial cameras are fixed on an optical platform with an angle of about 35 degrees between the cameras. Black and white speckles are prepared on the surface of the object to be measured by water transfer printing and fixed to the optical platform. The angle and position of the industrial cameras are adjusted so that the object to be measured is located in the center of the camera's field of view. The two cameras are focused on the object to be measured and connected to an electronic computer via wires. Step 2: When using a camera-projector system for measurement, a red pattern is projected by a projector, and the speckle pattern on the surface of the object to be measured is captured by an industrial camera to obtain a set of speckle image sequences; When using a dual-camera system for measurement, two industrial cameras are used to capture the speckle pattern on the surface of the object to be measured, and two sets of speckle image sequences are obtained; Step 3: Select an image from a set of speckle image sequences as a reference image, perform pixel-by-pixel digital image correlation with other images in the sequence, and obtain the sub-pixel position of the whole pixel of the reference image in other images; Step 4: Use a high-precision interpolation algorithm to solve the grayscale values of the sub-pixel positions of other images in the sequence, and take the grayscale average value to map and transform to replace the grayscale value of the corresponding point in the reference image; Step 5: Perform steps 3 to 4 on all images in the sequence to complete noise suppression for the entire image sequence. In step 6, the image sequence obtained in step 5 is returned to step 3 to continue iterative optimization until the absolute difference between the average pixel displacements of two adjacent iterations meets the convergence condition or the number of iterations is reached, that is, the image noise suppression is completed and a set of low-noise image sequences is obtained.

2. The method for suppressing speckle image noise based on digital image correlation method according to claim 1, characterized in that: Step 3: Select an image in the speckle image sequence as the reference image, perform pixel-by-pixel digital image correlation with other images in the sequence, and obtain the sub-pixel position of the whole pixel of the reference image in other images. The specific method is as follows: The base image is used as the reference image, and another image in the sequence is used as the deformed image. Assuming that the coordinates of a certain image point in the reference image are (u, v), a rectangular sub-region of size (2N+1)×(2M+1) is selected around the point as the center. N and M represent the pixel height and pixel width of the sub-region respectively. The deformed sub-regions of the same size are selected in the deformed image, and the similarity between the two sub-regions is measured by the correlation function, which is expressed as follows: Where, f(u i ,v j ) represents the point in the reference sub-region (u i ,v j ) grayscale value, g(u i ′,v′ j ) represents the point in the deformation sub-region (u i ′,v′ j ) grayscale value, f m and g m Denote the average grayscale values in the reference sub-region and the deformed sub-region, respectively. The subscripts i and j represent the position index relative to the sub-region center point. The center point (u′, v′) of the deformed sub-region with the smallest correlation coefficient is selected as the corresponding point of the reference image point (u, v).

3. The method for suppressing speckle image noise based on digital image correlation method according to claim 1, characterized in that: Step 4: Use a high-precision interpolation algorithm to solve the grayscale values of the sub-pixel positions of other images in the sequence, and take the grayscale average value to map and transform to replace the grayscale value of the corresponding point in the reference image. The specific interpolation method is: Assume that the coordinates of the point to be interpolated in the deformed image are (u′, v′), and with this point as the center, select a 4×4 neighborhood of integer pixels, and use i and j to number the 16 integer pixels, i = 0, 1, 2, 3, j = 0, 1, 2, 3, and then use bicubic interpolation to interpolate the grayscale. The expression is: Where a ij is the weight coefficient, which is determined by the cubic interpolation kernel function W(x): a ij =W(u′-u i ′)×W(v′-v′ j ) (3) Taking a = 0.5 and applying equations (3) and (4) in parallel, we can get the weight coefficient a of each pixel in the neighborhood: ij , substituting the weight coefficient into formula (2) can obtain the grayscale value of the sub-pixel position (u′, v′) in the deformed image.

4. The method for suppressing speckle image noise based on digital image correlation method according to claim 1, characterized in that: Step 6: Return the image sequence obtained in step 5 to step 3 and continue iterative optimization until the absolute difference between the average pixel displacements of two adjacent iterations meets the convergence condition or the number of iterations is reached. This means that image noise suppression is completed and a set of low-noise image sequences is obtained. The specific method is as follows: Return the image sequence obtained in step 5 to step 3 and continue iterative optimization until the absolute difference in the average pixel displacement of two adjacent iterations is less than 10 -4 If the number of pixel or iterations is greater than or equal to 7, image noise suppression is completed.

5. A speckle image noise suppression system based on a digital image correlation method, implementing the speckle image noise suppression method based on a digital image correlation method according to any one of claims 1 to 4 to achieve speckle image noise suppression based on the digital image correlation method, comprising an electronic computer, an industrial camera and industrial lens, a projector, a speckle pattern, an optical platform, an object to be measured, and five modules for respectively performing steps 2 to 6.

6. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for suppressing speckle image noise based on a digital image correlation method according to any one of claims 1 to 4 is implemented to achieve speckle image noise suppression based on a digital image correlation method.

7. A computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the computer program implements the speckle image noise suppression method based on the digital image correlation method according to any one of claims 1 to 4, thereby achieving speckle image noise suppression based on the digital image correlation method.

Images