Image vignetting correction method, image processing device and application

Abstract

The present application provides a kind of image vignetting correction method, image processing device and application, this method is under the condition of predetermined photographing environment, using camera to shoot template image under the condition of welding and carry out gray value sampling, surface fitting, with the maximum value of fitting surface equation divided by the surface fitting value of template image, and the result is normalized to obtain normalized dark corner compensation model, based on the normalized dark corner compensation model, the image to be corrected that camera shoots under the condition of predetermined photographing environment is corrected.The present application can quickly find the dark corner compensation model based on template image using simple surface fitting method, correct the image to be corrected, improve image quality, without obtaining complex geometric parameters of optical system in advance, has wide adaptability.The image processing device of the present application can be applied to droplet digital PCR system to improve the accuracy of PCR detection results, and can also be applied to single cell sorting system to facilitate the identification and positioning of nozzle, improve the efficiency of single cell sorting.

Description

Technical Field

[0001] This invention belongs to the fields of microfluidics, droplet digital PCR system development and biomedical image processing technology, and relates to an image vignetting correction method, image processing device and application. Background Technology

[0002] Camera lenses inevitably introduce vignetting during the imaging process, causing uneven brightness in the image, resulting in a brighter center and darker edges. To maintain image uniformity, appropriate image processing algorithms are needed to correct vignetting.

[0003] Droplet-based digital PCR (polymerase chain reaction) using inkjet printing technology generates droplets as microreaction containers within a microfluidic chip. DNA templates are amplified within these microreaction containers and then transferred to microchannels. An image acquisition system images the droplets generated by the chip, and image processing algorithms are used to statistically analyze the bright-field and fluorescence images to count the number of positive and negative droplets, achieving absolute quantification of digital PCR.

[0004] This system uses a CCD camera to image droplets in microchannels. The vignetting produced during camera imaging affects the accuracy of droplet segmentation within the channel sheet. Inherent limitations of the optical imaging system and angular deviations in the light received by the sensor during camera imaging can affect the uniformity of digital PCR fluorescence images. To reduce the impact of uneven camera light exposure on subsequent image processing algorithms, vignetting caused by the camera needs to be compensated for to improve image quality. Summary of the Invention

[0005] In view of the shortcomings of the prior art described above, the purpose of this invention is to provide an image vignetting correction method, an image processing device, and an application to solve the problem of poor image quality caused by uneven light sensitivity of the camera in the prior art.

[0006] To achieve the above and other related objectives, the present invention provides an image vignetting correction method, comprising the following steps:

[0007] Use the camera to capture a template image under a pre-defined shooting environment and in a focused state;

[0008] Perform grayscale sampling on the template image;

[0009] The sampled gray values ​​are fitted to a surface to obtain the equation of the fitted surface for the image gray values;

[0010] Divide the maximum value of the fitted surface equation by the surface fitting value of the template image, and normalize the result to obtain a normalized vignetting compensation model.

[0011] The image to be corrected is obtained by correcting the image captured by the camera under the predetermined shooting environment based on the normalized vignetting compensation model.

[0012] Optionally, the surface equation may be a polynomial fitting equation or a Gaussian surface fitting equation.

[0013] Optionally, the surface equation is a two-variable quadratic polynomial fitting equation.

[0014] Optionally, the calculation formula used to divide the maximum value of the fitted surface equation by the surface fitting value of the template image includes:

[0015]

[0016] Where g(x, y) represents the vignetting compensation model, x represents the position coordinates of the pixel in the length direction of the image, y represents the position coordinates of the pixel in the width direction of the image, c(x, y) represents the surface fitting value, and max(c(x, y)) represents the maximum value of the fitted surface equation.

[0017] Optionally, the calculation formula used to normalize the results to obtain the normalized vignetting compensation model includes:

[0018]

[0019] Where w(x, y) represents the normalized vignetting model, and max(g(x, y)) represents the maximum value of g(x, y).

[0020] Optionally, the template image is selected from one of the images to be corrected.

[0021] Optionally, the camera includes a charge-coupled device (CCD) camera.

[0022] The present invention also provides an image processing apparatus, the image processing apparatus including a processor and a memory connected to the processor, the memory being used to store a program, and the processor being used to run the program stored in the memory to perform the image vignetting correction method.

[0023] The present invention also provides an application of an image processing device, wherein the image processing device described above is applied to a droplet digital PCR system, the droplet digital PCR system using the image processing device to perform vignetting correction on the droplet fluorescence image obtained by camera scanning, and counts the number of positive and negative droplets based on the corrected droplet fluorescence image.

[0024] Optionally, the droplets are obtained by pushing liquid reagents into droplet-generating oil using a thermally printed chip.

[0025] The present invention also provides an application of an image processing device, wherein the image processing device described above is applied to a single-cell sorting system, the single-cell sorting system delivers a cell suspension to multiple nozzles of a thermally printed chip, the image processing device performs vignetting correction on the nozzle image captured by the camera, and analyzes the cell condition inside the nozzle based on the corrected nozzle image to perform single-cell sorting.

[0026] As described above, the image vignetting correction method of the present invention uses a camera to capture a template image under a predetermined shooting environment and a focusing condition. The template image is then sampled for grayscale values. A surface fitting is performed on the sampled grayscale values ​​to obtain an image grayscale value fitting surface equation. The maximum value of the fitting surface equation is divided by the surface fitting value of the template image, and the result is normalized to obtain a normalized vignetting compensation model. Based on the normalized vignetting compensation model, the image to be corrected captured by the camera under the predetermined shooting environment is corrected to obtain a corrected image. The image vignetting correction method of the present invention can quickly find a vignetting compensation model based on a template image using a simple surface fitting method, correcting the image to be corrected and improving image quality. It does not require prior acquisition of complex geometric parameters of the optical system and has wide applicability. The image processing device using the image vignetting correction method of the present invention can be applied to droplet digital PCR systems to perform vignetting correction on droplet fluorescence images to improve the accuracy of PCR detection results, and can also be applied to single-cell sorting systems to perform vignetting correction on nozzle images to facilitate nozzle identification and positioning, improving single-cell sorting efficiency. Attached Figure Description

[0027] Figure 1 The flowchart shown is a process for image vignetting correction according to the present invention.

[0028] Figure 2 The image displayed is a background image randomly captured by a charge-coupled device (CCD) camera under a pre-defined shooting environment and during focusing.

[0029] Figure 3 The display shows a normalized grayscale value fitted surface obtained by performing polynomial fitting on the template image.

[0030] Figure 4 The normalized vignetting compensation model is obtained by dividing the maximum value of the fitted surface equation by the fitted surface value of the image, and then normalizing the result.

[0031] Figure 5 The image shown is of the droplet before vignetting was performed in Example 3.

[0032] Figure 6 Displayed as Figure 5 The pixel fitting curve at the diagonal line from the top left corner to the bottom right corner.

[0033] Figure 7 The selected vignetting correction template image is displayed.

[0034] Figure 8 Displayed as based on Figure 7 The vignetting correction template is calculated from the vignetting correction template image shown.

[0035] Figure 9 Displayed as based on Figure 8 The vignetting correction template shown is used to correct the vignetting image.

[0036] Figure 10 Displayed as along Figure 9 Collect the grayscale values ​​of the image from the top left corner to the bottom right corner and plot the grayscale value curve.

[0037] Component designation explanation

[0038] Steps S1 to S5 Detailed Implementation

[0039] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention.

[0040] Please see Figures 1 to 10 It should be noted that the illustrations provided in this embodiment are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0041] Example 1

[0042] This embodiment provides an image vignetting correction method. Please refer to [link / reference]. Figure 1 The flowchart of this method is shown below, including the following steps:

[0043] S1: Use the camera to capture a template image under the specified shooting conditions and focusing conditions;

[0044] S2: Perform grayscale sampling on the template image;

[0045] S3: Perform surface fitting on the sampled gray values ​​to obtain the image gray value fitting surface equation;

[0046] S4: Divide the maximum value of the fitted surface equation by the surface fitting value of the template image, and normalize the result to obtain a normalized vignetting compensation model.

[0047] S5: Based on the normalized vignetting compensation model, the image to be corrected captured by the camera under the predetermined shooting environment is corrected to obtain a corrected image.

[0048] Specifically, in step S3, the surface equation can be a polynomial fitting equation or a Gaussian surface fitting equation. In this embodiment, a polynomial fitting equation is preferred because Gaussian surface fitting or Lagrange surface fitting requires a large fitting window and may produce large errors.

[0049] As an example, a bivariate quadratic polynomial is used for surface fitting to obtain relevant parameters. The corresponding surface fitting function formula is as follows:

[0050] c(x,y)=a0+a1x+a2y+a3x 2 +a4xy+a5y 2 ;

[0051] Where x represents the position coordinate of the pixel in the length direction of the image, y represents the position coordinate of the pixel in the width direction of the image, c(x, y) represents the surface fitting value, a0 is a constant, and a1, a2, a3, a4, and a5 are the coefficients of each term.

[0052] Specifically, in step S4, the calculation formula used to divide the maximum value of the fitted surface equation by the surface fitting value of the template image includes:

[0053]

[0054] Where g(x, y) represents the vignetting compensation model, x represents the position coordinates of the pixel in the length direction of the image, y represents the position coordinates of the pixel in the width direction of the image, c(x, y) represents the surface fitting value, and max(c(x, y)) represents the maximum value of the fitted surface equation.

[0055] The calculation formula used to normalize the results to obtain the normalized vignetting compensation model includes:

[0056]

[0057] Where w(x, y) represents the normalized vignetting model, and max(g(x, y)) represents the maximum value of g(x, y).

[0058] For example, please refer to Figure 2The image displayed is a background image randomly captured by a charge-coupled device (CCD) camera under pre-defined focusing conditions. This background image is used as the template image. It should be noted that the image has been stretched to grayscale for easier observation (this is unnecessary in actual processing), and vignetting is clearly visible after adjusting the image contrast.

[0059] It should be noted that the background image mentioned above does not contain a specific object to be analyzed. In other embodiments, the template image may also be selected from one of the images to be corrected that contains a specific object to be analyzed.

[0060] Please see Figure 3 The display shows the normalized grayscale value fitting surface obtained by polynomial fitting of the template image, where the z-axis represents the normalized image grayscale value, and the x and y axes represent the length and width coordinates of the image.

[0061] Please see Figure 4 The result is a normalized vignetting compensation model obtained by dividing the maximum value of the fitted surface equation by the fitted surface value of the image, and then normalizing the result. Using this vignetting compensation model to correct images captured by a camera can eliminate image vignetting.

[0062] The image vignetting correction method in this embodiment can perform real-time compensation on the acquired image. It has low complexity and does not require complex image processing algorithms. Furthermore, this method does not require prior acquisition of the geometric parameters of the optical system; image quality can be improved simply by fitting grayscale values ​​of a template image. It automatically calculates vignetting compensation models for different shooting scenarios, demonstrating excellent adaptability.

[0063] Example 2

[0064] This embodiment provides an image processing apparatus, which includes a processor and a memory connected to the processor. The memory is used to store a program, and the processor is used to run the program stored in the memory to perform the image vignetting correction method as described in Embodiment 1.

[0065] Example 3

[0066] This embodiment applies the image processing device described in Embodiment 2 to a droplet digital PCR system. The droplet digital PCR system uses the image processing device to perform vignetting correction on the droplet fluorescence image obtained by camera scanning, and counts the number of positive and negative droplets based on the corrected droplet fluorescence image.

[0067] As an example, the droplets are obtained by pushing liquid reagents into droplet-generating oil using a thermally printed chip.

[0068] Please see Figure 5The image shown is of the droplets before vignetting, where brighter droplets are positive and darker droplets are negative. Please refer to [link to image]. Figure 6 Displayed as Figure 5 The pixel fitting curves from the top left corner to the bottom right corner of the image show obvious vignetting.

[0069] Please see Figure 7 and Figure 8 ,in, Figure 7 Displayed as the selected vignetting correction template image. Figure 8 The image displayed is a vignetting correction template calculated based on the vignetting correction template image. The result of correcting the vignetting image is as follows: Figure 9 As shown, along Figure 9 Collect the grayscale values ​​of the image from the top left corner to the bottom right corner and plot the grayscale value curve as shown below. Figure 10 As shown in the figure, the grayscale value fitting curve of the corrected image shows that the image vignetting has been well compensated.

[0070] This embodiment addresses the uneven illumination phenomenon that occurs during droplet digital PCR image scanning. Before scanning the droplet fluorescence image, only one background image needs to be taken. Then, a simple polynomial surface fitting method can be used to quickly find the camera vignetting compensation model, which can then be used to compensate for the droplet fluorescence image obtained in subsequent scans in real time, thereby improving image quality.

[0071] This embodiment provides real-time compensation for droplet digital PCR image acquisition systems. The method is low in complexity and does not require sophisticated image processing algorithms. It obtains a vignetting compensation model of the camera under specific shooting conditions and uses this model to correct the real-time captured image. This eliminates the need for pre-obtaining complex geometric parameters of the optical system; only a background image of the camera under focusing conditions needs to be input as a template image. The grayscale value fitting method of the background image is used to improve the quality of the fluorescent droplet image. The vignetting compensation model is automatically calculated for different shooting scenarios, demonstrating good and wide adaptability. Using this image vignetting correction method to correct vignetting in droplet digital PCR fluorescence images results in more uniform image pixels, making it easier to distinguish between positive and negative droplet signals.

[0072] Example 4

[0073] This embodiment applies the image processing device described in Embodiment 2 to a single-cell sorting system. The single-cell sorting system delivers a cell suspension to multiple nozzles of a thermally printed chip. The image processing device performs vignetting correction on the nozzle images captured by the camera and analyzes the cell condition inside the nozzles based on the corrected nozzle images to perform single-cell sorting.

[0074] Specifically, the aforementioned image vignetting correction method is incorporated as part of the image processing algorithm in the single-cell sorting system. Vivid correction improves the vignetting angle of the image, which is beneficial for the subsequent identification and positioning of nozzles and improves the efficiency of single-cell sorting.

[0075] In summary, the image vignetting correction method of the present invention uses a camera to capture a template image under a predetermined shooting environment and a focusing condition. The template image is then sampled for grayscale values. A surface fitting is performed on the sampled grayscale values ​​to obtain an image grayscale value fitting surface equation. The maximum value of the fitting surface equation is divided by the surface fitting value of the template image, and the result is normalized to obtain a normalized vignetting compensation model. Based on the normalized vignetting compensation model, the image to be corrected captured by the camera under the predetermined shooting environment is corrected to obtain a corrected image. The image vignetting correction method of the present invention can quickly find a vignetting compensation model based on a template image using a simple surface fitting method, correcting the image to be corrected and improving image quality. It does not require prior acquisition of complex geometric parameters of the optical system and has wide applicability. The image processing device using the image vignetting correction method of the present invention can be applied to droplet digital PCR systems to perform vignetting correction on droplet fluorescence images to improve the accuracy of PCR detection results, and can also be applied to single-cell sorting systems to perform vignetting correction on nozzle images to facilitate nozzle identification and positioning, improving single-cell sorting efficiency. Therefore, this invention effectively overcomes the various shortcomings of the prior art and has high industrial application value.

[0076] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A method for correcting image vignetting, characterized in that, Includes the following steps: Use the camera to capture a template image under a pre-defined shooting environment and in a focused state; Perform grayscale sampling on the template image; The sampled gray values ​​are fitted to a surface to obtain the equation of the fitted surface for the image gray values; The maximum value of the surface equation fitted by the gray value of the image is divided by the surface fitting value of the template image, and the result is normalized to obtain a normalized vignetting compensation model. The image to be corrected is obtained by correcting the image captured by the camera under the predetermined shooting environment based on the normalized vignetting compensation model. The calculation formula used to divide the maximum value of the surface equation fitted by the grayscale value of the image by the surface fitting value of the template image includes: Where g(x, y) represents the vignetting compensation model, x represents the position coordinates of the pixel in the length direction of the image, y represents the position coordinates of the pixel in the width direction of the image, c(x, y) represents the surface fitting value of the template image, and max(c(x, y)) represents the maximum value of the surface equation fitted to the grayscale value of the image. The calculation formula used to normalize the results to obtain the normalized vignetting compensation model includes: Where w(x, y) represents the normalized vignetting compensation model, and max(g(x, y)) represents the maximum value of g(x, y).

2. The image vignetting correction method according to claim 1, characterized in that: The image grayscale value fitting surface equation is selected from either a polynomial fitting equation or a Gaussian surface fitting equation.

3. The image vignetting correction method according to claim 1, characterized in that: The image grayscale value fitting surface equation is a bivariate quadratic polynomial fitting equation.

4. The image vignetting correction method according to claim 1, characterized in that: The template image is selected from one of the images to be corrected.

5. The image vignetting correction method according to claim 1, characterized in that: The camera includes a charge-coupled device (CCD) camera.

6. An image processing apparatus, characterized in that: The image processing apparatus includes a processor and a memory connected to the processor. The memory is used to store a program, and the processor is used to run the program stored in the memory to perform the image vignetting correction method as described in any one of claims 1-5.

7. A method for applying an image processing device, characterized in that: The image processing device as described in claim 6 is applied to a droplet digital PCR system, wherein the droplet digital PCR system uses the image processing device to perform vignetting correction on the droplet fluorescence image obtained by camera scanning, and counts the number of positive and negative droplets based on the corrected droplet fluorescence image.

8. The application method of the image processing apparatus according to claim 7, characterized in that: The droplets are obtained by pushing liquid reagents into droplet-generating oil using a thermally printed chip.

9. A method for applying an image processing device, characterized in that: The image processing device as described in claim 6 is applied to a single-cell sorting system, wherein the single-cell sorting system delivers a cell suspension to multiple nozzles of a thermally printed chip, and the image processing device performs vignetting correction on the nozzle images captured by the camera, and analyzes the cell condition inside the nozzles based on the corrected nozzle images to perform single-cell sorting.

Images