Image enhancement method for printing plate inspection of a printing press

Abstract

This invention belongs to the field of image enhancement technology, specifically relating to an image enhancement method for inspecting printing plates in a printing press, comprising the following steps: S1, acquiring the original digital image of the printing plate to be inspected, performing multi-scale decomposition, detail remapping, and image reconstruction on the original digital image to obtain an enhanced image; S2, the multi-scale decomposition process is as follows: for any scale level from fine to coarse... l The input image is processed by calculating its local gradient covariance matrix, local information entropy map, and gradient consistency measure; a spatially variable filter kernel is constructed, and the input image is filtered using the spatially variable filter kernel to obtain the baseband image. This invention can enhance the contrast of weak defects while suppressing the amplification of background noise and irrelevant textures, resulting in an enhanced image with high signal-to-noise ratio, prominent defect features, and clear edge contours.

Description

Technical Field

[0001] This invention belongs to the field of image enhancement technology, specifically relating to an image enhancement method for inspecting printing plates for printing presses. Background Technology

[0002] In the printing industry, the quality of printing plates is a key factor determining the quality of printed products. Even minor defects, such as blemishes, scratches, missing dots, or deformation, can lead to the scrapping of large batches of printed materials. Therefore, quality inspection of printing plates is an indispensable part of the printing production process. Automated optical inspection systems based on digital image processing have gradually replaced manual visual inspection, becoming the mainstream inspection method. However, during the image acquisition stage, limitations imposed by the complex lighting environment of industrial sites, the reflective properties of the printing plate surface, and the inherent noise of imaging equipment often result in raw digital images with low contrast, blurred details, and a low signal-to-noise ratio. This causes subtle defect features to be submerged in complex background textures and noise, severely affecting the accuracy of subsequent defect segmentation and recognition algorithms. To address this issue, image enhancement technology is applied to the preprocessing stage of printing plate inspection. Existing enhancement methods, such as histogram equalization and Retinex algorithm, can improve the overall contrast of images to some extent. However, the former is prone to over-enhancing noise and losing local details, while the latter may introduce color distortion or halo artifacts. Therefore, they have limited applicability to printing plate inspection scenarios that require precise preservation of details and suppression of noise.

[0003] To better handle image details and structure, enhancement methods based on multi-scale decomposition have attracted attention, among which local Laplacian filtering is an advanced detail enhancement technique. By constructing a Laplacian pyramid, the image is decomposed into detail components at different scales. Nonlinear remapping is then applied to each level of detail component to reconstruct the image and achieve enhancement. Compared to sharpening algorithms, local Laplacian filtering effectively enhances local contrast while better suppressing halo artifacts. However, local Laplacian filtering still has some inherent drawbacks. During pyramid decomposition, spatially invariant isotropic filters, such as Gaussian filters, are typically used to generate the baseband image. These filters lack edge-preserving properties, blurring edges and fine structures during smoothing, resulting in impure detail information and affecting enhancement accuracy. Furthermore, detail remapping strategies are often relatively simple, such as designing gain functions solely based on local grayscale values ​​or variance. This fails to effectively distinguish between true edges, texture details, and random noise in the image. While enhancing target features, it often inevitably amplifies background noise, reducing the image's signal-to-noise ratio (SNR), which is particularly detrimental for detecting minute imperfections with already low SNR. Summary of the Invention

[0004] This invention provides an image enhancement method for printing plate inspection in printing presses, which solves the technical problems in the prior art where local Laplacian filtering blurs edges and fine structures due to the use of filters without edge preservation characteristics during pyramid decomposition, and the single detail remapping strategy easily amplifies background noise and reduces the signal-to-noise ratio.

[0005] In a first aspect, the present invention provides an image enhancement method for inspecting printing plates for a printing press, comprising the following steps:

[0006] S1. Obtain the original digital image of the printing plate to be inspected, and perform multi-scale decomposition, detail remapping and image reconstruction on the original digital image to obtain the enhanced image;

[0007] S2, the multi-scale decomposition process is as follows: for any scale level from fine to coarse... l Input image Processing: Calculate the input image Local gradient covariance matrix Local information entropy diagram And gradient consistency metric Construct a spatially variable filter kernel and use it to filter the input image. Filtering is performed to obtain the baseband image. Input image With baseband image The difference is used as the scale level. l Laplace detail ; for baseband images Downsampling is performed, and the resulting image is used as the input for the next scale level. ;

[0008] S3, the detail remapping process is as follows: for each scale level, calculate the input image... Local phase consistency map Based on the local phase consistency diagram Phase consistency value pc and gradient consistency metric Detail of Laplace Remapping is performed to obtain the remapping detail components. If the phase consistency value pc and the gradient consistency metric If all values ​​are greater than their respective high thresholds, then an S-shaped gain function is used for enhancement; if the phase consistency value pc is greater than the high threshold of the phase consistency value pc, then gradient consistency measurement... Not greater than gradient consistency metric If the low threshold is not greater than the high threshold of phase consistency value pc, then a logarithmic gain function is used for enhancement; if the phase consistency value pc is not greater than the high threshold of phase consistency value pc, then a soft threshold function is used for coefficient shrinkage.

[0009] Furthermore, in S2, anisotropy is determined by the local gradient covariance matrix. The eigenvectors determine the spatial support range, which is determined by the local information entropy map. The entropy value is adjusted in the opposite direction.

[0010] Furthermore, the image reconstruction process involves taking the input image at the coarsest scale level obtained from multi-scale decomposition and remapping it with the detail components after remapping at all scale levels. The image is enhanced by upsampling and accumulating the data step by step, from coarse to fine.

[0011] Furthermore, a spatially variable filter kernel is constructed, including:

[0012] For the input image For each pixel, calculate the local gradient covariance matrix within a 5×5 neighborhood window of that pixel. ;

[0013] For the input image Perform eigenvalue decomposition to obtain eigenvectors. and ,Will and The anisotropic direction serves as the spatially variable filter kernel;

[0014] Calculate the local information entropy map within a 5×5 neighborhood window. The spatial support range of the spatially variable filter kernel is set according to its entropy value, and the calculation formula is as follows: ,in It is a constant. represents the scale parameter of the spatially variable filter kernel in the principal direction.

[0015] Furthermore, if the phase consistency value pc and the gradient consistency metric... If all values ​​exceed their respective high thresholds, then an S-shaped gain function is used for enhancement, including:

[0016] When phase consistency value Greater than the first preset threshold and gradient consistency metric When the value is greater than the second preset threshold, the S-shaped gain function is used to adjust the Laplace detail components. Enhancement is performed to obtain remapped detail components. , The gain function The calculation method is as follows .

[0017] Furthermore, if the phase consistency value pc is greater than the high threshold of the phase consistency value pc, then the gradient consistency measure... Not greater than gradient consistency metric For low threshold values, a logarithmic gain function is used for enhancement, including:

[0018] When phase consistency value Greater than the first preset threshold and gradient consistency metric When the value is not greater than the second preset threshold, a logarithmic gain function is used to evaluate the Laplace detail components. Enhancement is performed to obtain remapped detail components. ,in It is a symbolic function.

[0019] Furthermore, if the phase consistency value pc is not greater than the high threshold of the phase consistency value pc, then a soft thresholding function is used for coefficient shrinkage, including:

[0020] When phase consistency value When the value is not greater than the first preset threshold, a soft thresholding function is used to evaluate the Laplace detail components. By performing coefficient shrinkage, the remapped detail components are obtained. ,in The shrinkage threshold is determined based on the local standard deviation of the image at the current scale level. To remap detail components, For Laplace detail components.

[0021] Furthermore, the input image at the coarsest scale level obtained from multi-scale decomposition is compared with the detail components after remapping across all scale levels. Upsampling and accumulation are performed step by step from coarse to fine to obtain the enhanced image, including:

[0022] Assume multi-scale decomposition This process involves several levels to obtain the input image at the coarsest scale level. And detail components after remapping all levels ;

[0023] Let the initial layer of the reconstructed image ;

[0024] According to Upsampling and accumulation are performed sequentially up to 0, and the calculation method is as follows:

[0025] ,in For bilinear interpolation upsampling operation, Enlarged to the size of the remapped detail component Same size, In the process of image reconstruction, the first l Reconstructed images at multiple levels;

[0026] The final image obtained This refers to image enhancement.

[0027] Furthermore, in S2, for the input image For each pixel in the algorithm, the horizontal gradient gx and vertical gradient gy of the pixel are calculated using the Sobel operator. A 5×5 window is taken around the pixel, and the local gradient covariance matrix is ​​calculated using the gradient values ​​of all pixels within the window. .

[0028] Furthermore, the method for obtaining the original digital image of the printing plate to be inspected includes: acquiring the surface image of the printing plate using an industrial camera, and uniformly converting the acquired color or grayscale image into a grayscale image as the original digital image.

[0029] The beneficial effects are as follows: In the multi-scale decomposition stage, a spatially variable filter kernel, jointly determined by the local gradient covariance matrix and local information entropy, is constructed, achieving a filtering process with edge-preserving characteristics. The anisotropy and spatial support range of the spatially variable filter kernel are controlled by the local structural characteristics of the image, which can avoid blurring of important structures such as the edges of printing plate defects and dots when smoothing the baseband image. In the detail remapping stage, by jointly utilizing local phase consistency and gradient consistency measures, the image content is differentiated. For highly structural defect features, ordinary textures, and noise areas, S-shaped strong gain, logarithmic moderate gain, and soft thresholding are applied respectively. This can enhance the contrast of weak defects while suppressing the amplification of background noise and irrelevant textures. The resulting enhanced image has a high signal-to-noise ratio, prominent defect features, and clear edge contours. Attached Figure Description

[0030] Figure 1 A flowchart of an image enhancement method for inspecting printing plates used in printing presses. Detailed Implementation

[0031] 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, not all, of the embodiments of the present invention. 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.

[0032] An embodiment of the image enhancement method for printing plate inspection of a printing press provided by the present invention:

[0033] like Figure 1 As shown, an image enhancement method for inspecting printing plates on a printing press includes the following steps:

[0034] S1: Obtain the original digital image of the printing plate to be inspected, and perform multi-scale decomposition, detail remapping, and image reconstruction on the original digital image to obtain an enhanced image.

[0035] The surface image of the printing plate is acquired using an industrial camera. The acquired color or grayscale images are then uniformly converted to grayscale images to serve as the raw digital image. The total number of levels in the multi-scale decomposition is determined, for example, 5 levels. The raw digital image is then used as the input image for level 0. Then, the decomposition begins, layer by layer.

[0036] S2, the multi-scale decomposition process is as follows: for any scale level from fine to coarse... l Input image Processing: Calculate the input image Local gradient covariance matrix Local information entropy diagram And gradient consistency metric Construct a spatially variable filter kernel, with anisotropy determined by the local gradient covariance matrix. The eigenvectors determine the spatial support range, which is determined by the local information entropy map. The entropy value is adjusted inversely; the input image is checked using a spatially variable filter. Filtering is performed to obtain the baseband image. Input image With baseband image The difference is used as the scale level. l Laplace detail ; for baseband images Downsampling is performed, and the resulting image is used as the input for the next scale level. .

[0037] For the input image For each pixel in the array, the horizontal gradient gx and vertical gradient gy of that pixel are calculated using operators such as Sobel. A local window, such as a 5×5 window, is taken around the pixel, and the local gradient covariance matrix is ​​calculated using the gradient values ​​of all pixels within the window. For local information entropy graphs Take a local window, such as a 7x7 window, around each pixel, and calculate the gray-level histogram of the pixels within the window. Calculate the information entropy of the local window using the Shannon entropy formula. The information entropy values ​​of all pixels constitute the information entropy map. For gradient consistency metrics Similarly, within a local window surrounding each pixel, the magnitude of the sum of the gradient vectors of all pixels within the window is calculated, and then divided by the sum of the magnitudes of the gradient vectors of all pixels within the window. The ratio is the gradient consistency measure of the pixel, with a value range between 0 and 1.

[0038] At each pixel, the local gradient covariance matrix of that pixel... Eigenvalue decomposition yields two eigenvalues ​​and two corresponding eigenvectors. These two eigenvectors indicate the principal and secondary directions of the local image structure, which are then used to determine the direction of an anisotropic Gaussian kernel. Simultaneously, based on the pixel's position in the information entropy map... The corresponding value in the entropy is used to determine the size of the Gaussian kernel. A higher entropy value indicates a more complex local region, so a smaller Gaussian kernel is used to preserve details; a lower entropy value indicates a flatter local region, so a larger Gaussian kernel is used for stronger smoothing. In this way, each pixel in the image corresponds to a spatially variable filter kernel with different shapes and sizes.

[0039] For the input image During filtering, the output value of each pixel is obtained by placing the spatially variable filter kernel corresponding to that pixel onto it and comparing it with the input image. The baseband image is obtained by weighted summation of the corresponding neighborhood pixel values. After traversing all pixels, the resulting image is the baseband image. Input image With baseband image Subtracting each pixel individually yields the difference image, which represents the Laplacian detail component of the level. .

[0040] To prevent aliasing caused by downsampling, the baseband image can be pre-sampled. Perform a slight Gaussian smoothing, extracting pixels every other row and column, reducing both width and height to half their original values. The resulting image is the next level. l Input image plus 1 The process is repeated until the preset total number of decomposition levels is reached.

[0041] In an optional embodiment, a spatially variable filter kernel is constructed, comprising:

[0042] For the input image For each pixel, calculate the local gradient covariance matrix within a 5×5 neighborhood window of that pixel. ;

[0043] For the input image Perform eigenvalue decomposition to obtain eigenvectors. and ,Will and The anisotropic direction serves as the spatially variable filter kernel;

[0044] Calculate the local information entropy map within a 5×5 neighborhood window. The spatial support range of the spatially variable filter kernel is set according to its entropy value, and the calculation formula is as follows: ,in It is a constant. represents the scale parameter of the spatially variable filter kernel in the principal direction.

[0045] For a pixel located on a vertical edge in an image, the gradient within the pixel's 5×5 neighborhood is mainly concentrated in the horizontal direction. The calculated local gradient covariance matrix at this point... This will exhibit significant anisotropy; for example, the matrix elements may approximate as [150, 2; 2, 5]. After performing eigenvalue decomposition on the matrix elements, the resulting principal eigenvectors... The approximate direction is vertical, that is, along the edge, while the secondary eigenvector... The direction is approximately horizontal, that is, perpendicular to the edge. Two mutually orthogonal vectors. and The anisotropic direction of subsequent filtering operations is defined, allowing filtering to be performed along the edge rather than across the edge, thus preserving edge sharpness.

[0046] Furthermore, the scale or size of the spatially variable filter kernel is determined by the local information entropy map. Adaptive adjustment. In the aforementioned edge regions, pixel values ​​change drastically, and the calculated local information entropy map... It will be relatively high, for example, in a local entropy graph. The value is 4.5. Assume a constant. If we set it to 15, then the scale parameter calculated according to the formula... Approximately 2.73. In a flat region of the image, the pixel value changes very little within a 5×5 neighborhood, resulting in a calculated local entropy map. It will be very low, for example, in a local entropy graph. The value is 0.2. The calculated scale parameter at this time... The value is 12.5. Therefore, a small-scale spatially variable filter kernel is used in edge and textured areas to preserve details, while a large-scale spatially variable filter kernel is used in flat areas to effectively smooth noise.

[0047] S3, the detail remapping process is as follows: for each scale level, calculate the input image... Local phase consistency map Based on the local phase consistency diagram Phase consistency value pc and gradient consistency metric Detail of Laplace Remapping is performed to obtain the remapping detail components. If the phase consistency value pc and the gradient consistency metric If all values ​​are greater than their respective high thresholds, then an S-shaped gain function is used for enhancement; if the phase consistency value pc is greater than the high threshold of the phase consistency value pc, then gradient consistency measurement... Not greater than gradient consistency metric If the low threshold is not greater than the high threshold of phase consistency value pc, then a logarithmic gain function is used for enhancement; if the phase consistency value pc is not greater than the high threshold of phase consistency value pc, then a soft threshold function is used for coefficient shrinkage.

[0048] For each scale level of the input image A set of multi-scale, multi-directional logarithmic Gabor wavelet filters is used to process the input image. Convolution is performed to calculate the local energy and local magnitude at each pixel. The ratio of these two values ​​is the phase consistency value pc for that pixel. All phase consistency values ​​pc constitute a phase consistency map. .

[0049] The high threshold for phase consistency is set at 0.6, and the high threshold for gradient consistency is set at 0.8. If the phase consistency value (pc) of a pixel is greater than 0.6 and the gradient consistency measurement... A value greater than 0.8 indicates a well-defined, strong edge, corresponding to the Laplacian detail component. The coefficient values ​​at the corresponding positions are mapped using an S-shaped function. The S-shaped function provides a weak boost to smaller coefficient values, a significant amplification to intermediate coefficient values, and a reduced amplification to very large coefficient values, in order to enhance the contrast of imperfections while avoiding over-enhancement.

[0050] If the phase consistency value pc of a certain pixel is greater than 0.6, while the gradient consistency measure... A value less than 0.2 indicates a region that is structurally significant but has inconsistent gradient directions, such as a fine texture. This is indicated by the Laplacian detail component. The coefficients at the corresponding positions are mapped using a logarithmic function, which provides a moderate level of non-linear gain to moderately enhance texture details.

[0051] If the phase coherence value (pc) of a pixel is not greater than 0.6, it indicates that it is not perceptually important and is likely noise. Therefore, the Laplacian detail component... The coefficients at the corresponding positions are subjected to soft thresholding, that is, the absolute value of the coefficient is first subtracted from a preset threshold, and if the result is negative, it is set to zero; otherwise, the result and the original sign are retained. This can effectively suppress noise coefficients.

[0052] In an optional embodiment, if the phase consistency value pc and the gradient consistency metric If all values ​​exceed their respective high thresholds, then an S-shaped gain function is used for enhancement, including:

[0053] When phase consistency value Greater than the first preset threshold and gradient consistency metric When the value is greater than the second preset threshold, the S-shaped gain function is used to adjust the Laplace detail components. Enhancement is performed to obtain remapped detail components. , The gain function is .

[0054] In an image of a printed document, the phase consistency value of pixels along the outline. Typically, the gradient is very high, such as 0.86, and the gradient directions within the neighborhood of a pixel are highly consistent; this is a measure of gradient consistency. It is also very high, such as 0.93. The simultaneous satisfaction of both conditions indicates that the pixel is located on a significant structural feature. Assume the Laplacian detail component of the pixel... A value of 18 indicates a relatively strong edge signal.

[0055] At this point, the sigmoid gain function will be used to calculate the enhancement gain. The calculation process for the gain function g is as follows: taking the Laplace detail components... The absolute value is approximately 0.018. The gain function g is approximately 1.96. Remapping detail components. Equals the gain function g multiplied by the original Laplace detail component The value is approximately 35.28. In this way, the details of such strong structural edges are significantly and non-linearly enhanced, making them more prominent and clear in the image. At the same time, the function tends to saturate the gain of excessively strong detail signals, thus avoiding over-enhancement.

[0056] In an optional embodiment, if the phase consistency value pc is greater than a high threshold of the phase consistency value pc, the gradient consistency measure... Not greater than gradient consistency metric For low threshold values, a logarithmic gain function is used for enhancement, including:

[0057] When phase consistency value Greater than the first preset threshold and gradient consistency metric When the value is not greater than the second preset threshold, a logarithmic gain function is used to evaluate the Laplace detail components. Enhancement is performed to obtain remapped detail components. ,in It is a symbolic function.

[0058] The image contains rich details in certain areas, therefore the phase consistency value The gradient consistency metric is relatively high, for example, 0.83. However, due to the complexity of the texture, the gradient directions within the neighborhood vary widely, resulting in a lack of high consistency. A lower value, such as 0.65, indicates that the region is rich in detail but not a simple straight line or curved edge. This is assuming the Laplacian detail component of a pixel... A value of -6 represents a medium level of texture detail.

[0059] Based on the judgment conditions, a logarithmic gain function is selected for processing. Remapping detail components. It is approximately -12.8. This shows the original Laplacian detail component, which was -6. It has been enhanced to -12.8. The logarithmic gain function has a greater boosting effect on details with smaller values, while the boosting effect on details with larger values ​​is weakened. It is very suitable for enhancing the visibility of textures, so that the texture's sense of layering and richness are improved, without producing an overly sharp effect as when dealing with strong edges.

[0060] In an optional embodiment, if the phase consistency value pc is not greater than the high threshold of the phase consistency value pc, then a soft thresholding function is used for coefficient shrinkage, including:

[0061] When phase consistency value When the value is not greater than the first preset threshold, a soft thresholding function is used to evaluate the Laplace detail components. By performing coefficient shrinkage, the remapped detail components are obtained. ,in The shrinkage threshold is determined based on the local standard deviation of the image at the current scale level.

[0062] In cases such as out-of-focus situations, there is no clear structural information, therefore the phase consistency value... The value will be lower than the first preset threshold of 0.8, for example, a measured value of 0.4. This indicates that the signal in the region is likely dominated by noise or has very weak texture. A shrinkage threshold is determined based on the local standard deviation of the region in the image. For example, if the local standard deviation of the image is 3.5, a shrinkage threshold can be set. It is 3.5.

[0063] Laplacian detail component for each pixel in the region Apply a soft thresholding function. Assume a Laplacian detail component for a pixel. A value of 2.5 is likely a noise signal. The absolute value of 2.5 is less than the contraction threshold. That is, 3.5, therefore minus The result is negative. Take the maximum value between the result and 0, which is 0. Remap the detail components. A value of 0 indicates that noise has been completely filtered out. Assume the Laplacian detail component of another pixel. A value of -5 might indicate a minor detail that needs to be preserved. An absolute value of 5 is greater than the shrinkage threshold of 3.5. minus The result is 1.5. Taking the largest value between 1.5 and 0, which is 1.5, and then multiplying by the Laplace detail component... The sign is -1, which yields the remapped detail components. The value is -1.5. This indicates that details are preserved, but the amplitude is reduced. This helps to suppress noise in smooth areas while retaining true details beyond the noise level.

[0064] S4, the image reconstruction process is as follows: the input image at the coarsest scale level obtained from multi-scale decomposition is remapped with the detail components after remapping at all scale levels. The image is enhanced by upsampling and accumulating the data step by step, from coarse to fine.

[0065] The last baseband image obtained from the decomposition, which is the input image at the coarsest scale level, is used as the initial image for reconstruction. The detail components after the coarsest scale remapping are then used. The remapped detail components Add the result to the initial image. Upsample the result so that its size is remapped to the detail component of the next higher level (i.e., the second coarser level). With the same dimensions, upsampling can be achieved by inserting zero-value rows and columns and performing interpolation filtering. The upsampling result is then remapped to the detail components of the second-coarsening level. Add. Repeat the above sampling and accumulation process, progressively from coarse to fine, until all levels of remapping detail components are completed. Once all additions are complete, the resulting image is the enhanced image of the printing plate.

[0066] In an optional embodiment, the input image at the coarsest scale level obtained from multi-scale decomposition is remapped to the detail components after remapping across all scale levels. Upsampling and accumulation are performed step by step from coarse to fine to obtain the enhanced image, including:

[0067] Assume multi-scale decomposition This process involves several levels to obtain the input image at the coarsest scale level. And detail components after remapping all levels ;

[0068] Let the initial layer of the reconstructed image ;

[0069] According to Upsampling and accumulation are performed sequentially up to 0, and the calculation method is as follows:

[0070] ,in For bilinear interpolation upsampling operation, Enlarged to the size of the remapped detail component Same size, In the process of image reconstruction, the first l Reconstructed images at multiple levels;

[0071] The final image obtained This refers to image enhancement.

[0072] The reconstruction process is a coarse-to-fine image synthesis workflow. Assume a 1024×1024 image is decomposed into three levels, where L equals 3. This results in a coarsest input image of size 128×128. and three detail component layers that have undergone detail remapping. Dimensions: 1024×1024 Size 512×512 and The dimensions are 256×256. The first step in reconstruction is to obtain the initial layer image of the reconstructed image. Input image set as the coarsest layer At this time, the initial layer image It is a 128×128 image that contains the most basic outline and color information of the image.

[0073] Next, from Reconstruction begins with a value equal to 1, and intermediate layer images are calculated. The initial layer image of the reconstructed image That is, a 128×128 image is upsampled using bilinear interpolation and enlarged to the same size as the image. The same size 512x512. The upsampled image is compared with the 512x512 detail component. By adding pixels one by one, the reconstructed intermediate layer image is obtained. Intermediate layer image In the initial layer image The image is then infused with enhanced details at a medium scale. The final image is calculated. , intermediate layer image That is, the 512×512 image is upsampled to 1024×1024, and then compared with the finest detail component of size 1024×1024. Add them together. The final image obtained is obtained. It is a 1024×1024 enhanced image that incorporates details from all scale levels that have been selectively enhanced or suppressed.

[0074] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. An image enhancement method for inspecting printing plates in a printing press, characterized in that, Includes the following steps: S1, Obtain the original digital image of the printing plate to be inspected; S2, performs multi-scale decomposition on the original digital image: for any scale level from fine to coarse. l Input image Processing: Calculate the input image Local gradient covariance matrix Local information entropy diagram And gradient consistency metric Construct a spatially variable filter kernel and use it to filter the input image. Filtering is performed to obtain the baseband image. Input image With baseband image The difference is used as the scale level. l Laplace detail ; for baseband images Downsampling is performed, and the resulting image is used as the input for the next scale level. The gradient consistency measure is as follows: within a local window surrounding each pixel, calculate the magnitude of the sum of the gradient vectors of all pixels within the window, and then divide it by the sum of the magnitudes of the gradient vectors of all pixels within the window. The ratio is the gradient consistency measure of the pixel, with a value range between 0 and 1. S3, Detail remapping process: For each scale level, calculate the input image... Local phase consistency map Based on the local phase consistency diagram Phase consistency value pc and gradient consistency metric Detail of Laplace Remapping is performed to obtain the remapping detail components. If the phase consistency value pc and the gradient consistency metric If all values ​​are greater than their respective high thresholds, then an S-shaped gain function is used for enhancement; if the phase consistency value pc is greater than the high threshold of the phase consistency value pc, then gradient consistency measurement... Not greater than gradient consistency metric If the threshold is low, a logarithmic gain function is used for enhancement; If the phase consistency value pc is not greater than the high threshold of the phase consistency value pc, then the soft threshold function is used to shrink the coefficients; S4, Image Reconstruction Process: The input image at the coarsest scale level obtained from multi-scale decomposition is remapped with the detail components after remapping at all scale levels. Upsampling and accumulation are performed step-by-step from coarse to fine to obtain the enhanced image: Let the multi-scale decomposition... This process involves several levels to obtain the input image at the coarsest scale level. And detail components after remapping all levels Let the initial layer of the reconstructed image be... According to from Upsampling and accumulation are performed sequentially up to 0, using the following formula: ,in For bilinear interpolation upsampling operation, Enlarged to the size of the remapped detail component Same size, In the process of image reconstruction, the first l Reconstructed images at multiple levels; the final image obtained. This refers to image enhancement.

2. The image enhancement method for inspecting printing plates for a printing press according to claim 1, characterized in that, In S2, anisotropy is determined by the local gradient covariance matrix. The eigenvectors determine the spatial support range, which is determined by the local information entropy map. The entropy value is adjusted in the opposite direction.

3. The image enhancement method for inspecting printing plates for a printing press according to claim 1, characterized in that, Constructing a spatially variable filter kernel includes: For the input image For each pixel, calculate the local gradient covariance matrix within a 5×5 neighborhood window of that pixel. ; For the input image Perform eigenvalue decomposition to obtain eigenvectors. and ,Will and The anisotropic direction serves as the spatially variable filter kernel; Calculate the local information entropy value within a 5×5 neighborhood window. The spatial support range of the spatially variable filter kernel is set according to its entropy value, and the calculation formula is as follows: ,in It is a constant. represents the scale parameter of the spatially variable filter kernel in the principal direction.

4. The image enhancement method for inspecting printing plates for a printing press according to claim 1, characterized in that, If the phase consistency value pc and the gradient consistency measure If all values ​​exceed their respective high thresholds, then an S-shaped gain function is used for enhancement, including: When phase consistency value Greater than the first preset threshold and gradient consistency metric When the value is greater than the second preset threshold, the S-shaped gain function is used to adjust the Laplace detail components. The pixel values ​​in the image are enhanced to obtain the remapped detail components. , The gain function The calculation method is as follows , It is the absolute value symbol.

5. The image enhancement method for inspecting printing plates for a printing press according to claim 1, characterized in that, If the phase consistency value pc is greater than the high threshold of the phase consistency value pc, then the gradient consistency measure... Not greater than gradient consistency metric For low threshold values, a logarithmic gain function is used for enhancement, including: When phase consistency value Greater than the first preset threshold and gradient consistency metric When the value is not greater than the second preset threshold, a logarithmic gain function is used to evaluate the Laplace detail components. The pixel values ​​in the image are enhanced to obtain the remapped detail components. ,in For symbolic functions, It is the absolute value symbol.

6. The image enhancement method for inspecting printing plates for a printing press according to claim 1, characterized in that, If the phase consistency value pc is not greater than the high threshold of the phase consistency value pc, then a soft thresholding function is used for coefficient shrinkage, including: When phase consistency value When the value is not greater than the first preset threshold, a soft thresholding function is used to evaluate the Laplace detail components. The pixel values ​​in the image are subjected to coefficient shrinkage to obtain the remapped detail components. ,in The shrinkage threshold is determined based on the local standard deviation of the image at the current scale level. To remap detail components, For Laplace detail weights, It is the absolute value symbol.

7. The image enhancement method for inspecting printing plates for a printing press according to any one of claims 1-6, characterized in that, In S2, for the input image For each pixel in the algorithm, the horizontal gradient gx and vertical gradient gy of the pixel are calculated using the Sobel operator. A 5×5 window is taken around the pixel, and the local gradient covariance matrix is ​​calculated using the gradient values ​​of all pixels within the window. .

8. The image enhancement method for inspecting printing plates for a printing press according to claim 7, characterized in that, Methods for obtaining the original digital image of the printing plate to be inspected include: acquiring the surface image of the printing plate using an industrial camera, and uniformly converting the acquired color or grayscale image into a grayscale image as the original digital image.

Images