X-ray detection image enhancement method and system

Abstract

The invention discloses an X-ray detection image enhancement method and system, and the method comprises the following image enhancement steps: carrying out the gray preprocessing of an image, carrying out image enhancement preprocessing, carrying out homomorphic filtering enhancement processing, carrying out USM sharpening processing, taking a filtered image Ifiltering as an input, carrying out the processing of an input image through Gaussian blur, generating an enhanced image, and finally carrying out the tone adjustment. And carrying out color gradation adjustment on the enhanced image Ishadow to obtain a final target image Itarget. According to the method, firstly, multiple parameters such as exposure, brightness and contrast are adjusted for the original image, so that the overall quality of the image can be improved more comprehensively. Compared with the prior art, the multi-dimensional adjustment provides greater flexibility, so that the image can reach the optimal basic state before entering the subsequent processing. And a foundation can be laid for subsequent image post-processing, image analysis, image classification and image library and map establishment.

Description

Technical Field

[0001] The present invention relates to the technical field of image post-processing, and in particular to the technical field of post-processing methods based on X-ray detection images, and specifically to an X-ray detection image enhancement method and system. Background Art

[0002] X-ray inspection image enhancement processing is an important research direction in the fields of medical imaging, material science and non-destructive testing. With the continuous development of X-ray imaging technology, the improvement of image quality plays a vital role in the accuracy of diagnosis and the effectiveness of detection. X-ray images often face problems such as low contrast, noise interference and lack of details, which directly affect the detection of material defects by non-destructive testing engineers. Therefore, image enhancement can not only improve the visualization of images, but also assist clinical decision-making and technical evaluation.

[0003] At present, X-ray image enhancement technologies mainly include histogram equalization, adaptive histogram equalization (CLAHE), filtering technology (such as high-pass filtering, denoising) and deep learning-based methods. These technologies have their own advantages and disadvantages. Histogram equalization can effectively improve the overall contrast, but may cause loss of image details; CLAHE can improve the contrast in local areas, but may cause artifacts when processing high-noise images. While filtering technology removes noise, it may also blur the edge details of the image. Although the enhancement method based on deep learning can achieve good results, it usually requires a large amount of labeled data for training, and the model is highly complex, making it difficult to apply in real time.

[0004] Although there are many image enhancement technologies, they still have some problems. First, many traditional methods have poor adaptability when processing different types of images and cannot achieve universal enhancement. Second, new noise or artifacts may be introduced during the image enhancement process, resulting in distortion of information. In addition, computational efficiency is also an important issue, especially in the implementation scenario of non-destructive testing, where complex enhancement algorithms are difficult to meet the requirements when real-time requirements are high. Therefore, future research needs to find a balance between improving image quality and maintaining details, while exploring more efficient and intelligent image enhancement technologies to better serve practical applications. Summary of the invention

[0005] In order to solve the enhancement problem of X-ray detection image post-processing, the present application provides an X-ray detection image enhancement method and system, which aims to solve the enhancement problem of X-ray detection image post-processing and provide a new image enhancement solution. The enhanced image can lay the foundation for subsequent image post-processing, image analysis, image classification and gallery, and atlas establishment, especially for AI intelligent gallery establishment and atlas database establishment and analysis. It is of great significance, can solve the problem of machine recognition deviation, and improve the efficiency and accuracy of machine processing, which is subversive for the convenience brought to the power equipment detection image post-processing and analysis industry.

[0006] In order to achieve the above purpose, the technical solution adopted in this application is:

[0007] An X-ray detection image enhancement method includes the following image enhancement steps:

[0008] Step STP100, grayscale preprocessing of the image, respectively, the input original image I Original Adjust the exposure, brightness, contrast, curve contrast and gamma value parameters and obtain the grayscale image I gray ;

[0009] Step STP200, image enhancement preprocessing, grayscale image I gray The pre-processed image I is obtained by sequentially performing HDR-like processing, multiple exposure fusion processing and adaptive histogram equalization processing. pret ;

[0010] Step STP300, homomorphic filtering enhancement processing, the pre-processed image I pret The homomorphic filtering based on Gaussian high-pass filter deformation filtering and the homomorphic filtering based on high-frequency enhancement filtering are respectively performed to obtain the filtered image I filtering ;

[0011] Step STP400, USM sharpening processing, the filtered image I filtering As input, use Gaussian blur to process the input image, blurring the image I blurred It can be obtained by the following formula:

[0012] I blurred (x,y)=I filtering (x,y)*H(x,y)

[0013] Among them, H(x,y) is the Gaussian kernel; high-frequency sharpening is then performed to calculate the original image I filtering with I blurred The difference between the high frequency components is calculated by the following formula:

[0014] I high_freq (x,y)=Ifiltering (x,y)-I blurred (x,y)

[0015] Step STP500, generating an enhanced image, and removing the high frequency component I high_freq (x,y) and the original image I filtering Superposition is performed to obtain enhanced image I sharpened

[0016] I sharpened (x,y)=I filtering (x,y)+k·I high_freq (x,y)

[0017] Where k is the multiplication factor of the sharpening strength, k = 1.0-3.0;

[0018] Step STP600, tone adjustment, will enhance the image I sharpened Perform color scale adjustment to obtain the final target image I target :

[0019]

[0020] Among them, max , in min ,out max ,out min They are the upper limit of window width, the lower limit of window width, the upper limit of output, and the lower limit of output respectively.

[0021] In order to better achieve the final enhancement effect, preferably, the homomorphic filtering of the Gaussian high-pass and high-frequency enhancement filters in step STP300 is implemented by the following steps:

[0022] Step STP310, logarithmic transformation of the image, the pre-processed image I pret The multiplication model is converted to an additive model, which enhances high-frequency information and reduces low-frequency information. The formula for logarithmic transformation is as follows:

[0023] I p ′ ret (x,y)=log(I(x,y)+1)

[0024] Among them, I p ′ ret (x,y) is the preprocessed image after logarithmic transformation;

[0025] Step STP320, frequency domain transformation of the image, the pre-processed image I pret From the spatial domain to the frequency domain, the image is converted to the frequency domain using the fast Fourier transform FFT. The formula of FFT is:

[0026]

[0027] Among them, F(u,v) is the frequency domain image;

[0028] Step STP330, design a high-pass filter to filter out low-frequency components and retain high-frequency components. The frequency response is defined as:

[0029]

[0030] Among them, H(u,v) is the frequency response, D(u,v) is the frequency distance, and D0 is the cutoff frequency;

[0031] Step STP340, applying a high pass filter, applying the frequency domain image in step STP320 to the high pass filter in step STP330 to obtain a filtered frequency domain image G(u,v)

[0032] G(u,v)=F(u,v)·H(u,v)

[0033] Then use the inverse fast Fourier transform IFFT to convert the frequency domain image G(u,v) back to the spatial domain image g(x,y).

[0034] That is, g(x,y)=IFFT(G(u,v))

[0035] Then the obtained spatial domain image g(u,v) is exponentially transformed to restore the original highlight model filtered image I filtering (x,y)=e g(x,y) -1.

[0036] Preferably, the image enhancement preprocessing in step STP200 includes the following steps:

[0037] Step STP210, grayscale image I gray Use gradient exposure to create multiple images with different exposure registrations Using the weight function w k Synthesizing HDR Images I HDR (x,y)

[0038]

[0039] Among them, I k (x, y) is the pixel value of the kth exposure image at position (x, y); w k is a weight function used to adjust the weight of each image to the final synthesized HDR image I HDR The influence of (x,y);

[0040] Step STP220, synthesizing the HDR image I HDR(x, y) performs multi-exposure fusion to obtain the fused image I MEF (x,y):

[0041]

[0042] Among them, I MEF (x,y) is the pixel value of the fused image at position (x,y);

[0043] γ is a correction factor used to adjust the image brightness and satisfies 1>γ>0;

[0044] λ is a fixed coefficient and satisfies λ>0.

[0045] In order to better lay the foundation for subsequent homomorphic filtering, preferably, the following steps are also included: performing local contrast enhancement processing on the fused image processed by step STP220, and converting the fused image I MEF (x, y) is divided into multiple array blocks, and each array block is subjected to histogram equalization to equalize the image I CLAHE (x,y):

[0046]

[0047] Among them, I CLAHE (x,y) is the pixel value of the image at position (x,y) after CLAHE processing;

[0048] L max , L min They are the maximum and minimum pixels of the current array block respectively;

[0049] C is the contrast limiting factor.

[0050] The present invention also provides an X-ray detection image enhancement system, comprising a storage module for storing software programs, and a processing module for calling and executing the software programs summarized in the storage module to execute the X-ray detection image enhancement method described above.

[0051] Beneficial effects:

[0052] 1. The present invention first adjusts multiple parameters of the original image, such as exposure, brightness, contrast, etc., to improve the overall quality of the image more comprehensively. Compared with the prior art, this multi-dimensional adjustment provides greater flexibility, so that the image can reach the best basic state before entering the subsequent processing.

[0053] 2. The present invention can effectively expand the dynamic range of the image and retain more details through HDR-like processing and multiple exposure fusion. In traditional image enhancement technology, highlight and shadow areas are often not fully processed, but this step of the present method can significantly improve the detail performance and layering of the image.

[0054] 3. The present invention introduces adaptive histogram equalization (CLAHE) to enhance local areas of the image. Compared with global equalization, this local contrast enhancement method can more effectively enhance image details and avoid image distortion caused by over-enhancement.

[0055] 4. The implementation of homomorphic filtering in the present invention includes two methods based on Gaussian high-pass filter and high-frequency enhancement filter. Such a design can better highlight the edges and details in the image, and compared with a single filtering method, it can better adapt to the needs of different scenes and improve the clarity and readability of the final image.

[0056] 5. After the image is fused, further local contrast enhancement is performed to make the details of the image more prominent. By dividing the image into multiple small blocks and performing equalization, the information loss caused by global processing can be effectively avoided, thereby improving the visibility and aesthetics of the image. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without paying any creative labor.

[0058] Figure 1 It is a flow chart of the present invention.

[0059] Figure 2 This is the original image of GIS detected by X-ray.

[0060] Figure 3 Through the present invention Figure 2 Enhanced image after enhancement processing. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various different configurations.

[0062] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application for which protection is sought, but merely represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0063] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0064] In the description of this application, it should be noted that if the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside", etc. appear, the orientation or position relationship indicated is based on the orientation or position relationship shown in the drawings, or the orientation or position relationship in which the product of the application is usually placed when in use. It is only for the convenience of describing this application and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as a limitation on this application. In addition, if the terms "first", "second", etc. appear in the description of this application, they are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0065] In addition, if the terms "horizontal" or "vertical" appear in the description of this application, it does not mean that the components are required to be absolutely horizontal or suspended, but can be slightly tilted. For example, "horizontal" only means that its direction is more horizontal than "vertical", and does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0066] In the description of this application, it should also be noted that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal connection of two components. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0067] Embodiment 1:

[0068] This embodiment provides an X-ray detection image enhancement method. The method flow is shown in the attached specification. Figure 1 As shown, Figure 2 As the original image I Original Input, specifically including the following image enhancement steps:

[0069] Step STP100, grayscale preprocessing of the image, respectively, the input original image I Original Adjust the exposure, brightness, contrast, curve contrast and gamma value parameters and obtain the grayscale image I gray ;

[0070] In this embodiment, the specific implementation method and calculation formula for adjusting the exposure, brightness, contrast, curve contrast and gamma value parameters are implemented using the following content; for the convenience of description, in this step, since the images processed according to different parameters have input images and output images, in this step, the input image with any parameter adjustment is used as I Original (x, y) is used to express the output image. To express, They represent exposure, brightness, contrast, curve contrast and gamma respectively;

[0071] 1) Exposure adjustment formula:

[0072]

[0073] e is the exposure factor;

[0074] 2)Brightness adjustment formula:

[0075]

[0076] b is the brightness coefficient

[0077] 3) Contrast adjustment formula:

[0078] Principle formula:

[0079]

[0080] (c is the contrast coefficient; t is a preset constant, the default is 1)

[0081] Actual formula:

[0082]

[0083] 4) Curve contrast formula:

[0084]

[0085] c is the curve contrast coefficient

[0086] 5) Gamma value adjustment formula:

[0087]

[0088] Step STP200, image enhancement preprocessing, grayscale image I grayThe pre-processed image I is obtained by sequentially performing HDR-like processing, multiple exposure fusion processing and adaptive histogram equalization processing. pret ;

[0089] Step STP300, homomorphic filtering enhancement processing, the pre-processed image I pret The homomorphic filtering based on the Gaussian high-pass filter deformation filtering and the homomorphic filtering based on the high-frequency enhancement filtering are respectively performed to obtain the filtered image I filtering ;

[0090] Step STP400, USM sharpening processing, the filtered image I filtering As input, use Gaussian blur to process the input image, blurring the image I blurred It can be obtained by the following formula:

[0091] I blurred (x,y)=I filtering (x,y)*H(x,y)

[0092] Among them, H(x,y) is the Gaussian kernel; high-frequency sharpening is then performed to calculate the original image I filtering with I blurred The difference between the high frequency components is calculated by the following formula:

[0093] I high_freq (x,y)=I filtering (x,y)-I blurred (x,y)

[0094] Step STP500, generating an enhanced image, and removing the high frequency component I high_freq (x,y) and the original image I filtering Superposition is performed to obtain enhanced image I sharpened , see Figure 3 As shown;

[0095] I sharpened (x,y)=I filtering (x,y)+k·I high_freq (x,y)

[0096] Where k is the multiplication factor of the sharpening strength, k = 1.0-3.0;

[0097] Step STP600, tone adjustment, will enhance the image I sharpened Perform color scale adjustment to obtain the final target image I target :

[0098]

[0099] Among them, max , inmin ,out max ,out min They are the upper limit of window width, the lower limit of window width, the upper limit of output, and the lower limit of output respectively.

[0100] Embodiment 2:

[0101] In order to better achieve the final enhancement effect, this embodiment is based on the above embodiment 1 and the attached Figure 1-3 On the basis of further optimizing and refining some steps, the homomorphic filtering of Gaussian high-pass and high-frequency enhancement filters in step STP300 is realized by the following steps:

[0102] Step STP310, logarithmic transformation of the image, the pre-processed image I pret The multiplication model is converted to an additive model, which enhances high-frequency information and reduces low-frequency information. The formula for logarithmic transformation is as follows:

[0103] I p ′ ret (x,y)=log(I(x,y)+1)

[0104] Among them, I p ′ ret (x,y) is the preprocessed image after logarithmic transformation;

[0105] Step STP320, frequency domain transformation of the image, the pre-processed image I pret From the spatial domain to the frequency domain, the image is converted to the frequency domain using the fast Fourier transform FFT. The formula of FFT is:

[0106]

[0107] Among them, F(u,v) is the frequency domain image;

[0108] Step STP330, design a high-pass filter to filter out low-frequency components and retain high-frequency components. The frequency response is defined as:

[0109]

[0110] Among them, H(u,v) is the frequency response, D(u,v) is the frequency distance, and D0 is the cutoff frequency;

[0111] Step STP340, applying a high pass filter, applying the frequency domain image in step STP320 to the high pass filter in step STP330 to obtain a filtered frequency domain image G(u,v)

[0112] G(u,v)=F(u,v)·H(u,v)

[0113] Then use the inverse fast Fourier transform IFFT to convert the frequency domain image G(u,v) back to the spatial domain image g(x,y).

[0114] That is, g(x,y)=IFFT(G(u,v))

[0115] Then the obtained spatial domain image g(u,v) is exponentially transformed to restore the original highlight model filtered image I filtering (x,y)=e g(x,y) -1.

[0116] Furthermore, the image enhancement preprocessing in step STP200 includes the following steps:

[0117] Step STP210, grayscale image I gray Use gradient exposure to create multiple images with different exposure registrations Using the weight function w k Synthesizing HDR Images I HDR (x,y)

[0118]

[0119] Among them, I k (x, y) is the pixel value of the kth exposure image at position (x, y); w k is a weight function used to adjust the weight of each image to the final synthesized HDR image I HDR The influence of (x,y);

[0120] Step STP220, synthesizing the HDR image I HDR (x, y) performs multi-exposure fusion to obtain the fused image I MEF (x,y):

[0121]

[0122] Among them, I MEF (x,y) is the pixel value of the fused image at position (x,y);

[0123] γ is a correction factor used to adjust the image brightness and satisfies 1>γ>0;

[0124] λ is a fixed coefficient and satisfies λ>0.

[0125] In order to better lay the foundation for subsequent homomorphic filtering, preferably, the following steps are also included: performing local contrast enhancement processing on the fused image processed by step STP220, and converting the fused image I MEF(x, y) is divided into multiple array blocks, and each array block is subjected to histogram equalization to equalize the image I CLAHE (x,y):

[0126]

[0127] Among them, I CLAHE (x,y) is the pixel value of the image at position (x,y) after CLAHE processing;

[0128] L max , L min They are the maximum and minimum pixels of the current array block respectively;

[0129] C is the contrast limiting factor.

[0130] This embodiment introduces adaptive histogram equalization (CLAHE), which can enhance local areas of the image. Compared with global equalization, this local contrast enhancement method can more effectively enhance the details of the image and avoid image distortion caused by excessive enhancement. The implementation of homomorphic filtering includes two methods based on Gaussian high-pass filter and high-frequency enhancement filter. Such a design can better highlight the edges and details in the image, and compared with a single filtering method, it can better adapt to the needs of different scenes and improve the clarity and readability of the final image.

[0131] Embodiment 3:

[0132] The present embodiment provides an X-ray detection image enhancement system, including a storage module for storing software programs, and a processing module for calling and executing the software programs summarized in the storage module to perform an X-ray detection image enhancement method described above.

[0133] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for enhancing an X-ray detection image, characterized in that: The following image enhancement steps are included: Step STP100, grayscale preprocessing of the image, respectively, the input original image I Original Adjust the exposure, brightness, contrast, curve contrast and gamma value parameters and obtain the grayscale image I gray ; Step STP200, image enhancement preprocessing, grayscale image I gray The pre-processed image I is obtained by sequentially performing HDR-like processing, multiple exposure fusion processing and adaptive histogram equalization processing. pret ; Step STP300, homomorphic filtering enhancement processing, the pre-processed image I pret The homomorphic filtering based on Gaussian high-pass filter deformation filtering and the homomorphic filtering based on high-frequency enhancement filtering are respectively performed to obtain the filtered image I filtering ; Step STP400, USM sharpening processing, the filtered image I filtering As input, use Gaussian blur to process the input image, blurring the image I blurred It can be obtained by the following formula: I blurred (x,y)=I filtering (x,y)*H(x,y) Among them, H(x,y) is the Gaussian kernel; high-frequency sharpening is then performed to calculate the original image I filtering with I blurred The difference between the high frequency components is calculated by the following formula: I high_freq (x,y)=I filtering (x,y)-I blurred (x,y) Step STP500, generating an enhanced image, and removing the high frequency component I high_freq (x,y) and the original image I filtering Superposition is performed to obtain enhanced image I sharpened I sharpened (x,y)=I filtering (x,y)+k·I high_freq (x,y) Where k is the multiplication factor of the sharpening strength, k = 1.0-3.0; Step STP600, tone adjustment, will enhance the image I sharpened Perform color scale adjustment to obtain the final target image I target : Among them, max , in min ,out max ,out min They are the upper limit of window width, the lower limit of window width, the upper limit of output, and the lower limit of output respectively.

2. The X-ray detection image enhancement method according to claim 1, characterized in that: The homomorphic filtering of the Gaussian high-pass and high-frequency enhancement filters in step STP300 is achieved by the following steps: Step STP310, logarithmic transformation of the image, the pre-processed image I pret The multiplication model is converted to an additive model, which enhances high-frequency information and reduces low-frequency information. The formula for logarithmic transformation is as follows: I p ′ ret (x,y)=log(I(x,y)+1) Among them, I p ′ ret (x,y) is the preprocessed image after logarithmic transformation; Step STP320, frequency domain transformation of the image, the pre-processed image I pret From the spatial domain to the frequency domain, the image is converted to the frequency domain using the fast Fourier transform FFT. The formula of FFT is: Among them, F(u,v) is the frequency domain image; Step STP330, design a high-pass filter to filter out low-frequency components and retain high-frequency components. The frequency response is defined as: Among them, H(u,v) is the frequency response, D(u,v) is the frequency distance, and D0 is the cutoff frequency; Step STP340, applying a high pass filter, applying the frequency domain image in step STP320 to the high pass filter in step STP330 to obtain a filtered frequency domain image G(u,v) G(u,v)=F(u,v)·H(u,v) Then use the inverse fast Fourier transform IFFT to convert the frequency domain image G(u,v) back to the spatial domain image g(x,y). That is, g(x,y)=IFFT(G(u,v)) Then the obtained spatial domain image g(u,v) is exponentially transformed to restore the original highlight model filtered image I filtering (x,y)=e g(x,y) -1.

3. The X-ray detection image enhancement method according to claim 1, characterized in that: The image enhancement preprocessing in step STP200 includes the following steps: Step STP210, grayscale image I gray Use gradient exposure to create multiple images with different exposure registrations Using the weight function w k Synthesizing HDR Images I HDR (x,y) Among them, I k (x, y) is the pixel value of the kth exposure image at position (x, y); w k is a weight function used to adjust the weight of each image to the final synthesized HDR image I HDR The influence of (x,y); Step STP220, synthesizing the HDR image I HDR (x, y) performs multi-exposure fusion to obtain the fused image I MEF (x,y): Among them, I MEF (x,y) is the pixel value of the fused image at position (x,y); γ is a correction factor used to adjust the image brightness and satisfies 1>γ>0; λ is a fixed coefficient and satisfies λ>0.

4. The X-ray detection image enhancement method according to claim 3, characterized in that: The following steps are also included: The fused image processed in step STP220 is subjected to local contrast enhancement processing, and the fused image I MEF (x, y) is divided into multiple array blocks, and each array block is subjected to histogram equalization to equalize the image I CLAHE (x,y): Among them, I CLAHE (x,y) is the pixel value of the image at position (x,y) after CLAHE processing; L max , L min They are the maximum and minimum pixels of the current array block respectively; C is the contrast limiting factor.

5. An X-ray detection image enhancement system, characterized in that: It comprises a storage module for storing software programs, and a processing module for calling and executing the software programs summarized in the storage module to execute the methods described in claims 1-4.

Images