A Method and System for Enhancing the Contrast of X-ray Images of Space High-temperature Materials

Abstract

The present invention discloses a method for enhancing the X-ray image contrast of a spatial high-temperature material. The method includes: receiving an X-ray image of a spatial high-temperature material; adding a mask to the X-ray image to obtain a region of interest image; removing the low-frequency components of the Fourier transform by performing Gaussian high-pass filtering on the region of interest image to obtain a high-frequency detail image; adding the high-frequency detail image to the region of interest image to form a region of interest image with enhanced edges; performing gray-scale weighted histogram equalization on the region of interest image with enhanced edges to obtain a contrast-enhanced image; performing second-order controllable kernel regression on the contrast-enhanced image to remove the block effect caused by contrast enhancement and the generated noise pollution, and obtaining a contrast-enhanced image with the block effect removed. The present invention improves the contrast and edge details and removes the block effect and noise pollution caused by contrast enhancement, making the growth process of the high-temperature material clearly visible and facilitating the analysis of the interface dynamics problem of material growth.

Description

Technical Field

[0001] The present invention relates to the field of image processing and space materials, and in particular to a method and system for enhancing the contrast of X-ray images of space high-temperature materials. Background Art

[0002] Space materials science has important economic, military and scientific significance for expanding the existing basic theories of materials science and guiding the synthesis of new ground-based materials. Real-time observation of experimental phenomena in space and ground-based materials science is crucial for the study of material growth mechanisms. To date, almost all types of new materials have been prepared in space, and space material experiments have developed from simple material addition to comprehensive applications such as simultaneous space preparation, online detection, real-time observation, and physical property determination.

[0003] The overall mission of my country's space station has planned a high-temperature material science experiment cabinet (high-temperature cabinet) for the research of space material science experiments. The high-temperature cabinet is designed with an X-ray transmission imaging real-time observation module, which can observe the formation and preparation process of high-temperature materials under microgravity conditions in real time online, and is mainly used for the analysis of solid-liquid interface changes and the study of interface dynamics. Since it is difficult to directly observe continuous growth units (interface structures such as atoms, molecules or aggregates) during crystal growth in experiments, it is often measured by comparing the relationship between interface movement speed and driving force function with theoretical models to obtain information related to crystal growth mechanism. This criterion is currently an effective way to study growth mechanism. Since space experiment opportunities are scarce and difficult to repeat, more than 95% of laboratories are ground-based experiments. In order to ensure the smooth development of space experiments, repeated tests are required on the ground. The space environment can provide microgravity environment and vacuum environment for high-temperature material science, resulting in different growth phenomena, but it will not affect the performance of image processing algorithms. Therefore, image algorithms are suitable for space and ground-based high-temperature material experiments.

[0004] The research on image contrast enhancement technology is mainly divided into two categories: spatial domain and frequency domain technology. Research in the spatial domain mainly focuses on the principle and implementation of histogram equalization, while research in the frequency domain mainly focuses on the application and implementation of frequency filtering. Summary of the invention

[0005] The purpose of the present invention is to overcome the defects of the prior art, and proposes a method for enhancing the contrast of X-ray images of high-temperature materials in space, and also proposes a system for enhancing the contrast of X-ray images of high-temperature materials in space.

[0006] In order to achieve the above object, the present invention proposes a method for enhancing the contrast of a space high temperature material X-ray image, the method comprising:

[0007] Receive X-ray images of high-temperature materials in space;

[0008] Add a mask to the X-ray image to obtain an image of the region of interest;

[0009] By performing Gaussian high-pass filtering on the image in the domain of interest, the low-frequency components of Fourier transform are removed to obtain a high-frequency detail image;

[0010] Adding the region of interest image to the high-frequency detail image to form the region of interest image with edge emphasis;

[0011] Perform grayscale weighted histogram equalization processing on the edge-emphasized domain of interest image to obtain a contrast-enhanced image;

[0012] The contrast enhanced image is subjected to second-order controllable kernel regression to remove the block effect and noise pollution caused by contrast enhancement, and obtain a contrast enhanced image without block effect.

[0013] As an improvement of the above method, before receiving the space high temperature material X-ray image, the method further includes: converting the captured video of the space high temperature material X-ray into an image.

[0014] As an improvement of the above method, the region of interest image f(x, y) satisfies the following formula:

[0015] f(x,y)=I(x,y)R(x,y)

[0016] Among them, I(x,y) is the X-ray image, R(x,y) is the established image mask, and (x,y) represents the pixel coordinates.

[0017] As an improvement of the above method, the method performs Gaussian high-pass filtering on the image in the domain of interest to remove the low-frequency components of Fourier transform to obtain a high-frequency detail image; specifically, the method includes:

[0018] Calculate the frequency domain representation of the spatial domain image through Fourier transform;

[0019] Performing Gaussian high-pass filtering on the image in the domain of interest using a Gaussian high-pass filter with a set cutoff frequency to obtain a frequency domain representation of the Gaussian high-pass filtered image that retains high-frequency details;

[0020] The Gaussian high-pass filter image represented in the frequency domain is inversely Fourier transformed to obtain the spatial domain representation and the high-frequency detail image.

[0021] As an improvement of the above method, the high-frequency detail image is added with the region of interest image to form the region of interest image with edge emphasis; specifically, the method includes:

[0022] The domain of interest image and the high-frequency detail image are multiplied by coefficients α and β respectively and then added to obtain an edge-emphasized image of interest, where 0<α≤1, β>0.

[0023] As an improvement of the above method, the grayscale weighted histogram equalization processing is performed on the edge-emphasized region of interest image to obtain a contrast-enhanced image; specifically comprising:

[0024] Step 1) calculating the grayscale cumulative distribution function of the edge-emphasized region of interest image, dividing the entire image into three levels: bright, medium or dark, and obtaining a divided image;

[0025] Step 2) Calculate the grayscale histogram of the block image, determine the number of peaks, and then calculate the cumulative distribution function H of each peak i (k), where i represents the i-th peak, n k is the number of gray values ​​k;

[0026]

[0027] Step 3) From the cumulative distribution function H i The cumulative distribution function H of the maximum peak selected in (k) max (k);

[0028] Step 4) Calculate the brightness level of the peak value of the i-th peak according to the following formula:

[0029]

[0030] Step 5) traverse the brightness level level of each peak, and for the peak with brightness level level 1, classify the brightness level of level = 1 into a new bright, medium or dark level; go to step 2); if the brightness level level of each peak is 0, use the bright, medium and dark levels of the entire image, and go to step 6);

[0031] Step 6) performing histogram equalization on different areas of divided brightness levels;

[0032] Step 7) multiply each histogram-equalized brightness area by a weight factor to obtain a grayscale weighted histogram equalization image.

[0033] As an improvement of the above method, the contrast enhanced image is subjected to second-order controllable kernel regression to remove the block effect and noise pollution caused by contrast enhancement, and obtain a contrast enhanced image with the block effect removed; specifically, the method includes:

[0034] The controllable kernel regression function is calculated according to the following formula:

[0035]

[0036] where det(·) represents the determinant of the matrix, C pis the covariance matrix based on local grayscale differences, h is the global smoothing parameter, u p is a constant set to 1, H p is a smoothing matrix, representing a pixel in the sliding window, x p is the neighboring pixel of pixel a in the sliding window, the superscript T indicates transposition, and satisfies the following formula:

[0037]

[0038]

[0039] Among them, Z x1 (·) and Z x2 (·) is the first-order derivative of the image along the horizontal and vertical directions, w q is the qth local analysis window, x j is the j-th pixel belonging to the local analysis window;

[0040] Substituting the high-frequency enhanced grayscale weighted histogram equalization image into the controllable kernel regression function, the image with the block effect removed and the noise pollution reduced is obtained, and the contrast enhanced image with the block effect removed is obtained.

[0041] A space high temperature material X-ray image contrast enhancement system, the system comprises: a receiving module, a mask adding module, a Gaussian high pass filtering module, an edge emphasis module, an equalization processing module and a controllable kernel regression module; wherein,

[0042] The receiving module is used to receive X-ray images of high-temperature materials in space;

[0043] The mask adding module is used to add a mask to the X-ray image to obtain an image of a region of interest;

[0044] The Gaussian high-pass filter module is used to remove the low-frequency components of Fourier transform by performing Gaussian high-pass filter processing on the image in the domain of interest, so as to obtain a high-frequency detail image;

[0045] The edge emphasis module is used to add the region of interest image to the high-frequency detail image to form an edge-emphasized region of interest image;

[0046] The equalization processing module is used to perform grayscale weighted histogram equalization processing on the edge-emphasized domain of interest image to obtain a contrast-enhanced image;

[0047] The controllable kernel regression module is used to perform second-order controllable kernel regression on the contrast enhanced image to remove the block effect and noise pollution caused by contrast enhancement, so as to obtain a contrast enhanced image with the block effect removed.

[0048] Compared with the prior art, the advantages of the present invention are:

[0049] 1. The method of the present invention can enhance the contrast and improve the edge details of bismuth ferrite X-ray images, making the analysis of bismuth ferrite interface growth dynamics and the study of growth mechanism real and feasible;

[0050] 2. The method of the present invention optimizes the histogram equalization by dividing the X-ray image into brightness levels and assigning different weights to the brightness levels according to the calculation results, so that the grayscale coverage of the X-ray image is wider and the grayscale details are more obvious;

[0051] 3. The method of the present invention combines spatial histogram equalization, frequency domain filtering and weight optimization, enriches the grayscale information and texture details of X-ray images, amplifies weak image changes, and provides a real and reliable real-time online observation method for space material science experiments;

[0052] 4. The method of the present invention removes the block effect and reduces the noise pollution to a great extent while retaining the contrast enhancement through the controllable kernel regression of the image;

[0053] 5. The method and system of the present invention can overcome the problem that the images directly acquired by the X-ray transmission imaging system have low contrast and blurred edge details, making it difficult to observe the growth process of high-temperature materials in real time online. This method is used to enhance the contrast and improve the edge details and remove the block effect and noise pollution caused by the contrast enhancement, so that the growth process of high-temperature materials can be clearly seen, which is convenient for analyzing the interface dynamics of material growth. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 This is a flow chart of a method for enhancing contrast of a space high temperature material X-ray image according to Embodiment 1 of the present invention;

[0055] Figure 2 is a technical schematic diagram of Embodiment 1 of the present invention;

[0056] Figure 3 is an experimental system diagram of Example 1 of the present invention;

[0057] Figure 4 is the original image directly obtained in Example 1 of the present invention;

[0058] Figure 5 yes Figure 4 Grayscale histogram of ;

[0059] Figure 6 yes Figure 4 The image of the area of ​​interest;

[0060] Figure 7 yes Figure 6 Grayscale histogram of ;

[0061] Figure 8 is a set of high-frequency detail images of Example 1 of the present invention;

[0062] Fig. 9 is a set of edge-emphasized images according to Embodiment 1 of the present invention;

[0063] Fig.10 is the output image after grayscale weighted histogram equalization of embodiment 1 of the present invention;

[0064] Fig.11 yes Fig.10 Enhanced image of

[0065] Fig.12 yes Fig.11 Grayscale histogram of ;

[0066] Fig.13 yes Fig.11 The output image is obtained by removing blocking artifacts and noise. DETAILED DESCRIPTION

[0067] The method of the present invention can enhance the contrast and improve the edge details of the bismuth ferrite growth image obtained by the X-ray imaging system. The method first converts the video shot by the X-ray imaging system into an image frame; adds a mask to the X-ray image to eliminate the influence of the invalid background on the image enhancement, and obtains the domain of interest of the image; performs Gaussian high-pass filtering on the image of the domain of interest to filter out low-frequency components and retain high-frequency details; multiplies the high-pass filtered image by a coefficient and adds the image of the domain of interest by a coefficient to obtain an edge-enhanced enhanced image; performs grayscale weighted histogram equalization on the edge-enhanced image to obtain a contrast-enhanced and edge-enhanced X-ray image for real-time online observation and analysis of solid-liquid interface dynamics in material growth experiments.

[0068] The technical solution of the present invention is described in detail below with reference to the accompanying drawings and embodiments.

[0069] Example 1

[0070] like Figure 1 As shown, embodiment 1 of the present invention proposes a method for enhancing the contrast of X-ray images of high-temperature materials in space; Figure 2 This is a technical schematic diagram of an example of this method. The specific steps of the method are as follows:

[0071] Step 1) receiving an X-ray image of a high-temperature material in space;

[0072] Step 2) adding a mask to the X-ray image to obtain an image of a region of interest;

[0073] Step 3) performing Gaussian high-pass filtering on the image in the domain of interest to remove the low-frequency components of the Fourier transform and obtain a high-frequency detail image;

[0074] Step 4) adding the region of interest image to the high-frequency detail image to form an edge-emphasized region of interest image;

[0075] Step 5) performing grayscale weighted histogram equalization processing on the edge-emphasized region of interest image to obtain a contrast-enhanced image;

[0076] Step 6) Performing a second-order controllable kernel regression on the contrast-enhanced image to remove the block effect and noise pollution caused by the contrast enhancement, and obtaining a contrast-enhanced image with the block effect removed.

[0077] Figure 3 The system diagram is an example of a method and system for contrast enhancement of space high-temperature material X-ray images based on controllable kernel regression.

[0078] For step 1), press Figure 3 The experiment was conducted using the experimental system shown in the figure. The distance between the platinum crucible and the detector was 30 mm. The distance between the two was strictly controlled. A distance that was too close would damage the X-ray detector, and a distance that was too far would affect the quality of X-ray imaging. The platinum crucible was heated by a DC power supply to gradually melt the bismuth ferrite material sample in the crucible. After the bismuth ferrite was completely melted, it was slowly cooled to gradually solidify. The X-ray transmission imaging system recorded the entire experimental process in the form of a video. The X-ray transmission imaging video was converted into an X-ray image by writing a program. The final original image is shown in the figure below. Figure 4 As shown in the figure, it contains a lot of invalid background information, and the effective image content has low contrast and unclear edge details. Figure 4 The grayscale histogram and Figure 5 It can be seen that the grayscale distribution of the original X-ray image is very uneven.

[0079] For step 2), a mask R(x,y) is added to the original X-ray image I(x,y) to obtain the image of interest f(x,y). The pixel grayscale of the image content that you want to retain remains unchanged, and the pixel grayscale of the image content that you do not want to retain is set to zero, so that too much invalid image content can be effectively avoided. For the processed X-ray image, the area to be processed is outlined in the form of a box, circle, ellipse, irregular polygon, etc., that is, the domain of interest. The domain of interest is an image area selected from the original X-ray image. This area is the focus of processing, and at the same time avoids the influence of invalid images. This area is the focus of image analysis. Extracting the domain of interest can reduce processing time and increase processing accuracy.

[0080] by Figure 4Taking the image shown in the figure as an example, the specific operation is as follows: locate the center of the circular platinum crucible and create a circular image mask with a radius of 45 pixels. The region of interest here is a circle, but it is not limited to a circle. This mask can keep the image content inside the circular crucible unchanged, and all the pixel grayscales outside the crucible are set to zero, minimizing the impact of invalid background on image contrast enhancement, and obtaining the following Figure 6 The grayscale histogram of the image of the domain of interest shown in Figure 7 As shown, it can be seen that the influence of invalid image information is effectively avoided, but the phenomenon of uneven grayscale distribution is not changed.

[0081] f(x,y)=I(x,y)R(x,y)

[0082] For step 3), ideal high-pass filter, Butterworth high-pass filter and Gaussian high-pass filter can be selected. The result obtained by Gaussian filter is smoother than the filtering effect of the first two filters. The zero-frequency component of Fourier transform is removed from the X-ray image of the domain of interest by Gaussian high-pass filtering, and the average intensity of the background is reduced to nearly black, retaining high-frequency details. Specifically including:

[0083] The representation of the spatial domain image in the frequency domain is calculated by DFT:

[0084]

[0085] Use the intercept distance origin as D 0 A Gaussian high-pass filter with a value of =50 is used to perform Gaussian high-pass filtering on the image in the domain of interest to obtain a Gaussian high-pass filtered image with clear edges and less distortion.

[0086]

[0087] G(u,v)=H hp (u,v)F(u,v)

[0088] Calculate the inverse Fourier transform to obtain the spatial domain representation g(x,y) of the frequency domain image G(u,v).

[0089]

[0090] Use different cutoff frequencies D 0 Get different Gaussian high-pass filtered images, such as Figure 8 shown.

[0091] For step 4), the high-frequency detail image after Gaussian high-pass filtering is multiplied by the coefficient β and the X-ray image of the region of interest is multiplied by the coefficient α to obtain the edge-emphasized X-ray image of the region of interest. Specifically:

[0092] I(x,y)=αf(x,y)+βg(x,y)

[0093] 0<α≤1, β>0, for this embodiment 1, α=0.8, β=1, we get Fig. 9 Edge emphasized image shown.

[0094] For step 5), calculate the grayscale cumulative distribution function of the entire image and divide the entire image into different levels. Fig. 9 The grayscale of the image shown is concentrated in two grayscale ranges. One grayscale range is the area where the black background is concentrated, and the other grayscale range is the area where the bismuth ferrite in the platinum crucible is located. There are two peaks.

[0095] Calculate the grayscale histogram of the block image, determine the number of peaks, and then calculate the cumulative distribution function of each peak:

[0096]

[0097] Calculate the cumulative distribution function of the maximum peak:

[0098] H max (k) = max(H i (k))

[0099] The peak value can be judged as a brightness level:

[0100]

[0101] The image is divided into two brightness levels, and each brightness level is divided into three grayscale ranges: bright, medium, and dark. Then histogram equalization is used to process the divided brightness areas separately, so that pixels with similar grayscales can better enhance the contrast. k Different regions of the image are histogram equalized separately.

[0102] Variance is used to measure the degree of deviation between a random variable and its mean value, and can be used as an indicator of the shape of a histogram. When the grayscale in the region is evenly distributed near the brightness level, the variance is small; when the grayscale in the region is concentrated on both sides, the variance is large. Use variance as a function of weighted calculation.

[0103]

[0104] Among them, σ 2 It represents the variance, N is the total number of pixels in the block image, n j The gray level is y j The number of points, m i is the average gray value of the area, and the final weighting factor is calculated as:

[0105]

[0106] Each histogram equalized brightness area is multiplied by the weight factor, and finally a grayscale weighted histogram equalization image is obtained:

[0107]

[0108] Finally, we get Fig.10 The output image after grayscale weighted histogram equalization is shown in Fig.11 Displayed Fig.10 After processing, the enhanced image Fig.12 It is the corresponding grayscale histogram, from which we can see the effective coverage of the grayscale over the entire grayscale range.

[0109] For step 6), the contrast-enhanced image obtained after grayscale weighted histogram equalization is processed by controllable kernel regression to obtain an image with block effects and noise pollution removed:

[0110]

[0111] For this embodiment 1, h = 0.5, u = 1, where det(·) represents the determinant of the matrix, C p is the covariance matrix based on local grayscale differences, h is the global smoothing parameter, u p is a constant set to 1, H p is a smoothing matrix, representing a pixel in the sliding window, x p is the neighboring pixel of pixel a in the sliding window, the superscript T indicates transposition, and satisfies the following formula:

[0112]

[0113]

[0114] Among them, Z x1 (·) and Z x2 (·) is the first-order derivative of the image along the horizontal and vertical directions, w q is the qth local analysis window, x j is the j-th pixel belonging to the local analysis window;

[0115] Substitute the image after high-frequency enhanced grayscale weighted histogram equalization into the controllable kernel regression function to obtain an image with removed block effects and reduced noise pollution, such as Fig.13 shown.

[0116] Example 2

[0117] Embodiment 2 of the present invention proposes a space high temperature material X-ray image contrast enhancement system, which is implemented based on the method of embodiment 1. The system includes: a receiving module, a mask adding module, a high-frequency detail image acquisition module, an edge emphasis module, an equalization processing module and a controllable kernel regression module; wherein,

[0118] The receiving module is used to receive X-ray images of high-temperature materials in space;

[0119] The mask adding module is used to add a mask to the X-ray image to obtain an image of a region of interest;

[0120] The high-frequency detail image acquisition module is used to obtain a high-frequency detail image by performing Gaussian high-pass filtering on the image in the domain of interest to remove the low-frequency components of Fourier transform;

[0121] The edge emphasis module is used to add the region of interest image to the high-frequency detail image to form an edge-emphasized region of interest image;

[0122] The equalization processing module is used to perform grayscale weighted histogram equalization processing on the edge-emphasized domain of interest image to obtain a contrast-enhanced image;

[0123] The controllable kernel regression module is used to perform second-order controllable kernel regression on the contrast enhanced image to remove the block effect and noise pollution caused by contrast enhancement, so as to obtain a contrast enhanced image with the block effect removed.

[0124] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention is described in detail with reference to the embodiments, it should be understood by those skilled in the art that any modification or equivalent replacement of the technical solutions of the present invention does not depart from the spirit and scope of the technical solutions of the present invention and should be included in the scope of the claims of the present invention.

Claims

1. A method for enhancing the X-ray image contrast of a space high-temperature material, the method comprises: Receiving an X-ray image of a space high-temperature material; Adding a mask to the X-ray image to obtain a region of interest image; Removing the low-frequency components of the Fourier transform by performing Gaussian high-pass filtering on the region of interest image to obtain a high-frequency detail image; Adding the high-frequency detail image to the region of interest image to form a region of interest image with edge emphasis; Performing gray-weighted histogram equalization on the region of interest image with edge emphasis to obtain a contrast-enhanced image; Performing second-order controllable kernel regression on the contrast-enhanced image to remove the blocking effect caused by contrast enhancement and the generated noise pollution, and obtaining a contrast-enhanced image with the blocking effect removed; The performing second-order controllable kernel regression on the contrast-enhanced image to remove the blocking effect caused by contrast enhancement and the generated noise pollution, and obtaining a contrast-enhanced image with the blocking effect removed; specifically includes: Calculating a controllable kernel regression function according to the following formula: where det(·) represents taking the determinant of a matrix, C p is the covariance matrix based on local gray - level differences, h is the global smoothing parameter, u p is a constant set to 1, H p is the smoothing matrix, representing a pixel in the sliding window, x p is the adjacent pixel of pixel a in the sliding window, the superscript T represents transpose, and satisfies the following formula: Among them, Z x1 (·) and Z x2 (·) are the first-order derivatives of the image along the horizontal and vertical directions, w q is the q-th local analysis window, and x j is the j-th pixel belonging to the local analysis window; Substituting the image after high-frequency enhancement gray-weighted histogram equalization into the controllable kernel regression function to obtain an image with the blocking effect removed and noise pollution weakened, and obtaining a contrast-enhanced image with the blocking effect removed.

2. The method for enhancing the X-ray image contrast of a space high-temperature material according to claim 1, wherein, Before receiving the X-ray image of the space high-temperature material, it further includes: converting the video of the X-ray of the space high-temperature material taken into an image.

3. The method for enhancing the X-ray image contrast of a space high-temperature material according to claim 1 or 2, wherein, The region of interest image f(x, y) satisfies the following formula: f(x, y) = I(x, y)R(x, y) wherein, I(x, y) is the X-ray image, R(x, y) is the established image mask, and (x, y) represents pixel coordinates.

4. The method for enhancing the X-ray image contrast of a space high-temperature material according to claim 3, wherein, The removing the low-frequency components of the Fourier transform by performing Gaussian high-pass filtering on the region of interest image to obtain a high-frequency detail image; specifically includes: Calculating the frequency-domain representation of the spatial-domain image through Fourier transform; Performing Gaussian high-pass filtering on the region of interest image using a Gaussian high-pass filter with a set cut-off frequency to obtain the frequency-domain representation of the Gaussian high-pass filtered image with high-frequency details retained; Performing inverse Fourier transform on the frequency-domain representation of the Gaussian high-pass filtered image to obtain the spatial-domain representation, and obtaining a high-frequency detail image.

5. The method for enhancing the X-ray image contrast of a space high-temperature material according to claim 4, wherein, The adding the high-frequency detail image to the region of interest image to form a region of interest image with edge emphasis; specifically includes: Multiplying the region of interest image and the high-frequency detail image by coefficient α and coefficient β respectively and then adding them to obtain an edge-emphasized interest image, where 0 < α ≤ 1 and β > 0.

6. The method for enhancing the X-ray image contrast of a space high-temperature material according to claim 5, wherein, The performing gray-weighted histogram equalization on the region of interest image with edge emphasis to obtain a contrast-enhanced image; specifically includes: Step 1) Calculate the gray cumulative distribution function of the edge-emphasized region of interest image, divide the entire image into three levels: bright, medium, or dark, and obtain the segmented image; Step 2) Calculate the grayscale histogram of the segmented image, determine the number of peaks, and then calculate the cumulative distribution function H i (k), where i represents the i-th peak and n k is the number of pixels with grayscale value k; Step 3) Select the cumulative distribution function H i (k) with the maximum peak from the cumulative distribution function H max (k); Step 4) Calculate the brightness level level of the peak value of the i-th peak according to the following formula: Step 5) Traverse the brightness level level of each peak. For the peak with the brightness level lecel being 1, divide the brightness level with level = 1 into a new bright, medium, or dark level; go to Step 2); if the brightness level level of each peak is 0, then adopt the bright, medium, and dark levels of the entire image and go to Step 6); Step 6) Perform histogram equalization on different regions divided by the brightness level respectively; Step 7) Multiply each brightness region after histogram equalization by a weight factor to obtain a gray-scale weighted histogram equalized image.

7. A system for the method of enhancing the contrast of X-ray images of spatial high-temperature materials according to claim 1, characterized in that the system includes: a receiving module, a mask adding module, a Gaussian high-pass filtering module, an edge emphasizing module, an equalization processing module, and a controllable kernel regression module; wherein, the receiving module is used to receive X-ray images of spatial high-temperature materials; the mask adding module is used to add a mask to the X-ray image to obtain a region of interest image; the Gaussian high-pass filtering module is used to remove the low-frequency components of the Fourier transform by performing Gaussian high-pass filtering on the region of interest image to obtain a high-frequency detail image; the edge emphasizing module is used to add the region of interest image to the high-frequency detail image to form an edge-emphasized region of interest image; the equalization processing module is used to perform gray-scale weighted histogram equalization processing on the edge-emphasized region of interest image to obtain a contrast-enhanced image; the controllable kernel regression module is used to perform second-order controllable kernel regression on the contrast-enhanced image to remove the block effect caused by contrast enhancement and the generated noise pollution, and obtain a contrast-enhanced image with the block effect removed.

Images