CN116993628B - A CT image enhancement system for tumor radiofrequency ablation guidance - Google Patents

Abstract

The invention relates to the field of image processing, and specifically relates to a CT image enhancement system for tumor radiofrequency ablation guidance. The system includes: collecting tumor radiofrequency ablation CT images; and according to each edge line in the neighborhood window of each central pixel point and The relationship between the central pixel points is used to obtain the central boundary line in the neighborhood window; based on the curvature radius vector of each central boundary line in the neighborhood window of each central pixel point and the corner point information on both sides, the suspected tumor line of each central boundary line is obtained Coefficient; the tumor emphasis coefficient of each central pixel is obtained based on the tumor line suspicion coefficient and gradient difference of each central boundary line in the neighborhood window of each central pixel; the tumor enhancement coefficient of each edge point in the edge contour binary map is obtained After the coefficients and the grayscale information of each pixel in the tumor radiofrequency ablation CT image are enhanced, the grayscale value of each pixel is enhanced to complete the intelligent enhancement of the tumor radiofrequency ablation CT image. It achieves targeted enhancement of tumors and improves algorithm accuracy.

Description

A CT image enhancement system for tumor radiofrequency ablation guidance

Technical field

The present application relates to the field of image processing, and specifically to a CT image enhancement system for tumor radiofrequency ablation guidance.

Background technique

With the development of imaging technology, in order to achieve accurate positioning and puncture of tumors in the body under CT guidance, compared with traditional image positioning methods, spiral CT technology has the characteristics of faster scanning speed and clearer images, which overcomes the Since the instability of the respiratory movement of the chest and abdomen makes it difficult to determine the location where the needle enters, the use of spiral CT technology is of great significance in the treatment of tumors in the body. The traditional local mean square error enhancement algorithm has poor processing effect on CT images. It cannot focus on enhancing the tumor area in the image and is not targeted.

In summary, the present invention proposes a CT image enhancement system for tumor radiofrequency ablation guidance, which enhances tumor radiofrequency ablation CT images and performs tumor feature analysis on tumor radiofrequency ablation CT images by improving the local mean square error enhancement algorithm. And obtain the difference between the tumor and surrounding blood vessels to extract the tumor edge line, calculate the enhancement control coefficient of the center pixel in each neighborhood window to affect the enhancement effect, and avoid the feature of the enhancement algorithm that does not enhance the tumor area , improving the accuracy of the enhancement effect.

Contents of the invention

In order to solve the above technical problems, the present invention provides a CT image enhancement system for tumor radiofrequency ablation guidance, which system includes:

Image acquisition module: collects tumor radiofrequency ablation CT images;

Image enhancement module: Perform edge detection on tumor radiofrequency ablation CT images to obtain an edge contour binary map. Taking each edge point in the edge contour binary map as the center pixel, obtain the neighborhood window of each center pixel;

According to the relationship between each edge line in the neighborhood window of each center pixel point and the center pixel point, the center boundary line in the neighborhood window is obtained;

A corner detection algorithm is used to extract the corner information on both sides of each central boundary line in the neighborhood window of each central pixel; each central boundary line is obtained based on the curvature of the boundary points on each central boundary line in the neighborhood window of each central pixel. A sequence of curvature radius vectors;

Based on the curvature radius vector sequence of each central boundary line in the neighborhood window of each central pixel and the corner point information on both sides of each central boundary line, the suspected tumor line of each central boundary line in the neighborhood window of each central pixel is obtained Coefficient, the central boundary line with the largest suspected tumor line coefficient in the neighborhood window of each central pixel is recorded as the final boundary line of the neighborhood window of each central pixel;

Calculate the gray gradient pattern on both sides of the final boundary line within the neighborhood window of each central pixel;

According to the gray gradient law on both sides of the final boundary line in the neighborhood window of each central pixel point, the gradient difference of the final boundary line in the neighborhood window of the central pixel point is obtained;

The tumor emphasis coefficient of each central pixel is obtained based on the tumor line suspicion coefficient and gradient difference of the final boundary line in the neighborhood window of each central pixel;

According to the tumor enhancement coefficient of each edge point in the edge contour binary map and the grayscale information of each pixel in the tumor radiofrequency ablation CT image, the grayscale value of each pixel after enhancement is obtained to complete the intelligent enhancement of the tumor radiofrequency ablation CT image.

Preferably, the specific steps of obtaining the central boundary line in the neighborhood window based on the relationship between each edge line in the neighborhood window of each central pixel point and the central pixel point are:

According to the edge line in the neighborhood window of each center pixel that passes through the center pixel and intersects with the neighborhood window to form a closed area, it is recorded as the central boundary line in the neighborhood window.

Preferably, the specific steps of using the corner detection algorithm to extract the corner information on both sides of each central boundary line in the neighborhood window of each central pixel are as follows:

According to each central pixel,

Obtain corner points on both sides of each central boundary line in the corresponding neighborhood window;

The number of corner points on the side with the most corner points on both sides is recorded as the first corner point number, and the number of corner points on the side with the least corner points on both sides is recorded as the second corner point number;

The ratio of the number of first corner points to the number of second corner points is recorded as the corner point information on both sides of each central boundary line in the neighborhood window of each central pixel point.

Preferably, the specific steps of obtaining the curvature radius vector sequence of each central boundary line based on the curvature of the boundary points on each central boundary line in the neighborhood window of each central pixel point are:

Calculate the curvature radius vector of each boundary point on each central boundary line of each central pixel neighborhood window,

Based on the boundary points on the central boundary line in the first neighborhood window from left to right and top to bottom as the initial starting point, arrange the boundary points on the central boundary line in the neighborhood window into a sequence to obtain the neighborhood window Curvature radius vector sequence of each center boundary line within.

Preferably, the curvature radius vector sequence of each central boundary line in the neighborhood window of each central pixel point and the corner point information on both sides of each central boundary line are obtained. The specific steps for tumor line suspicion coefficient are:

Based on the curvature radius vector sequence of each central boundary line in the neighborhood window of each central pixel,

Obtain the curvature radius variance of each vector module length in the curvature radius vector sequence of each central boundary line, and obtain the curvature radius direction variance of each vector direction in the curvature radius vector sequence of each central boundary line,

The tumor line suspicion coefficient of each central boundary line in the neighborhood window of each central pixel is proportional to the sum of the curvature radius variance and the curvature radius direction variance of the corresponding central boundary line, and is proportional to the corner points on both sides of the corresponding central boundary line. Information is directly proportional.

Preferably, the specific steps of recording the central boundary line with the largest tumor line suspicion coefficient in each central pixel's neighborhood window as the final boundary line of each central pixel's neighborhood window are:

Obtain the tumor line suspicion coefficient of each central boundary line within the neighborhood window of each central pixel,

The central boundary line corresponding to the maximum value of the tumor line suspicion coefficient in the neighborhood window is recorded as the final boundary line.

Preferably, the expression for calculating the gray gradient pattern on both sides of the final boundary line within the neighborhood window of each central pixel is:

In the formula, is the number of boundary points on the final boundary line within the neighborhood window of the center pixel,/> is the final boundary line within the neighborhood window of the center pixel/> The boundary point is the concave side area of the starting point and along the boundary point/> The number of pixels on the straight line in the direction of the curvature radius,/> is an exponential function with natural constants as the base,/> ,/> is the final boundary line within the neighborhood window of the center pixel/> The boundary point is the concave side area of the starting point and along the boundary point/> /> on the straight line where the direction of the curvature radius is located No., No./> The gray value of two adjacent pixels,/> It is the gray gradient law on the concave side of the final boundary line of the center pixel neighborhood window.

Preferably, the specific steps of obtaining the gradient difference of the final boundary line in the neighborhood window of the center pixel based on the gray gradient rules on both sides of the final boundary line in the neighborhood window of each central pixel are as follows:

Based on the neighborhood window of each central pixel,

The side with the largest grayscale gradient pattern on both sides of the corresponding final boundary line is recorded as the first gradient area.

The side with the smallest grayscale gradient pattern on both sides of the corresponding final boundary line is recorded as the second gradient area.

The ratio of the grayscale gradient rules of the first gradient area and the second gradient area is recorded as the gradient difference of the final boundary line in the neighborhood window of the central pixel point.

Preferably, the specific steps of obtaining the tumor emphasis coefficient of each central pixel based on the tumor line suspicion coefficient and the gradient difference of the final boundary line in the neighborhood window of each central pixel are as follows:

According to the final boundary line of the neighborhood window of each central pixel point,

Obtain the grayscale mean value corresponding to the areas on both sides of the final boundary,

The tumor emphasis coefficient of each central pixel is proportional to the absolute value of the difference between the tumor line suspicion coefficient of the final boundary line within the neighborhood window of the central pixel, the gradient difference, and the gray mean value of the areas on both sides of the final boundary line.

Preferably, based on the tumor enhancement coefficient of each edge point in the edge contour binary map and the grayscale information of each pixel in the tumor radiofrequency ablation CT image, the expression for the enhanced grayscale value of each pixel is:

In the formula, is a linear normalization function,/> is the pixel point in the edge contour binary image,/> is any pixel in the tumor radiofrequency ablation CT image,/> is the baseline value of the tumor emphasis coefficient for each pixel in the tumor radiofrequency ablation CT image,/> is the tumor emphasis coefficient of the center pixel in the pixel neighborhood window in the edge contour binary map, Radiofrequency ablation of tumors in CT images/> The average gray value of the pixel in the neighborhood window with the pixel as the center pixel at the position,/> For the pixels in the tumor radiofrequency ablation CT image/> The gray value,/> For the pixels in the tumor radiofrequency ablation CT image/> Enhanced grayscale value.

The present invention at least has the following beneficial effects:

The present invention corrects the enhancement control coefficient in the local mean square error enhancement algorithm, constructs a characteristic index based on the characteristics around the tumor edge point in the tumor radiofrequency ablation CT image, evaluates the enhancement coefficient of each point in the image with respect to the tumor edge point, and avoids the need for enhancement algorithms. There is no feature of enhancing the tumor area, which improves the accuracy of enhancement.

This invention analyzes the neighborhood window centered on each edge point in the edge contour binary map of the tumor radiofrequency ablation CT image, extracts the differential features between the tumor and surrounding blood vessels, and combines corner point information and grayscale information and the curvature radius information of the pixel points, extracting the edge lines of suspected tumors that meet the conditions in the neighborhood window, realizing the preliminary positioning of the tumor edge lines in the image, and eliminating interference from other situations; the present invention combines the central boundary in the neighborhood window The tumor line suspicion coefficient of the line and the central boundary line are the grayscale information on both sides of the dividing line to obtain the tumor emphasis coefficient in the neighborhood window. According to the tumor emphasis coefficient in the neighborhood window, the enhancement of each pixel in the tumor radiofrequency ablation CT image is obtained. The final gray value accurately evaluates the degree of suspicion of the tumor edge line; the present invention obtains an enhanced image based on the tumor enhancement coefficient and gray information of each edge point in the image, thereby realizing adaptive enhancement of tumor radiofrequency ablation CT images.

The invention improves the pertinence of the enhancement range and has better enhancement effect.

Description of the drawings

In order to more clearly explain the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the drawings in the following description The drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without exerting creative efforts.

Figure 1 is a flow chart of a CT image enhancement system for tumor radiofrequency ablation guidance provided by the present invention.

Detailed ways

In order to further elaborate on the technical means and effects adopted by the present invention to achieve the intended purpose of the invention, the following is a CT image enhancement system for tumor radiofrequency ablation guidance proposed according to the present invention in conjunction with the drawings and preferred embodiments. The specific implementation, structure, characteristics and efficacy are described in detail as follows. In the following description, different terms "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Additionally, the specific features, structures, or characteristics of one or more embodiments may be combined in any suitable combination.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field to which the invention belongs.

A specific solution of a CT image enhancement system for tumor radiofrequency ablation guidance provided by the present invention will be described in detail below with reference to the accompanying drawings.

An embodiment of the present invention provides a CT image enhancement system for tumor radiofrequency ablation guidance. The system includes an image acquisition module and an image enhancement module.

Specifically, a CT image enhancement system for tumor radiofrequency ablation guidance in this embodiment provides the following intelligent enhancement guidance method for tumor radiofrequency ablation CT. Please refer to Figure 1. The method includes the following steps:

Step S001: The image acquisition module collects tumor radiofrequency ablation CT images of the treatment body through spiral CT.

This embodiment mainly collects tumor radiofrequency ablation CT images of the treatment body based on spiral CT, analyzes the image information and extracts features, and realizes tumor suspicion based on the distribution of the tumor and surrounding blood vessels in the neighborhood window of each edge point in the image. Edge line processing. In this embodiment, the tumor radiofrequency ablation CT image of the treatment body is first collected through spiral CT.

Since some noise inevitably appears in tumor radiofrequency ablation CT images, these noises will cause certain interference in the enhancement and positioning of the tumor area. Considering that there will be noise interference during the image acquisition process, the collected tumor radiofrequency ablation CT images are processed using median filtering technology to eliminate the effects of noise and image motion interference. It should be noted that the median filtering technology is a well-known technology and can be implemented by existing technology. It is not within the protection scope of this embodiment and will not be explained in detail here.

According to the above method of this embodiment, the tumor image of the treatment body is collected to obtain the radiofrequency ablation CT image of the tumor.

Step S002, the image enhancement module analyzes the tumor characteristics in the neighborhood window in which each edge point is the center pixel in the edge binary image, extracts the edge line in the neighborhood window, and analyzes the tumor edge at each point in the image. Point enhancement coefficient realizes the enhancement of tumor radiofrequency ablation CT images.

For tumor radiofrequency ablation CT images, it is considered that the tumor area is not significant in tumor radiofrequency ablation CT images and is interfered by other factors such as surrounding blood vessels. Therefore, this embodiment will set up an image enhancement module to perform feature analysis on each area in the tumor radiofrequency ablation CT image, and extract tumor features from the neighborhood window in which each edge point is the center pixel in the edge binary image. And calculate the tumor enhancement coefficient of each point in the tumor radiofrequency ablation CT image to achieve precise enhancement of the tumor area in the tumor radiofrequency ablation CT image and avoid the impact of low enhancement effect of the tumor area caused by untargeted tumor area.

The image enhancement module in this embodiment mainly includes the following steps:

The Canny operator is used to detect all edges in tumor radiofrequency ablation CT images, and the edge contour binary map is obtained. Since the tumor edges in tumor radiofrequency ablation CT images may not be clear, it is necessary to enhance the tumor area in the image to show its contrast compared with other areas, so as to help doctors more accurately identify patients' tumors. Radiofrequency ablation CT The possibility of tumors appearing in the images.

Because the edge of the tumor has irregular characteristics compared to blood vessels, and the edge of the tumor is located in the direction of the edge curvature radius, not only the tumor area will appear, but there will also be some blood vessel areas, and the distribution of such blood vessels is relatively uniform, compared with the tumor Areas with more vascular bifurcations.

In response to this situation, for each edge point in the edge contour binary image, calculate its surrounding Edge texture distribution within neighborhood windows. /> The value of is set by the implementer, here it is set to/> . Calculate the curvature radius vector of each boundary point on the central boundary line that passes through the center pixel in the neighborhood window and forms a closed area with the neighborhood window/> , based on the first boundary point on the boundary line from left to right and from top to bottom as the initial starting point, the curvature radius vector sequence of each central boundary line in the neighborhood window is obtained/> . Use the Harris corner detection algorithm to calculate the number of corner points on both sides of each central boundary line in the neighborhood window. If there is a large difference in the number of corner points appearing on both sides, then the central boundary line in this neighborhood window may be It is the suspected edge line of tumor. From this, the tumor line suspicion coefficient of each central boundary line in the neighborhood window of each central pixel is obtained, and the expression is:

In the formula, is the number of corner points with the largest number on both sides of the central boundary line in the center pixel neighborhood window,/> is the minimum number of corner points on both sides of the central boundary line in the center pixel neighborhood window,/> Indicates the proportional coefficient between the number of corner points on the side with the largest number and the number of corner points on the side with the least number. The larger the value, the more uneven the distribution of blood vessels on both sides of the line, and the more likely one side is a tumor area;/> is the variance of the curvature radius of each boundary point on the central boundary line in the neighborhood window of the central pixel point. The larger the value, the larger the value, the smaller the curvature radius size distribution of the boundary points on the central boundary line in the neighborhood window is relatively irregular, that is, The central boundary line in the neighborhood window is more curved than the blood vessel line;/> It is the variance of the curvature radius direction of each boundary point on the central boundary line in the neighborhood window of the central pixel point. The distribution of the curvature radius direction of the central boundary line in the neighborhood window is also more chaotic than that of the blood vessel edge line;/> is the tumor line suspicion coefficient of the central boundary line in the neighborhood window of the central pixel, /> The larger the value, the more likely the central boundary line in the neighborhood window is the tumor edge line.

Repeat the above method in this embodiment to obtain the tumor line suspicion coefficient in the neighborhood window where each edge point in the edge contour binary map is the central pixel point; obtain the center with the largest tumor line suspicion coefficient in the neighborhood window of the center pixel point The boundary line serves as the final boundary line.

Because when the edge moves closer to the central area of the tumor, its gray value will change, and this change in gray value will gradually increase; while the gray value on the blood vessel side will not show this gradient pattern.

In response to this situation, the grayscale gradient on each side of the final boundary line in the neighborhood window is calculated to show a regular increasing degree of change, and the grayscale gradient pattern on both sides of the final boundary line in the neighborhood window of the central pixel is obtained. Taking the concave side of the final boundary line as an example, calculate the grayscale gradient pattern on this side.

In the formula, is the number of boundary points on the final boundary line within the neighborhood window of the center pixel,/> is the final boundary line within the neighborhood window of the center pixel/> The boundary point is the concave side area of the starting point and along the boundary point/> The number of pixels on the straight line in the direction of the curvature radius,/> is an exponential function with natural constants as the base,/> ,/> is the final boundary line within the neighborhood window of the center pixel/> The boundary point is the concave side area of the starting point and along the boundary point/> /> on the straight line where the direction of the curvature radius is located No., No./> The gray value of two adjacent pixels is calculated/> /> Changes in grayscale between adjacent pixels; the value obtained by summing is less than 0, indicating that the grayscale on the concave side of each boundary point on the final boundary line of the neighborhood window generally changes incrementally, indicating that the area on this side may It is the area where the gray value of the tumor edge decreases. /> is the gray gradient pattern on the concave side of the final boundary line of the central pixel neighborhood window,/> The smaller the value, the grayscale changes on the straight line along the curvature radius direction of each boundary point on the final boundary line on the concave side are increasing, which means that this side may be a tumor area.

Repeat the above steps to obtain the gray gradient pattern on the convex side of the final boundary line within the neighborhood window of the central pixel.

Since the tissue density of tumors is usually greater than that of surrounding normal tissues, in tumor radiofrequency ablation CT images, tumors are generally high-density disease areas, and the brightness difference between the tumor area and other surrounding areas is large.

In response to this situation, combined with the final boundary line in the neighborhood window , calculate the mean gray value on both sides of the final boundary line. The larger the difference, the more it indicates that the gray distribution on both sides of the final boundary line is different, that is, one of the two sides may be a tumor area. At the same time, obtain the gray gradient pattern of the side with the largest value on both sides of the final boundary line/> and the gray gradient law of the smallest side , the gradient difference shows a regular increase in the grayscale gradient on one side of the tumor. The tumor emphasis coefficient of each central pixel is obtained based on the tumor line suspicion coefficient of the final boundary line in the neighborhood window of each central pixel and the grayscale information on both sides of the final boundary line. The specific expression is:

In the formula, is the tumor line suspicion coefficient of the final boundary line in the neighborhood window,/> is the gray average value on one side of the final boundary line in the neighborhood window,/> is the gray average value on the other side of the final boundary line in the neighborhood window,/> , by calculating the grayscale difference on both sides, the larger the value, the greater the difference, indicating that the two sides of the edge line are different areas, that is, the possibility of tumors appearing on one side is greater;/> is the gray gradient law of the side with the largest value in the final boundary line of the center pixel neighborhood window, is the gray gradient pattern on the side with the smaller value in the final boundary line of the center pixel neighborhood window. The greater the ratio of the gradient difference, the greater the difference in the gray gradient pattern on both sides of the final boundary line. This shows that the side with the largest grayscale gradient is more likely to be a tumor area;/> is the tumor emphasis coefficient of the central pixel neighborhood window,/> The larger the value, the more similar the area around the central pixel is to the tumor edge area, and the more likely the central pixel is to be the tumor edge point.

According to the tumor emphasis coefficient of each pixel in the obtained edge contour binary map, by changing the enhancement control coefficient in the local mean square error-based enhancement algorithm, the tumor radiofrequency ablation CT image is modified according to the low-frequency and high-frequency components of each point in the image. The tumor edge points in the image are enhanced.

In the formula, is a linear normalization function,/> is the pixel point in the edge contour binary image,/> is any pixel in the tumor radiofrequency ablation CT image,/> is the baseline value of the tumor emphasis coefficient for each pixel in the tumor radiofrequency ablation CT image,/> is the tumor emphasis coefficient of the center pixel in the pixel neighborhood window in the edge contour binary map, Radiofrequency ablation of tumors in CT images/> The average gray value of the pixel in the neighborhood window with the pixel as the center pixel at the position,/> The pixels in the tumor radiofrequency ablation CT image/> The gray value,/> The pixels in the tumor radiofrequency ablation CT image/> Enhanced grayscale value.

It should be noted, Take tumor radiofrequency ablation CT image/> The average gray value of the pixels in the 5x5 neighborhood window with the pixel as the center pixel at the position is used to determine/> The pixel at the position is a high-frequency or low-frequency component, so that the local characteristics of each pixel are enhanced to a corresponding degree.

By analyzing the situation around each edge point in the edge contour binary map, the algorithm focuses on enhancing the tumor edge in the tumor radiofrequency ablation CT image when enhancing the image, thereby completing intelligent enhancement of the tumor radiofrequency ablation CT image.

To sum up, this embodiment modifies the enhancement control coefficient in the local mean square error enhancement algorithm, constructs a feature index based on the characteristics around the tumor edge point in the tumor radiofrequency ablation CT image, and evaluates the enhancement of each point in the image with respect to the tumor edge point. coefficient, which avoids the feature of the enhancement algorithm that does not target the tumor area and improves the accuracy of enhancement.

This embodiment analyzes the neighborhood window centered on each edge point in the edge contour binary map of the tumor radiofrequency ablation CT image, extracts the differential features between the tumor and surrounding blood vessels, and combines corner point information and grayscale information and the curvature radius information of pixels, extract the edge lines of suspected tumors that meet the conditions in the neighborhood window, realize the preliminary positioning of the tumor edge lines in the image, and eliminate interference from other situations; this embodiment combines the edge lines of the suspected tumors in the neighborhood window The tumor line suspicion coefficient of the central boundary line and the central boundary line are the grayscale information on both sides of the dividing line to obtain the tumor emphasis coefficient in the neighborhood window. According to the tumor emphasis coefficient in the neighborhood window, each pixel in the tumor radiofrequency ablation CT image is obtained. The gray value after point enhancement accurately assesses the degree of suspicion of the tumor edge line; this embodiment obtains an enhanced image based on the tumor enhancement coefficient and gray value information of each edge point in the image, thereby realizing automatic reconstruction of tumor radiofrequency ablation CT images. Adaptation enhances.

This embodiment improves the pertinence of the enhancement range and has better enhancement effect.

It should be noted that the above-mentioned order of the embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. Specific embodiments of this specification have been described above. Additionally, the processes depicted in the figures do not necessarily require the specific order shown, or sequential order, to achieve desirable results. Multitasking and parallel processing are also possible or may be advantageous in certain implementations.

Each embodiment in this specification is described in a progressive manner. The same and similar parts between the various embodiments can be referred to each other. Each embodiment focuses on its differences from other embodiments.

The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention shall be included in the protection scope of the present invention. Inside.

Claims (5)

1. A CT image enhancement system for tumor radio frequency ablation guidance, the system comprising:

and an image acquisition module: collecting a tumor radio frequency ablation CT image;

an image enhancement module: edge detection is carried out on the tumor radio frequency ablation CT image to obtain an edge contour binary image, and each edge point in the edge contour binary image is taken as a central pixel point to obtain a neighborhood window of each central pixel point;

obtaining a central boundary line in the neighborhood window according to the relation between each edge line in each central pixel point neighborhood window and the central pixel point;

extracting corner information on two sides of each central boundary line in each central pixel point neighborhood window by adopting a corner detection algorithm; obtaining a curvature radius vector sequence of each center boundary line according to the curvature of the boundary point on each center boundary line in each center pixel point neighborhood window;

obtaining a tumor line suspected coefficient of each central boundary line in each central pixel point neighborhood window according to the curvature radius vector sequence of each central boundary line in each central pixel point neighborhood window and the corner information on two sides of each central boundary line, and recording the central boundary line with the maximum tumor line suspected coefficient in each central pixel point neighborhood window as the final boundary line of each central pixel point neighborhood window;

calculating the gray level gradient rule of the concave side and the convex side of the final boundary line in the neighborhood window of each central pixel point; obtaining the gradient difference of the final boundary line in the neighborhood window of the central pixel point according to the gray gradient rules at two sides of the final boundary line in the neighborhood window of each central pixel point;

obtaining a tumor emphasis coefficient of each center pixel point according to a tumor line suspected coefficient and a gradient difference of a final boundary line in a neighborhood window of each center pixel point;

according to the tumor enhancement coefficient of each edge point in the edge contour binary image and the gray level information of each pixel point in the tumor radio frequency ablation CT image, the gray level value of each pixel point after enhancement is obtained, and intelligent enhancement of the tumor radio frequency ablation CT image is completed;

the specific steps of obtaining the central boundary line in the neighborhood window according to the relation between each edge line in each central pixel point neighborhood window and the central pixel point are as follows:

marking an edge line in the neighborhood window which passes through the central pixel point and intersects with the neighborhood window to form a closed area as a central boundary line in the neighborhood window according to the condition that the neighborhood window of each central pixel point is met;

the specific steps of obtaining the tumor line suspected coefficient of each central boundary line in each central pixel point neighborhood window according to the curvature radius vector sequence of each central boundary line in each central pixel point neighborhood window and the corner point information on two sides of each central boundary line are as follows:

based on the curvature radius vector sequence of each center boundary line in each center pixel point neighborhood window,

obtaining the curvature radius variance of each vector module length in the curvature radius vector sequence of each center boundary line, obtaining the curvature radius direction variance of each vector direction in the curvature radius vector sequence of each center boundary line,

the suspected coefficient of the tumor line of each central boundary line in each central pixel point neighborhood window is in direct proportion to the sum of the curvature radius variance and the curvature radius direction variance of the corresponding central boundary line and in direct proportion to the corner point information on two sides of the corresponding central boundary line;

the expression of the tumor line suspected coefficient of each central boundary line in each central pixel point neighborhood window is as follows:

in the method, in the process of the application,for the number of corner points with the largest number in the two sides of the central boundary line in the neighborhood window of the central pixel point, +.>For the least number of corner points in the central boundary line in the central pixel neighborhood window, +.>For the variance of the radius of curvature where each boundary point is located on the central boundary line in the neighborhood window of the central pixel point,/for the center pixel point>For the variance of the radius of curvature direction of each boundary point on the central boundary line in the neighborhood window of the central pixel point, +.>The suspected coefficient of the tumor line is the boundary line of the center in the neighborhood window of the center pixel point;

the expression for calculating the gray level gradient rule of the concave side of the final boundary line in each central pixel point neighborhood window is as follows:

in the method, in the process of the application,for the number of boundary points on the final boundary line within the neighborhood window of center pixel points,/>is the first +.o on the final boundary line in the neighborhood window of the central pixel point>Concave side area with boundary point as starting point and along boundary point +.>The number of pixels on a straight line along the radius of curvature direction of (a) is +.>As an exponential function based on natural constants, < +.>、/>Is the first +.o on the final boundary line in the neighborhood window of the central pixel point>Concave side area with boundary point as starting point and along boundary point +.>Is the first +.on the straight line where the radius of curvature direction of (a) is located>Person, th->Gray values of two adjacent pixels, < >>A gray level gradient rule of the concave side of the final boundary line of the neighborhood window of the central pixel point;

the specific steps of obtaining the tumor emphasis coefficient of each central pixel point according to the suspected coefficient and the gradient difference of the tumor line of the final boundary line in the neighborhood window of each central pixel point are as follows:

based on the final boundary line of each center pixel neighborhood window,

the gray average value corresponding to the areas at both sides of the final boundary is obtained,

the tumor emphasis coefficient of each central pixel point is in direct proportion to the suspected coefficient of the tumor line of the final boundary line in the neighborhood window of the central pixel point, the gradient difference and the absolute value of the difference value of the gray average values of the areas at two sides of the final boundary line;

the specific expression of the tumor emphasis coefficient of each center pixel point is as follows:

in the method, in the process of the application,tumor line suspected coefficient for final boundary line in neighborhood window, < ->For the gray mean value on the side of the final border line in the neighborhood window, < >>For the gray mean value on the other side of the final boundary line in the neighborhood window, < >>Is the gray level gradient rule of the maximum value side in the final boundary line of the neighborhood window of the central pixel point,/for the maximum value side>Is the gray level gradient rule of the side with the minimum value in the final boundary line of the neighborhood window of the central pixel point,/V>Is the middle warmerTumor emphasis coefficients for a cardiac pixel neighborhood window;

the expression of the gray value of each pixel point after the gray information of each pixel point in the tumor radio frequency ablation CT image is enhanced according to the tumor enhancement coefficient of each edge point in the edge contour binary image is as follows:

in the method, in the process of the application,as a linear normalization function>For pixel point in the edge contour binary diagram, < ->For any pixel point in tumor radio frequency ablation CT image,>for the tumor emphasis coefficient benchmark value of each pixel point in the tumor radio frequency ablation CT image, the +.>Tumor emphasis coefficient for central pixel point in pixel point neighborhood window in edge contour binary image, < +.>For in tumor radio frequency ablation CT image +.>The gray average value of the pixel point in the neighborhood window taking the pixel point as the central pixel point at the position,pixel points in radio frequency ablation CT image of tumor>Gray value of +.>Pixel point in CT image for tumor radio frequency ablationEnhanced gray values.

2. The CT image enhancement system for tumor radio frequency ablation guidance according to claim 1, wherein the specific step of extracting corner information on both sides of each center boundary line in each center pixel point neighborhood window by using a corner detection algorithm comprises the following steps:

based on the each of the center pixel points,

respectively acquiring angular points at two sides of each central boundary line in the corresponding neighborhood window;

the corner number of the most side of the corner points in the two sides is marked as a first corner number, and the corner number of the least side of the corner points in the two sides is marked as a second corner number;

and (3) recording the ratio of the number of the first corner points to the number of the second corner points as corner point information on two sides of each central boundary line in each central pixel point neighborhood window.

3. A CT image enhancement system for tumor radio frequency ablation guidance according to claim 1, wherein the specific step of obtaining the vector sequence of the radius of curvature vector of each center boundary line according to the curvature of each boundary point on each center boundary line in each center pixel point neighborhood window is as follows:

calculating the curvature radius vector of each boundary point on each central boundary line of each central pixel point neighborhood window,

and according to the boundary points on the central boundary lines in the first neighborhood window from left to right and from top to bottom as initial starting points, arranging the boundary points on the central boundary lines in the neighborhood window into a sequence to obtain a curvature radius vector sequence of each central boundary line in the neighborhood window.

4. The CT image enhancement system for tumor radio frequency ablation guidance according to claim 1, wherein the specific step of marking the central boundary line with the largest tumor line suspected coefficient in each central pixel point neighborhood window as the final boundary line of each central pixel point neighborhood window is as follows:

obtaining the tumor line suspected coefficient of each central boundary line in each central pixel point neighborhood window,

and taking a central boundary line corresponding to the maximum value of the tumor line suspected coefficient in the neighborhood window as a final boundary line.

5. The CT image enhancement system for tumor radio frequency ablation guidance according to claim 1, wherein the specific step of obtaining the gradient difference of the final boundary line in the neighborhood window of the central pixel point according to the gradient rule of the final boundary line on both sides of the neighborhood window of each central pixel point comprises the following steps:

based on each central pixel point neighborhood window,

the side with the largest gray gradient rule among the two sides of the corresponding final boundary line is marked as a first gradient region,

the side with the smallest gray gradient rule of the two sides of the corresponding final boundary line is marked as a second gradient region,

and marking the ratio of the gray level gradient rule of the first gradient region to that of the second gradient region as the gradient difference of the final boundary line in the neighborhood window of the central pixel point.