Confocal endoscope video image enhancement method based on principal component analysis

Abstract

The invention discloses a confocal endoscope video image enhancement method based on principal component analysis. The confocal endoscope video image enhancement method comprises the following steps: extracting an effective area of an endoscopic image; the effective area is divided into a plurality of sub-areas, and the sub-areas are overlapped; calculating a principal component of the sub-region image data based on the time sequence; calculating correction factors of the sub-image areas; combining the correction factors of the sub-image areas; and reconstructing the endoscopic image through the correction factor to obtain an enhanced reconstructed image. According to the method, the correction factor can be calculated adaptively according to the local region features of the image based on the video stream, so that the honeycomb structure of the confocal endoscopic scanning system is removed, and the imaging quality of the reconstructed image is improved.

Description

Technical Field

[0001] The invention belongs to the technical field of optical fiber bundle image processing, in particular to a confocal endoscope video image enhancement method based on principal component analysis. Background Art

[0002] Fluorescence confocal endoscopic imaging technology is a technology that combines fiber optic endoscopy technology with confocal scanning microscopy. It can perform non-invasive tissue examination on living tissue and is widely used in the field of medical imaging.

[0003] Fluorescence confocal endoscopic imaging technology uses a confocal scanning system to achieve imaging through a fiber bundle containing multiple optical fibers. Fiber bundles play an important role in endoscopic imaging and provide a way to visualize the cellular level for the diagnosis of a series of diseases in vivo. Typically, a fiber bundle contains a large number of fiber cores (usually up to 30,000), which can transmit and acquire optical signals simultaneously. However, the core spacing (i.e., cladding) causes information loss between fibers, and the distribution structure of the fibers appears in the acquired image, presenting a honeycomb periodic structure, affecting the readability of the image. Usually, a reference image is used to calculate the correction parameters of the image at the beginning, and a fixed correction coefficient is used to correct the image. However, during long-term use, due to external forces and other reasons, the position of the fiber interface end face of the confocal scanning system will change slightly, defocus will occur in some areas of the image, and the position and intensity of the honeycomb pattern will change. This makes it necessary to use a self-learning algorithm that can calculate the correction parameters in real time during use to eliminate the confocal honeycomb artifacts.

[0004] In the current processing method ("Optical fiber bundle image processing method and device. Publication number: CN107678153A Publication date: 2018.02.09 Application number: 201710959003.1"), before the confocal scanning system is started, the correction parameters are calculated using the reference image, and the image is corrected according to the correction parameters. The operation is relatively complicated, and after long-term use, the honeycomb pattern will reappear due to the displacement of the interface end face. Summary of the invention

[0005] The purpose of the present invention is to address the problems existing in the above-mentioned prior art and provide a confocal endoscopic video image enhancement method based on principal component analysis, which performs local analysis on the image based on time sequence and removes the image honeycomb structure.

[0006] The technical solution to achieve the purpose of the present invention is: a confocal endoscope video image enhancement method based on principal component analysis, the method comprising the following steps:

[0007] Extracting effective areas of endoscopic images;

[0008] Divide the effective area into several sub-areas, and the sub-areas overlap each other;

[0009] Calculate the principal components of the sub-region image data based on the time series;

[0010] Calculating correction factors for sub-image regions;

[0011] Correction factors for merged sub-image regions;

[0012] The endoscopic image is reconstructed using the correction factor to obtain an enhanced reconstructed image.

[0013] Furthermore, the extracting of the effective area of ​​the endoscopic image specifically includes:

[0014] Set the threshold σ;

[0015] For each endoscopic image , based on the threshold σ segmentation to obtain the corresponding mask ;

[0016] Perform AND operation on all masks and obtain the maximum connected domain of the masks;

[0017] Perform morphological dilation on the operation result to obtain the mask of the effective area.

[0018] Furthermore, dividing the effective area into a plurality of sub-areas specifically includes:

[0019] According to the preset sub-region width wd and the sub-region overlap size overlap, the effective area of ​​the endoscopic image P with a width w and a height h is divided into N small square sub-regions, where the width is divided into Small areas, divided by height Small area:

[0020]

[0021]

[0022]

[0023] The i-th small square sub-region image Starting position , And width ,high for:

[0024]

[0025]

[0026]

[0027]

[0028]

[0029]

[0030] According to the mask of the valid area, each small square sub-region i contains Valid pixels, filter out small square sub-regions with a preset number of valid pixels to form a region set R to be analyzed;

[0031] Extract the serial numbers of all valid pixels in each sub-region i in the region set R relative to the starting position of the sub-region, and form a position sequence from small to large .

[0032] Furthermore, the calculating the principal component of the sub-region image data based on the time series specifically includes:

[0033] For the tth endoscopic image, according to the position sequence of the sub-regions , get the pixel value of the corresponding position, and form a length of Vector ;

[0034] In the time domain, PCA analysis is performed on the data of the sub-region:

[0035] (1) Data preprocessing: normalize the raw data of the same sub-region i at each moment and calculate the logarithm:

[0036]

[0037] In the formula, represents the preprocessing result corresponding to the sub-region i of the t-th endoscopic image, is a fixed minimum value used to prevent the logarithmic term from being zero;

[0038] (2) Calculate the average value of the sub-area data within a period of time :

[0039]

[0040] In the formula, represents the total number of endoscopic images;

[0041] (3) Calculate the covariance matrix :

[0042]

[0043] (4) Obtaining the covariance matrix Take the eigenvalue With the eigenvector e:

[0044]

[0045]

[0046] In the formula, represents the jth eigenvector obtained by decomposing the matrix of the i-th sub-region, represents the jth eigenvalue obtained after decomposing the matrix of the i-th sub-region, j=1,2,..., , Indicates the number of eigenvectors or eigenvalues;

[0047] (5) Extracting principal component weights :

[0048] .

[0049] Furthermore, the calculating of the correction factor of the sub-image area specifically includes:

[0050] For all The minimum value in Perform mean-shift cluster analysis to extract cluster centers ;

[0051] Reduce the interference effect of artifacts on small area images It is expressed as:

[0052]

[0053] According to the position number of the data in the area Will Converted into fiber influence coefficient , the calculation formula is:

[0054] .

[0055] Furthermore, the correction factor of the merged sub-image area specifically includes:

[0056] Compute points in overlapping regions to sub-region images The closest distance to the border :

[0057]

[0058] In the formula, represents the points in the overlapping area;

[0059] Calculate sub-region image The weight parameter :

[0060]

[0061]

[0062] Calculate the total weight parameter of the overlapping area :

[0063]

[0064] Calculate the final correction factor for each overlapping area:

[0065] .

[0066] Further, reconstructing the endoscopic image by using the correction factor to obtain an enhanced reconstructed image specifically includes:

[0067] Correct the original endoscopic image using the correction factor:

[0068]

[0069] In the formula, express the corresponding corrected endoscopic image;

[0070] Obtaining endoscopic images The maximum and minimum values ​​of Perform normalization.

[0071] On the other hand, a confocal endoscope video image enhancement system based on principal component analysis is provided, the system comprising:

[0072] The first module is used to extract the effective area of ​​the endoscopic image;

[0073] The second module is used to divide the effective area into a plurality of sub-areas, and the sub-areas overlap each other;

[0074] The third module is used to calculate the principal components of the sub-region image data based on the time series;

[0075] The fourth module is used to calculate the correction factor of the sub-image area;

[0076] A fifth module is used to merge the correction factors of the overlapping areas of the sub-image areas;

[0077] The sixth module is used to reconstruct the endoscopic image by using the correction factor to obtain an enhanced reconstructed image.

[0078] On the other hand, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the confocal endoscopic video image enhancement method based on principal component analysis when executing the computer program.

[0079] On the other hand, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the confocal endoscopic video image enhancement method based on principal component analysis is implemented.

[0080] Compared with the prior art, the present invention has the following significant advantages:

[0081] (1) This method can automatically calculate the impact of fiber loss based on the video stream and correct the image accordingly. There is no need to calibrate the light intensity of the fiber bundle at the beginning. In addition, the processing effect of this method on honeycomb patterns will increase with the length of the video.

[0082] (2) This method can adaptively calculate the correction factor based on the local area characteristics of the image based on the video stream, thereby removing the honeycomb structure existing in the confocal endoscopy scanning system and improving the imaging quality of the reconstructed image.

[0083] (3) Compared with the neural network method, this method does not require the use of labeled images for training, is more interpretable, simpler to calculate, and more efficient.

[0084] The present invention is further described in detail below in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] Figure 1 The present invention is a flowchart of a confocal endoscope video image enhancement method based on principal component analysis in one embodiment.

[0086] Figure 2 is a reference image of a video stream obtained in an embodiment.

[0087] Figure 3 The effective area mask image is obtained according to the video stream in an embodiment.

[0088] Figure 4 is a schematic diagram of segmenting sub-regions according to a mask image in an embodiment, wherein Figure 4 (a) is a schematic diagram of dividing the effective area into several sub-areas. Figure 4 (b) is a schematic diagram of the relative position coordinates of the pixels in one of the sub-regions. Figure 4 (c) is a schematic diagram of the brightness of pixels in the sub-region. Figure 4 (d) is the position sequence composed of the sub-region coordinate position and brightness With brightness value vector Schematic diagram of Figure 4 In the small square area (b), the coordinates of each small square are (1,0), (1,1), (2,0), (2,1), (2,2), (2,3), (3,0), (3,1), (3,2), (3,3), (3,4), (3,5), (4,0), (4,1), (4,2), (4,3), (4,4), (4,5), (4,6), (5,0), (5,1), (5,2), (5,3), (5,4), (5,5), (5,6), (6,0), (6,1), (6,2), (6,3), (6,4), (6,5), (6,6), (7,0), (7,1), (7,2), (7,3), (7,4), (7,5), (7,6), (8,0), (8,1), (8,2), (8,3), (8,4), (8,5), (8,6), (9,0), (9,1), (9,2), (9,3), (9,4), (9,5), (9,6), (9,7), (10,0), (10,1), (10,2), (10,3), (10,4), (10,5), (10,6), (10,7), (10,8), (10,9), (10,10), (11,0), (11,1), (11,2), (11,3), (11,4), (11,5), (11,6), (11,7), (11,8), (11,9), (11,10), (11,11).

[0089] Figure 5 is a sub-region image in an embodiment, where Figure 5 (a) to (c) are images of a sub-region at different times.

[0090] Figure 6 The figure is a corrected image obtained by sub-region principal component analysis in one embodiment.

[0091] Figure 7 is an overall correction parameter coefficient image in one embodiment.

[0092] Figure 8 In one embodiment, Figure 1 The image obtained after correction. DETAILED DESCRIPTION

[0093] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0094] It should be noted that if the embodiments of the present invention involve directional indications (such as up, down, left, right, front, back...), the directional indications are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0095] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present invention, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or suggesting their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0096] In one embodiment, in combination Figure 1 , provides a confocal endoscope video image enhancement method based on principal component analysis, the method comprising the following steps:

[0097] S01, extracting the effective area of ​​the endoscopic image;

[0098] S02, dividing the effective area into a plurality of sub-areas, and the sub-areas overlap each other;

[0099] S03, calculating the principal components of the sub-region image data based on the time series;

[0100] S04, calculating the correction factor of the sub-image area;

[0101] S05, correction factor for merged sub-image area;

[0102] S06, reconstructing the endoscopic image using the correction factor to obtain an enhanced reconstructed image;

[0103] S07, putting the newly captured image into the time series image queue to be analyzed, and repeating step S03 at intervals.

[0104] Here, the confocal endoscope uses a galvanometer to scan the laser to the end face of the optical fiber bundle. The laser is transmitted to the surface of the object through the optical fiber, and the fluorescent signal generated by the biological tissue is collected through the optical fiber to generate an image. The scanning range of the galvanometer is generally rectangular, while the optical fiber bundle is a circular cross-section. The real effective area of ​​the image only exists within the effective cross-section of the light bundle (such as Figure 2As shown in the figure, in order to reduce the amount of calculation and the interference of image information in irrelevant areas, it is necessary to first extract the effective area of ​​the image. Since there is no fluorescence signal returned in the invalid area, the brightness value of its point on the image is basically zero.

[0105] Further, in one of the embodiments, the step of extracting the effective area of ​​the endoscopic image in S01 specifically includes:

[0106] S011, set threshold σ;

[0107] S012, for each endoscopic image , based on the threshold σ segmentation to obtain the corresponding mask ;

[0108] S013, performing AND operation on all masks;

[0109] S014, perform morphological processing (corrosion operation) on the operation result to eliminate the influence of small holes and obtain the mask of the effective area (such as Figure 3 as shown).

[0110] Further, in one embodiment, the step of dividing the effective area into a plurality of sub-areas in S02 specifically includes:

[0111] S021, according to the preset sub-region width wd and the sub-region overlap size overlap, the effective area of ​​the endoscopic image P with a width w and a height h is divided into N small square sub-regions, wherein the width is divided into Small areas, divided by height Small area:

[0112]

[0113]

[0114]

[0115] The i-th small square sub-region image Starting position , And width ,high for:

[0116]

[0117]

[0118]

[0119]

[0120]

[0121]

[0122] S022, according to the mask of the effective area, each small square sub-area i contains valid pixels, and select a small square sub-region with a preset number of valid pixels (such as Figure 4 As shown), constitute the region set R to be analyzed;

[0123] S023, extract the serial numbers of all valid pixel points in each sub-region i in the region set R relative to the starting position of the sub-region, and form a position sequence from small to large .

[0124] Furthermore, in one embodiment, the principal component algorithm is a method of data dimensionality reduction to map high-dimensional data into a low-dimensional space so that the variance between the coordinates of the midpoints in the low-dimensional space on the new coordinate axis (principal component) is as large as possible. PCA is widely used in data analysis in various industries and can be used for image data compression and feature dimensionality reduction of neural networks. In the present invention, the honeycomb influence coefficient can be approximated by performing principal component analysis in the time domain.

[0125] The step of calculating the principal components of the sub-region image data based on the time series in S03 specifically includes:

[0126] For the tth endoscopic image, the sub-image region data is converted into a one-dimensional vector. The process is as follows: Figure 3 As shown, according to the position sequence of the sub-region , read the pixel value at the corresponding position, and form a length of Vector , Metrix2Vector represents the operation of matrix steering;

[0127] Metrix2Vector(Mi,Pvi)

[0128] In the time domain, PCA analysis is performed on the data of the sub-region:

[0129] (1) Data preprocessing: normalize the raw data of the same sub-region i at each moment and calculate the logarithm:

[0130]

[0131] In the formula, represents the preprocessing result corresponding to the sub-region i of the t-th endoscopic image, is a fixed minimum value (1e-6) to avoid the logarithmic term being zero;

[0132] (2) Calculate the average value of the sub-area data within a period of time :

[0133]

[0134] In the formula, represents the total number of endoscopic images;

[0135] (3) Calculate the covariance matrix :

[0136]

[0137] (4) Covariance matrix Perform singular value SVD decomposition to obtain Eigenvalues ​​( ) and the eigenvector :

[0138]

[0139]

[0140] (5) Select the vector with the largest eigenvalue As the principal component, calculate the principal component weight of each sub-region image :

[0141] .

[0142] Further, in one embodiment, the step of calculating the correction factor of the sub-image area in S04 specifically includes:

[0143] The impact of artifacts on small area images can be expressed as follows:

[0144]

[0145] in, is the original signal, Due to the small size of each region, the honeycomb artifacts are often seen in a large part of the time series images (such as Figure 5 (shown) In space, it can be approximated as a constant distribution. Since the size of the sub-region is larger than the size of a single optical fiber, the signal is basically lossless in the center of the optical fiber. The maximum value of the coefficient transmitted within the area is 1, then:

[0146]

[0147]

[0148] When the pixel value in a small area When the distribution is approximately constant, the processed data and interference effects The relationship can be expressed as:

[0149]

[0150] therefore Minimum value of With principal component Will be more concentrated, using the mean shift clustering algorithm to extract the cluster center , then the vector of interference influence It can be expressed as:

[0151]

[0152] Calculated like Figure 6 As shown, according to the sequence number of the data in the area The Fi vector can be converted into an interference impact matrix :

[0153] i, )

[0154] Converted into optical fiber influence coefficient, i.e. correction coefficient , correction factor It can be calculated as follows:

[0155] .

[0156] Further, in one of the embodiments, the correction factors of merging the sub-image areas in S05 are specifically: merging the correction factors of the overlapping areas of the sub-image areas (because each sub-area may overlap with other sub-areas, the calculation of the correction coefficients of the overlapping areas needs to be mixed and superimposed based on weights, and the farther from the area boundary, the greater the weight), specifically including:

[0157] S051, calculate the point to sub-region image in the overlapping area The closest distance to the border

[0158]

[0159] In the formula, Represents the coordinates of the points in the overlapping area relative to the origin of the i-th sub-area;

[0160] S052, calculate sub-region image The weight parameter :

[0161]

[0162]

[0163] S053, set the total image parameter weight Q to zero:

[0164]

[0165] Where x and y are the coordinates of the pixel relative to the origin of the entire image area.

[0166] S054, the weight of the sub-region Perform overlay calculation to calculate the total weight parameter of the overlapping area :

[0167]

[0168] S055, calculate the final correction factor for each overlapping area:

[0169] .

[0170] Further, in one embodiment, the step of reconstructing the endoscopic image by using the correction factor to obtain an enhanced reconstructed image in S06 specifically includes:

[0171] S061, after calculating the overall correction factor (such as Figure 7 As shown), the original endoscopic image is corrected using the correction factor:

[0172]

[0173] In the formula, express the corresponding corrected endoscopic image;

[0174] S062, obtain endoscopic image The maximum and minimum values ​​of Normalization is performed to ensure the consistency of image brightness before and after correction. The final processed image result is as follows: Figure 8 shown.

[0175] Further, in one embodiment, S07 specifically includes:

[0176] S071, forming a queue for image analysis. Using a first-in-first-out strategy, newly captured images are placed into the queue to be analyzed, and the first image placed into the queue is removed from the queue.

[0177] S072, when the number of newly entered images exceeds half, repeat step S03 to recalculate the correction parameter C of the image so that it can adapt to the influence caused by the change of honeycomb artifacts.

[0178] In one embodiment, a confocal endoscope video image enhancement system based on principal component analysis is provided, the system comprising:

[0179] The first module is used to extract the effective area of ​​the endoscopic image;

[0180] The second module is used to divide the effective area into a plurality of sub-areas, and the sub-areas overlap each other;

[0181] The third module is used to calculate the principal components of the sub-region image data based on the time series;

[0182] The fourth module is used to calculate the correction factor of the sub-image area;

[0183] A fifth module is used to merge the correction factors of the overlapping areas of the sub-image areas;

[0184] The sixth module is used to reconstruct the endoscopic image by using the correction factor to obtain an enhanced reconstructed image.

[0185] For the specific definition of the confocal endoscopic video image enhancement system based on principal component analysis, please refer to the definition of the confocal endoscopic video image enhancement method based on principal component analysis above, which will not be repeated here. Each module in the above-mentioned confocal endoscopic video image enhancement system based on principal component analysis can be implemented in whole or in part through software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0186] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the following is achieved:

[0187] Extracting effective areas of endoscopic images;

[0188] Divide the effective area into several sub-areas, and the sub-areas overlap each other;

[0189] Calculate the principal components of the sub-region image data based on the time series;

[0190] Calculating correction factors for sub-image regions;

[0191] Correction factors for merged sub-image regions;

[0192] The endoscopic image is reconstructed using the correction factor to obtain an enhanced reconstructed image.

[0193] For the specific definition of each step, please refer to the definition of the confocal endoscopic video image enhancement method based on principal component analysis mentioned above, which will not be repeated here.

[0194] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the computer program implements:

[0195] Extracting effective areas of endoscopic images;

[0196] Divide the effective area into several sub-areas, and the sub-areas overlap each other;

[0197] Calculate the principal components of the sub-region image data based on the time series;

[0198] Calculating correction factors for sub-image regions;

[0199] Correction factors for merged sub-image regions;

[0200] The endoscopic image is reconstructed using the correction factor to obtain an enhanced reconstructed image.

[0201] For the specific definition of each step, please refer to the definition of the confocal endoscopic video image enhancement method based on principal component analysis mentioned above, which will not be repeated here.

[0202] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A confocal endoscope video image enhancement method based on principal component analysis, characterized in that: The method comprises the following steps: Extracting effective areas of endoscopic images; Divide the effective area into several sub-areas, and the sub-areas overlap each other; Calculate the principal components of the sub-region image data based on the time series; Calculating correction factors for sub-image regions; Correction factors for merged sub-image regions; The endoscopic image is reconstructed using the correction factor to obtain an enhanced reconstructed image.

2. The confocal endoscope video image enhancement method based on principal component analysis according to claim 1 is characterized in that: The extracting of the effective area of ​​the endoscopic image specifically includes: Set the threshold σ; For each endoscopic image , based on the threshold σ segmentation to obtain the corresponding mask ; Perform AND operation on all masks and obtain the maximum connected domain of the masks; Perform morphological dilation on the operation result to obtain the mask of the effective area.

3. The confocal endoscope video image enhancement method based on principal component analysis according to claim 1 is characterized in that: The dividing of the effective area into a plurality of sub-areas specifically includes: According to the preset sub-region width wd and the sub-region overlap size overlap, the effective area of ​​the endoscopic image P with a width w and a height h is divided into N small square sub-regions, where the width is divided into Small areas, divided by height Small area: ； ； ； The i-th small square sub-region image Starting position , And width ,high for: ； ； ； ； ； ； According to the mask of the valid area, each small square sub-region i contains Valid pixels, filter out small square sub-regions with a preset number of valid pixels to form a region set R to be analyzed; Extract the serial numbers of all valid pixels in each sub-region i in the region set R relative to the starting position of the sub-region, and form a position sequence from small to large .

4. The confocal endoscope video image enhancement method based on principal component analysis according to claim 3 is characterized in that: The calculating the principal component of the sub-region image data based on the time series specifically includes: For the tth endoscopic image, according to the position sequence of the sub-regions , get the pixel value of the corresponding position, and form a length of Vector ; In the time domain, PCA analysis is performed on the data of the sub-region: (1) Data preprocessing: normalize the raw data of the same sub-region i at each moment and calculate the logarithm: ； In the formula, represents the preprocessing result corresponding to the sub-region i of the t-th endoscopic image, is a fixed minimum value used to prevent the logarithmic term from being zero; (2) Calculate the average value of the sub-area data within a period of time : ； In the formula, represents the total number of endoscopic images; (3) Calculate the covariance matrix : ； (4) Obtaining the covariance matrix Take the eigenvalue With the eigenvector e: ； ； In the formula, represents the jth eigenvector obtained by decomposing the matrix of the i-th sub-region, represents the jth eigenvalue obtained after decomposing the matrix of the i-th sub-region, j=1,2,..., , Indicates the number of eigenvectors or eigenvalues; (5) Extracting principal component weights : 。 5. The confocal endoscope video image enhancement method based on principal component analysis according to claim 4 is characterized in that: The calculating of the correction factor of the sub-image area specifically includes: For all The minimum value in Perform mean-shift cluster analysis to extract cluster centers ; Reduce the interference effect of artifacts on small area images It is expressed as: ； According to the position number of the data in the area Will Converted into fiber influence coefficient , the calculation formula is: 。 6. The confocal endoscope video image enhancement method based on principal component analysis according to claim 5 is characterized in that: The correction factor of the merged sub-image area specifically includes: Compute points in overlapping regions to sub-region images The closest distance to the border : ； In the formula, represents the points in the overlapping area; Calculate sub-region image The weight parameter : ； ； Calculate the total weight parameter of the overlapping area : ； Calculate the final correction factor for each overlapping area: 。 7. The confocal endoscope video image enhancement method based on principal component analysis according to claim 6 is characterized in that: The step of reconstructing the endoscopic image by using the correction factor to obtain an enhanced reconstructed image specifically includes: Correct the original endoscopic image using the correction factor: ； In the formula, express the corresponding corrected endoscopic image; Obtaining endoscopic images The maximum and minimum values ​​of Perform normalization.

8. A confocal endoscope video image enhancement system based on the method according to any one of claims 1 to 7, characterized in that: The system comprises: The first module is used to extract the effective area of ​​the endoscopic image; The second module is used to divide the effective area into a plurality of sub-areas, and the sub-areas overlap each other; The third module is used to calculate the principal components of the sub-region image data based on the time series; The fourth module is used to calculate the correction factor of the sub-image area; A fifth module is used to merge the correction factors of the overlapping areas of the sub-image areas; The sixth module is used to reconstruct the endoscopic image by using the correction factor to obtain an enhanced reconstructed image.

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images