Image processing-based intelligent panoramic processing method and system for borehole peeping

Abstract

The present application relates to the technical field of borehole peeping video processing, in particular to a borehole peeping intelligent panoramic processing method and system based on image processing, which comprises the following steps: obtaining a borehole peeping video to obtain the rectangular coordinates of all pixel points of different radius circles; constructing a pixel rectangular strip corresponding to each circle according to the rectangular coordinates of all pixel points of each circle; the order of pixel points from 0 degrees to 360 degrees in the circle is the order of pixel points from left to right in the pixel rectangular strip; each circle obtains a pixel rectangular strip; the pixel rectangular strips are spliced from top to bottom in the order of circle radius from small to large to obtain a ladder-shaped reconstructed image; the ladder-shaped reconstructed image is subjected to interpolation processing to obtain a reconstructed rectangular image; the reconstructed rectangular image is subjected to repeated area detection to remove overlapping areas to obtain a panoramic image of the borehole. A high-quality panoramic image is generated to provide technical support for geological exploration in tunnel engineering.

Description

Technical Field

[0001] This invention relates to the field of borehole peep video processing technology, and in particular to a method and system for intelligent panoramic processing of borehole peep videos based on image processing. Background Technology

[0002] In tunnel engineering, borehole inspection technology is an important geological exploration method used to obtain information about the internal condition of the surrounding rock or the rock mass in front of the tunnel face, in order to help technicians evaluate and design the tunnel construction plan.

[0003] The borehole information obtained through borehole inspection is presented in the form of continuous video. This method acquires a massive amount of data, requiring specialized personnel for analysis and interpretation. When analyzing the internal conditions of the rock mass using borehole inspection technology, technicians need to view the inspection video to identify fractures, faults, and other features within the borehole. However, this method of analyzing borehole video cannot provide a comprehensive and intuitive understanding of the borehole's interior and may lead to misjudgments or omissions.

[0004] Common panoramic processing methods for borehole voyeurism videos typically only involve a simple unfolding of the video. However, during borehole voyeurism, manual operation cannot guarantee a completely uniform and stable insertion speed of the voyeuristic probe into the borehole. Therefore, when there are pauses or positional shifts in the video, ordinary panoramic processing methods will produce overlapping or erroneous images. Furthermore, because borehole voyeurism videos are generally in low brightness, the images extracted by traditional panoramic processing methods are not clear, resulting in low-quality images after panoramic processing. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an intelligent panoramic processing method and system for borehole observation based on image processing. Using the distance the borehole observation probe enters the borehole as a standard, video frames are extracted from the observation video at fixed intervals and saved as images. The images are then processed for sharpening, standardization, and contour recognition. The annular surrounding rock image in the image is transformed into a rectangular image through coordinate transformation and pixel assignment. The rectangular images processed at different distances are then stitched together. Finally, an image duplication region detection algorithm is used to detect and remove duplicate regions between each pair of images, forming a panoramic image of the entire borehole.

[0006] On the one hand, an intelligent panoramic processing method for borehole observation based on image processing is provided, including:

[0007] The borehole inspection video was acquired, and image enhancement processing was performed on each frame of the inspection video. Edge detection processing was performed on the enhanced image to determine the outer and inner radii of the annular surrounding rock.

[0008] Taking a single pixel in the image as a unit, starting from the outer circle of the ring-shaped surrounding rock, the number of pixels is reduced by one unit each time until the inner circle of the ring-shaped surrounding rock is reached. Each reduction yields a circle, and finally, circles with different radii are obtained. Each pixel in each circle is converted from polar coordinates to rectangular coordinates in order from 0 degrees to 360 degrees, thus obtaining the rectangular coordinates of all pixels in each circle.

[0009] Based on the Cartesian coordinates of all pixels in each circle, construct a pixel rectangle corresponding to each circle; the order of the pixels from 0 degrees to 360 degrees in the circle is the order of the pixels from left to right in the pixel rectangle; each circle results in a pixel rectangle.

[0010] The pixel rectangles are stitched together from top to bottom in order of increasing circle radius to obtain a stepped reconstructed image. The stepped reconstructed image is then interpolated to obtain a reconstructed rectangular image. The reconstructed rectangular image is then subjected to duplicate region detection to remove overlapping regions, resulting in a panoramic image of the borehole.

[0011] On the other hand, an intelligent panoramic processing system for borehole observation based on image processing is provided, including:

[0012] The image processing module is configured to: acquire borehole peep video, perform image enhancement processing on each frame of the peep video; perform edge detection processing on the enhanced image to determine the outer and inner radii of the annular surrounding rock;

[0013] The coordinate transformation module is configured to: take a single pixel in the image as a unit, start from the outer circle of the annular surrounding rock, reduce by one unit of pixels each time, until the inner circle of the annular surrounding rock is reached, and obtain a circle for each reduction, and finally obtain circles with different radii; convert each pixel of each circle from polar coordinates to rectangular coordinates in order from 0 degrees to 360 degrees, to obtain the rectangular coordinates of all pixels of each circle;

[0014] The rectangle construction module is configured to: construct a pixel rectangle for each circle based on the Cartesian coordinates of all pixels in each circle; the order of the pixels in the circle from 0 degrees to 360 degrees is the left-to-right pixel order in the pixel rectangle; each circle results in one pixel rectangle.

[0015] The panoramic image reconstruction module is configured to: stitch pixel rectangular bars sequentially from top to bottom according to the circle radius from small to large to obtain a stepped reconstructed image; perform interpolation processing on the stepped reconstructed image to obtain a reconstructed rectangular image; and perform duplicate region detection on the reconstructed rectangular image to remove overlapping regions, thereby obtaining a panoramic image of the borehole.

[0016] The above technical solution has the following advantages or beneficial effects:

[0017] Through image processing technology, coordinate transformation, pixel assignment, and duplicate region detection, borehole observation videos can be processed efficiently and intelligently to generate high-quality panoramic images, providing strong technical support for geological exploration in tunnel engineering. Attached Figure Description

[0018] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0019] Figure 1 This is a schematic diagram of borehole inspection in Example 1;

[0020] Figure 2 This is a schematic diagram of the annular surrounding rock in Example 1;

[0021] Figure 3 This is a schematic diagram of the image reconstruction process in Example 1;

[0022] Figures 4(a) and 4(b) are schematic diagrams of image repetition detection in Example 1;

[0023] The components include: 1. cable; 2. push rod; 3. limit roller; 4. probe; and 5. drill hole. Detailed Implementation

[0024] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0025] Example 1: This example provides an intelligent panoramic processing method for borehole observation based on image processing, including:

[0026] S101: Acquire borehole inspection video, perform image enhancement processing on each frame of the inspection video; perform edge detection processing on the enhanced image to determine the outer and inner radii of the annular surrounding rock;

[0027] S102: Taking a single pixel in the image as a unit, starting from the outer circle of the ring-shaped surrounding rock, reduce by one unit of pixels each time until the inner circle of the ring-shaped surrounding rock is reached. Each reduction yields a circle, and finally, circles with different radii are obtained. Convert each pixel of each circle from polar coordinates to rectangular coordinates in order from 0 degrees to 360 degrees to obtain the rectangular coordinates of all pixels in each circle.

[0028] S103: Based on the Cartesian coordinates of all pixels in each circle, construct a pixel rectangle corresponding to each circle; the order of the pixels from 0 degrees to 360 degrees in the circle is the order of the pixels from left to right in the pixel rectangle; each circle yields a pixel rectangle;

[0029] S104: The pixel rectangles are stitched together from top to bottom in order of increasing circle radius to obtain a stepped reconstructed image; the stepped reconstructed image is interpolated to obtain a reconstructed rectangular image; the reconstructed rectangular image is subjected to duplicate region detection to remove overlapping regions, resulting in a panoramic image of the borehole.

[0030] Furthermore, such as Figure 1 As shown, S101: acquiring borehole endoscopic video includes: installing a housing at the front end of the push rod 2, installing a probe 4 at the front end of the housing, installing limit rollers 3 on both sides of the housing, pushing the push rod 2 to push the housing forward at a constant speed in the borehole 5 under the action of external force, the limit rollers assisting the housing to be in the center position of the borehole, and the camera on the top of the head to collect video during the forward movement. The camera transmits the collected video to the host computer through cable 1.

[0031] Further, S101: performing image enhancement processing on each frame of the voyeuristic video, including:

[0032] Treat the original image as I1(x) i ,y i ), expressed by the following formula:

[0033] I1(x i ,y i )=L1(x i ,y i )*R1(x i ,y i );

[0034] Where R1(x) i ,y i L1(x) represents the reflection component; i ,y i ) represents the illumination component, and * represents the convolution operation.

[0035] New illumination components are obtained by filtering the original image using a two-dimensional Gaussian function.

[0036]

[0037] Wherein, L2(x i ,y i ) represents the filtered illumination component, and σ represents the standard deviation of the Gaussian function.

[0038] By taking multiple different standard deviations σ in the Gaussian function, the multi-scale reflection components R can be obtained. MSR (x i ,y i ):

[0039]

[0040] Among them, w j For weights at different scales, w j for Considering the high, medium, and low scales, σ takes three values: 64, 128, and 256.

[0041] By transformation, the reflection component R2(x) is made... i ,y i The pixel values ​​are in the range of 0 to 255:

[0042]

[0043] Thus, the enhanced image I2(x,y) can be further obtained:

[0044] ln(I2(x i ,y i ))=ln(L2(x i ,y i ))+ln(R2(x i ,y i )).

[0045] It should be understood that low-light images extracted from voyeuristic videos are enhanced to improve their clarity and to identify black circular cavities in borehole images.

[0046] Further, step S101: performing edge detection processing on the enhanced image to determine the outer and inner radii of the annular surrounding rock, including:

[0047] The enhanced image I2(x,y) is converted to grayscale and transformed into a grayscale image I_gray(x,y);

[0048] Calculate the horizontal and vertical gradients of the grayscale image, and then calculate the gradient magnitude value for each pixel based on the horizontal and vertical gradients.

[0049] Based on the gradient magnitude value of each pixel, determine the contour edge points; based on the contour edge points, determine the outer and inner radii of the ring-shaped surrounding rock.

[0050] Further, the calculation of the horizontal and vertical gradients of the grayscale image, and the calculation of the gradient magnitude value for each pixel based on the horizontal and vertical gradients, include:

[0051] Using the Sobel operator, calculate the horizontal gradient G of the grayscale image I_gray(x,y). x (x,y) and vertical gradient G y (x,y):

[0052]

[0053] Among them, S x (i,j) and S y (i,j) is the convolution kernel of the Sobel operator:

[0054]

[0055] Calculate the gradient magnitude G(x,y) for each pixel in the image:

[0056]

[0057] Further, the step of determining the contour edge points based on the gradient magnitude value of each pixel, and determining the outer and inner radii of the annular surrounding rock based on the contour edge points, includes:

[0058] Set a threshold to mark pixels above the threshold as contour edge points E(x,y):

[0059]

[0060] Where θ is a threshold, determined through manual optimization;

[0061] Contour E(x) i ,y j For a circular cavity, calculate the interpolated coordinates at different x-points on the same y-coordinate, and take the maximum value as the radius of the circular cavity:

[0062] r = max|E(x) i ,y j )-E(x i ,y j+n )|;

[0063] In subsequent image processing, the circular cavity region is ignored, and only the annular surrounding rock portion is processed, where the inner radius r of the annular surrounding rock is... j =r+1; for example Figure 2 As shown;

[0064] All processed images are standardized by translation and cropping to make them all squares of the same size (the number of pixels on each side must be odd). The center point of the square is the inner circle of the annular rock, and the center of the circular cavity in the image is ensured to be located in the center of the image to avoid obtaining null values ​​during the subsequent polar coordinate to rectangular coordinate conversion. Half the side length of the square is used as the outer radius r0 of the annular rock.

[0065] Further, S102: Convert each pixel on the circles of different radii from polar coordinates to rectangular coordinates in order from 0° to 360°, to obtain the rectangular coordinates of all pixels on the circles of different radii, including:

[0066] Treating a single pixel in the image as a unit, with the image center (x) r ,y r Using the origin as the origin and the radius r0 of the outer ring of the surrounding rock as the polar coordinate radius, the radius increases by 1° each time within the range of 0-360°. Using the conversion formula between polar and rectangular coordinates, the rectangular coordinates of all pixels located within the pixel ring of radius r0 are calculated, and the pixel value of that point is obtained based on its coordinates. Conversion formula:

[0067] x = x r +r0cos(α)

[0068] y = y r +r0sin(α).

[0069] If the values ​​of x and y are not integers, they will be rounded down.

[0070] Furthermore, S103: such as Figure 3 As shown, based on the Cartesian coordinates of all pixels in each circle, a pixel rectangle is constructed for each circle; the order of the pixels from 0° to 360° within the circle is the left-to-right pixel order within the pixel rectangle; each circle yields a pixel rectangle, including:

[0071] The pixels are horizontally arranged in the order of acquisition from 0 to 360°, with coordinates (0, 0), (1, 0)...(1, 0), and assigned corresponding pixel values.

[0072] Starting from the outer radius r0 of the annular surrounding rock, the process of repeatedly acquiring and horizontally rearranging pixels by decreasing by one unit each time is repeated. When the radius is r i At this point, the rearranged coordinates are: (0, i), (1, i)...(m, i), until they are reduced to the inner radius r of the annular surrounding rock. j Up to this point, the rearranged coordinates are: (0, n), (1, n)...(n, j).

[0073] Further, S104: Splice the pixel rectangular bars in order from the smallest to the largest circle radius, from top to bottom, to obtain a stepped reconstructed image; perform interpolation processing on the stepped reconstructed image to obtain a reconstructed rectangular image, including:

[0074] In the annular surrounding rock area, the radius r j < r i < r0, according to the circumference calculation formula of the circle, it can be obtained that: n < m < l. Therefore, the image constructed according to this result is a stepped reconstructed image. To construct a rectangular image, interpolation processing needs to be performed on the stepped reconstructed image, and pixel values are assigned to the interpolated pixel points;

[0075] Based on the number of pixels l + 1 obtained from the pixel circle with a radius of r0, perform interpolation on other rows with the number of pixels less than l + 1. For example, when the radius is r i The interpolation number and interpolation interval of the pixel circle are:

[0076] a i = l - m;

[0077] Among them, a i is the interpolation number, rounded to an integer; b i is the interpolation interval, and m is the total number of pixel points obtained from the pixel circle with a radius of r <00​​​​​​​​​​​​​​​​​​​​​​​​​​​The image is reconstructed based on the rearranged pixel coordinates and pixel values, arranged in the order extracted from the video. Starting from the first image, the overlapping parts between every two adjacent images are detected and removed. Then all images are stitched together to form a panoramic image of the entire borehole.

[0085] Further, as shown in Figures 4(a) and 4(b), the step of detecting and removing the overlapping portion between every two adjacent images includes:

[0086] In the order of image extraction from the drilling video, S1 was captured earlier than S2;

[0087] The reconstructed image is converted to grayscale, and the image S1 is converted to grayscale using b. j The overlapping portion in row S2 is retrieved. Starting from row d0 in S2, the overlap between each pixel in row d0 and b is calculated. j The interpolation ΔG for each pixel in the row i×0 :

[0088] ΔG i×0 =|G(a i ,b j )-G(c i ,d0)|;

[0089] Among them, G(a i ,b j ) represents each pixel (a i ,b j The grayscale value of ); i∈[0,n];

[0090] When pixel (a) i ,b j ) and pixels (c i Interpolation ΔG between d0) i×0 When the value is less than 20, the pixel (a) is determined. i ,b j ) and pixels (c i Let d0 be a similar point, and calculate b. j The similarity X between row d0 and row d0 i×0 :

[0091]

[0092] Where p is the number of similar points in row d0;

[0093] If S1's b j The similarity X between row d0 and row d0 in S2 i×0 If the percentage is less than 95%, it is determined to be non-repeating, and the calculation of b for S1 continues. j The repetition rate X between row d1 and row d1 in S2 i×1Determine if there is a repetition until the first detection of b in S1. j row and a certain row d in S2 c The repetition rate X between i×c When it reaches 95%, it is determined that S2 is from d. c line to line d0 (d c d c-1 d c-2 If , ..., d0) belong to the repeating region, delete the repeating region, and then concatenate S1 and S2. If b of S1 j If no row in S1 is repeated from any row in S2, then S1 and S2 are concatenated directly.

[0094] Example 2

[0095] This embodiment provides an intelligent panoramic processing system for borehole observation based on image processing, including:

[0096] The image processing module is configured to: acquire borehole peep video, perform image enhancement processing on each frame of the peep video; perform edge detection processing on the enhanced image to determine the outer and inner radii of the annular surrounding rock;

[0097] The coordinate transformation module is configured to: take a single pixel in the image as a unit, start from the outer circle of the annular surrounding rock, reduce by one unit of pixels each time, until the inner circle of the annular surrounding rock is reached, and obtain a circle for each reduction, and finally obtain circles with different radii; convert each pixel of each circle from polar coordinates to rectangular coordinates in order from 0 degrees to 360 degrees, to obtain the rectangular coordinates of all pixels of each circle;

[0098] The rectangle construction module is configured to: construct a pixel rectangle for each circle based on the Cartesian coordinates of all pixels in each circle; the order of the pixels in the circle from 0 degrees to 360 degrees is the left-to-right pixel order in the pixel rectangle; each circle results in one pixel rectangle.

[0099] The panoramic image reconstruction module is configured to: stitch pixel rectangular bars sequentially from top to bottom according to the circle radius from small to large to obtain a stepped reconstructed image; perform interpolation processing on the stepped reconstructed image to obtain a reconstructed rectangular image; and perform duplicate region detection on the reconstructed rectangular image to remove overlapping regions, thereby obtaining a panoramic image of the borehole.

[0100] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for intelligent panoramic processing of a borehole peep based on image processing, characterized in that, The method comprises the following steps: obtaining a borehole endoscopic video, and performing image enhancement processing on each frame image in the endoscopic video; performing edge detection processing on the enhanced image to determine the outer circle radius and the inner circle radius of the annular surrounding rock; taking a single pixel point in the image as a unit, starting from the outer circle of the annular surrounding rock, and reducing the pixel point by one unit each time until the inner circle of the annular surrounding rock is reached, obtaining a circle each time the pixel point is reduced, and finally obtaining circles with different radii; converting each pixel point of the circles with different radii into rectangular coordinates in the order from 0 degrees to 360 degrees to obtain the rectangular coordinates of all pixel points of the circles with different radii; constructing a pixel rectangle corresponding to each circle according to the rectangular coordinates of all pixel points of each circle; the order of the pixel points in the circle from 0 degrees to 360 degrees is the order of the pixel points from left to right in the pixel rectangle; each circle obtains a pixel rectangle; splicing the pixel rectangles in the order from small to large according to the circle radii from top to bottom to obtain a stepped reconstruction image; performing interpolation processing on the stepped reconstruction image to obtain a reconstructed rectangular image; performing repeated area detection on the reconstructed rectangular image to remove the overlapping area to obtain a panoramic image of the borehole; performing image enhancement processing on each frame image in the endoscopic video, including: The original image is regarded as is represented by the following equation: ； wherein, is a reflection component; is an illumination component, is a convolution operation; performing filtering processing on the original image through a two-dimensional Gaussian function to obtain a new light component: ； wherein, is the filtered illumination component, is the standard deviation of the Gaussian function; By taking multiple different standard deviation values in the Gaussian function one obtains a multi-scale reflection component : ； wherein, is a weight of different scales, is ; considering high, medium, and low three scales, Here take 64, 128, 256 three values; By converting the reflected component of the pixel value to be in the range 0-255: ； Thus further enhanced image is obtained : 。 2. The image processing-based borehole peeping intelligent panoramic processing method according to claim 1, characterized in that, performing edge detection processing on the enhanced image to determine the outer circle radius and the inner circle radius of the annular surrounding rock, including: the enhanced image gray scale processing is performed to convert the image into a gray scale image ; calculating the horizontal gradient and the vertical gradient of the gray image, calculating the gradient amplitude value of each pixel point according to the horizontal gradient and the vertical gradient; determining the contour edge point according to the gradient amplitude value of each pixel point; determining the outer circle radius and the inner circle radius of the annular surrounding rock according to the contour edge point.

3. The image processing-based borehole peeping intelligent panoramic processing method according to claim 2, characterized in that, The calculation of the horizontal gradient and the vertical gradient of the gray image, and the calculation of the gradient amplitude value of each pixel point according to the horizontal gradient and the vertical gradient, include: Using the Sobel operator, the horizontal gradient and the vertical gradient of the gray-scale image are calculated : ； ； wherein and is a convolution kernel of the Sobel operator: ； calculating a gradient magnitude value for each pixel in the image : 。 4. The image processing-based borehole peeping intelligent panoramic processing method according to claim 2, characterized in that, determining the contour edge point according to the gradient amplitude value of each pixel point; determining the outer circle radius and the inner circle radius of the annular surrounding rock according to the contour edge point, including: a threshold is set, and pixel points higher than the threshold are marked as contour edge points : ； wherein is a threshold value determined by means of manual optimization; contour is circular, the maximum of the interpolated coordinates at the same coordinates is taken as the radius of the circular cavity: coordinates is taken as the radius of the circular cavity: ； wherein the inner radius of the annular surrounding rock is ; All the processed images are normalized, and all the images are made into consistent square by translation and clipping, and the center point of the square is the dot of the inner circle of the ring-shaped surrounding rock, and half of the side length of the square is taken as the outer circle radius of the ring-shaped surrounding rock .

5. The image processing-based intelligent boreview panoramic processing method according to claim 1, wherein, converting each pixel point of the circles with different radii into rectangular coordinates in the order from 0 degrees to 360 degrees to obtain the rectangular coordinates of all pixel points of the circles with different radii, including: Take a single pixel point in the image as a unit, take the center of the image as the origin, take the outer circle radius of the ring-shaped surrounding rock as the polar radius, and take each increase of 1° in the range of 0-360° as a step, calculate the rectangular coordinates of all pixel points on the pixel circle with a radius of , and obtain the pixel value of the point according to the coordinates. The conversion formula is:​ ； wherein x , y If the value is not an integer, it is rounded.

6. The image processing-based bore viewing intelligent panoramic processing method according to claim 1, characterized in that, constructing a pixel rectangle corresponding to each circle according to the rectangular coordinates of all pixel points of each circle; the order of the pixel points in the circle from 0 degrees to 360 degrees is the order of the pixel points from left to right in the pixel rectangle; each circle obtains a pixel rectangle, including: The pixel points are horizontally arranged according to the acquisition order from 0-360°, and the coordinates are (0, 0), (1, 0)…(360, 0), and the corresponding pixel values are assigned to them. l , 0), and the corresponding pixel values are assigned to them. from the outer radius of the annular surrounding rock Start each decrease of one unit of repeated pixel acquisition and horizontal arrangement of reorganization process, when the radius is , the rearranged coordinates are: (0, i ), (1, i )… (m, i ), until the reduction to the inner radius of the annular surrounding rock , at this time the rearranged coordinates are: (0, n ), (1, n )… (m, n , j ).

7. The image processing-based bore viewing intelligent panoramic processing method according to claim 1, characterized in that, splicing the pixel rectangles in the order from small to large according to the circle radii from top to bottom to obtain a stepped reconstruction image; performing interpolation processing on the stepped reconstruction image to obtain a reconstructed rectangular image, including: In the annular surrounding rock area, the radius According to the circumference formula of the circle, we can get: n m Therefore, the image constructed according to this result is a stepped reconstruction image, and interpolation processing is required for the stepped reconstruction image to construct a rectangular image, and pixel assignment is required for the interpolated pixel points.​ The pixel number obtained by the pixel circle with a radius of The pixel number obtained by the pixel circle with a radius of The pixel number obtained by the pixel circle with a radius of The pixel number obtained by the pixel circle with a radius of The pixel number obtained by the pixel circle with a radius of ； ； wherein, is the number of interpolation, rounded to the nearest integer; is the interpolation interval, is the total number of pixels obtained by the pixel ring with a radius of pixels. From the starting point Start interpolation for each interval of pixels, if the number of pixels after interpolation is not full after interpolation, then directly interpolate after the original point the pixel value at the interpolation pixel point is the average value of the pixel values of the four pixel points on the top, bottom, left and right sides: ； According to the interpolated pixel point coordinates and pixel values, the image can be reconstructed, and the size of the image is .

8. The image processing-based bore viewing intelligent panoramic processing method according to claim 1, characterized in that, the repeated area detection on the reconstructed rectangular image to remove the overlapping area to obtain a panoramic image of the borehole, including: According to the rearranged pixel point coordinates and pixel values, the image is reconstructed, and the images are arranged according to the extraction order in the video, starting from the first image, the overlapping part between every two adjacent images is detected and removed, and then all the images are spliced to form a panoramic image of the entire borehole; The detection of the overlapping part between every two adjacent images and the removal thereof include: In the order of extracting images in the borehole video, the S1 shooting time is earlier than the S2; The reconstructed image is subjected to a gray-scale processing, and the overlapping part in the line search S2 is searched from the line in S2, and the interpolation value of each pixel point in the line is calculated from each pixel point in the line : ； wherein represents the gray value of each pixel point ; ; When pixel and pixels Interpolation between When the value is less than 20, the pixel is determined. and pixels For similar points, calculate Action and Row similarity : ； wherein is the number of similar points in the row; If S1 Rows in S2 Row similarity If the percentage is less than 95%, it is considered non-repeating, and the calculation of S1 continues. row and S2 Repetition rate between lines X i×1 Determine if there is a duplicate until S1 is detected for the first time. row and a certain row in S2 d c Repetition rate between X i×c When it reaches 95%, it is determined that S2 is from d c Arrive If the row belongs to a duplicate region, delete the duplicate region, then concatenate S1 and S2. If S1... If no row in S1 is repeated from any row in S2, then S1 and S2 are concatenated directly.

9. An image processing-based intelligent panoramic processing system for borehole peeping, characterized in that, The image processing module is configured to: acquire a borehole endoscopic video, and perform image enhancement processing on each frame of image in the endoscopic video; Edge detection processing is performed on the enhanced image to determine the outer circle radius and the inner circle radius of the annular surrounding rock; the image enhancement processing on each frame of image in the endoscopic video includes: A new light component is obtained by performing filtering processing on the original image based on a two-dimensional Gaussian function: The original image is regarded as is represented by the following equation: ； wherein, is a reflection component; is an illumination component, is a convolution operation; The coordinate conversion module is configured to: take a single pixel point in the image as a unit, start from the outer circle of the annular surrounding rock, and reduce one unit of pixel point at a time until the inner circle of the annular surrounding rock is reached, obtaining a circle each time, and finally obtaining circles with different radii; each pixel point of each circle is converted from polar coordinates to rectangular coordinates in order from 0 degrees to 360 degrees, to obtain the rectangular coordinates of all pixel points of each circle; ； wherein, is the filtered illumination component, is the standard deviation of the Gaussian function; By taking multiple different standard deviation values in the Gaussian function one obtains a multiscale reflection component : ； wherein, are weights of different scales, are ; considering high, medium and low three scales, Here take 64, 128, 256 three values; By converting the reflected component of the pixel value to be in the range 0-255: ； to further obtain an enhanced image : ； The rectangular strip construction module is configured to: construct a pixel rectangular strip corresponding to each circle according to the rectangular coordinates of all pixel points of each circle; the order of 0 degrees to 360 degrees of the pixel points in the circle is the order of the pixel points from left to right in the pixel rectangular strip; each circle obtains a pixel rectangular strip; The panoramic image reconstruction module is configured to: splice the pixel rectangular strips in order from small to large according to the circle radii, from top to bottom, to obtain a stepped reconstructed image; perform interpolation processing on the stepped reconstructed image to obtain a reconstructed rectangular image; and perform repeated area detection on the reconstructed rectangular image to remove the overlapping area, to obtain a panoramic image of the borehole. ​

Images