CN109584259B

CN109584259B - Quartz crucible bubble layered counting device and method

Info

Publication number: CN109584259B
Application number: CN201910108532.XA
Authority: CN
Inventors: 赵谦; 史凌云; 李蓉蓉; 李小龙; 唐振华
Original assignee: Individual
Current assignee: Zhao Qian
Priority date: 2019-01-18
Filing date: 2019-01-18
Publication date: 2021-10-01
Anticipated expiration: 2039-01-18
Also published as: CN109584259A

Abstract

The invention discloses a quartz crucible bubble layered counting device and a quartz crucible bubble layered counting method, wherein the device comprises a data processing controller, a micro projector and an image acquisition device, the image acquisition device comprises a mechanical arm and an electronic magnifier, and light projected by the micro projector is blue-violet light; the method comprises the following steps: firstly, collecting and transmitting quartz crucible images; secondly, denoising and balancing the amplified image of the quartz crucible; thirdly, layering quartz crucible bubble images; and fourthly, acquiring bubble outlines of the quartz crucible bubble layered images. The quartz crucible inspection device is reasonable in design, bubble images of the transparent layer in the quartz crucible are obtained through the electronic magnifier, and the bubble images are subjected to layered processing, so that whether the production process of the quartz crucible reaches the standard or not is judged, manpower, material resources and financial resources are saved, and the automation degree of quartz crucible inspection is improved.

Description

Quartz crucible bubble layered counting device and method

Technical Field

The invention belongs to the technical field of quartz crucible bubble detection, and particularly relates to a quartz crucible bubble layered counting device and method.

Background

Quartz crucibles are widely used in many important fields, particularly in the production of silicon single crystals as materials for the production of semiconductor electronic components, where quartz crucibles are the main equipment, and where bubbles in the transparent layer of quartz crucibles greatly affect the pulling result, it is very important to select experimental quartz crucibles with a small number of bubbles. The cross section of a quartz crucible generally has a double-layer structure: the inner layer is colorless and is called as a transparent layer, and the transparent layer is characterized in that the transparent layer has less bubble content and the bubbles are spherical and uniformly distributed; the outer layer is white, called opaque layer, and is characterized by being opaque, containing a large number of bubbles, and the composite bubbles are dense and obviously larger than the transparent layer. In the cross section of the quartz crucible, a particularly obvious boundary line between the two layers can be obviously observed, and the obvious difference of bubbles between the opaque layer and the transparent layer can be observed by using a magnifier, so that the bubbles are changed from large bubbles in the opaque layer to small bubbles in the transparent layer.

The quartz crucible used in the conventional crystal pulling method is generally a quartz crucible. During crystal growth, prolonged exposure of the quartz crucible interior sidewall to the high temperature silicon melt results in reaction of the silicon melt with the quartz crucible and dissolution of the inner surface of the quartz crucible sidewall. This exposes the bubbles in the side walls of the quartz crucible, with the result that the silicon melt continues to dissolve into the walls of the quartz crucible and thus into the walls of the bubbles. At some point, the walls of the bubbles are broken and the silicon melt may sink into the walls, releasing gas from the bubble interior and releasing quartz particles from the quartz crucible and or bubble sidewalls into the melt. In doing so, the quartz particles can disrupt the single crystal structure, thereby limiting the yield of single crystal grown from the crystal. At present, in the production of quartz crucibles, the detection work of the quartz crucibles is an extremely important link, and meanwhile, the detection direction of the quartz crucibles starts to develop towards the field detection. At present, people mainly check and judge detection results manually, but are easily influenced by various factors such as subjective quality, personal emotion and vision. The manual detection not only takes long time, but also is easy to make mistakes, and has great influence on the accuracy of the test result. In the market at present, the demand of the quartz crucible is huge, and meanwhile, the development mode of the factory in China is converted to the automation direction, so that the change of the detection mode of the quartz crucible is imperative.

Disclosure of Invention

The invention aims to solve the technical problem that the defects in the prior art are overcome, and the quartz crucible bubble layering counting device is reasonable in design and low in cost, obtains the bubble image of the transparent layer in the quartz crucible through the electronic magnifier, layers the bubble image, and is convenient for judging whether the production process of the quartz crucible meets the requirements, so that the device can save manpower, material resources and financial resources, improve the production benefit of manufacturers, avoid errors caused by human factors, reduce the error rate in manual detection, improve the automation degree of quartz crucible detection, improve the objectivity and the accuracy of the detection result of the quartz crucible, and further improve the detection accuracy of the quartz crucible.

In order to solve the technical problems, the invention adopts the technical scheme that: the utility model provides a quartz crucible bubble layering counting assembly which characterized in that: the image acquisition device comprises a mechanical arm and an electronic magnifier clamped on the mechanical arm, wherein the electronic magnifier and the mechanical arm are connected with the data processing controller, and light projected by the miniature projector is blue-violet light.

The quartz crucible bubble layering counting device is characterized in that: the data processing controller is a computer.

The quartz crucible bubble layering counting device is characterized in that: the mechanical arm is a six-degree-of-freedom mechanical arm, the magnification range of the electronic magnifier is 1-500 times, the electronic magnifier is connected with the data processing controller through a USB transmission line, and the mechanical arm is controlled by the data processing controller.

Meanwhile, the invention also discloses a quartz crucible bubble layered counting method which has the advantages of simple method steps, reasonable design, convenient realization, high detection accuracy and good use effect, and is characterized by comprising the following steps:

step one, collecting and transmitting quartz crucible images:

step 101, a data processing controller controls a mechanical arm to rotate, and the mechanical arm rotates to drive an electronic magnifier to move, so that the vertical central line of the electronic magnifier is perpendicular to the inner wall of a quartz crucible to be measured; a micro projector is arranged on the side edge of the quartz crucible to be measured, and horizontally projects blue-violet light to the edge opening of the quartz crucible to be measured;

step 102, manually adjusting the electronic magnifier to focus until bubbles on the inner wall of the quartz crucible can be clearly observed by people;

103, the data processing controller controls the electronic magnifier to shoot bubble images of the inner wall of the quartz crucible to be detected, and sends the shot bubble amplified images of the inner wall of the quartz crucible to the data processing controller; wherein, the magnification of the electronic magnifier is not less than 30 times;

step two, denoising and balancing the quartz crucible amplified image:

step 201, calling a gray processing module by using the data processing controller, and performing gray processing on the amplified bubble image on the inner wall of the quartz crucible to obtain a gray image of the bubbles in the quartz crucible;

step 202, adopting the data processing controller to call a weighted mean filtering module to filter the quartz crucible bubble gray level image to obtain a denoised quartz crucible bubble gray level image;

step 203, calling a gray value balancing module by using the data processing controller, and carrying out gray value balancing on the denoised quartz crucible bubble gray image to obtain a quartz crucible bubble balancing image;

step three, layering quartz crucible bubble images:

step 301, calling an Otsu algorithm module by using the data processing controller, and performing primary image segmentation on the quartz crucible bubble equalization image to obtain a first layer of bubble image; wherein the gray value threshold value of the primary image segmentation is 122-195;

step 302, calling an Otsu algorithm module by using the data processing controller, and performing secondary image segmentation on the quartz crucible bubble equalization image to obtain a second layer bubble image; wherein the gray value threshold value of the secondary image segmentation is 196-230;

step 303, calling an Otsu algorithm module by using the data processing controller, and performing image segmentation on the quartz crucible bubble equalization image for three times to obtain a third layer of bubble image; wherein the gray value threshold value of the three-time image segmentation is 231-255; wherein the bubble areas in the first layer of bubble image, the second layer of bubble image and the third layer of bubble image are white areas, and the background areas in the first layer of bubble image, the second layer of bubble image and the third layer of bubble image are black areas;

step four, acquiring bubble outlines of the quartz crucible bubble layered images: the method for acquiring the bubble outlines in the first layer of bubble image, the second layer of bubble image and the third layer of bubble image by adopting the data processing controller is the same, and the method for acquiring the bubble outlines in the l-th layer of bubble image by adopting the data processing controller is as follows:

step 401, calling a binary image edge tracking filtering algorithm module by using the data processing controller, and performing edge extraction on the l-th layer of bubble image to obtain an l-th layer of bubble edge contour image;

step 402, adopting the data processing controller to call a morphological algorithm to perform open operation, and performing open operation processing on the l-th layer bubble edge outline image to obtain an l-th layer bubble outline image after the open operation; wherein l is a positive integer, and the values of l are 1, 2 and 3.

The above method is characterized in that: after the first-layer bubble outline image after the opening operation is obtained in the fourth step, judging the non-overlapping area and the overlapping area of the bubbles in the fifth step, wherein the judging methods of the non-overlapping area and the overlapping area of the bubbles in the first-layer bubble image, the second-layer bubble image and the third-layer bubble image are the same by using the data processing controller, and the non-overlapping area and the overlapping area of the bubbles in the first-layer bubble image are judged by using the data processing controller, which is specifically as follows:

step 501, marking bubble outline connected areas on the first layer of bubble images by adopting a data processing controller and a region growing method to obtain a plurality of bubble outline connected areas; wherein the number of the bubble profile communication areas is f_l，f₁Representing the number of connected areas of the bubble contours in the first layer bubble image, f₂Representing the number of connected areas of the bubble profile in the second layer bubble image, f₃Representing the number of bubble outline connected areas in the third layer of bubble images;

step 502, adopting the data processing controller to call a minimum rectangle extraction module, and processing an nth bubble outline connected region to obtain a minimum external rectangle of the nth bubble outline connected region; wherein the minimum bounding rectangleOne side length is parallel to the row direction of the first layer of bubble image, the second layer of bubble image or the third layer of bubble image, n is a positive integer, and the value range of n is 1-f_l，f_lIs a positive integer not less than 1;

step 503, calling a pixel coordinate extraction module by using the data processing controller to obtain pixel coordinates of an upper left corner pixel point, an upper right corner pixel point, a lower left corner pixel point and a lower right corner pixel point of the minimum circumscribed rectangle; adopting the data processing controller to call a pixel coordinate conversion image coordinate module, and converting the pixel coordinates of the upper left corner pixel point, the upper right corner pixel point, the lower left corner pixel point and the lower right corner pixel point of the minimum external rectangle to obtain the image coordinate of the upper left corner pixel point of the minimum external rectangle

Image coordinates of upper right corner pixel

Image coordinates of lower left corner pixel points

And image coordinates of lower right corner pixel points

Step 504, the data processing controller is according to formula

Obtaining the circularity E of the n-th bubble outline communication area_n；

Step 505, judging 0.5 by using the data processing controller<E_nIf it is not more than 1, when it is 0.5<E_nWhen the bubble outline is not more than 1, the bubble outline communication area is a non-overlapped bubble outline; otherwise, when 0.5<E_nWhen the bubble outline is not more than 1, the bubble outline communication area is an overlapped bubble outline;

step 506, addRepeating the steps 501 to 505 until f in the first layer bubble image is completed_lJudging the connected area of the bubble outlines to obtain an overlapped bubble outline and a non-overlapped bubble outline in the first layer of bubble image; wherein the number of overlapped bubble outlines in the first layer bubble image is C_dlThe number of non-overlapping bubble contours in the first layer bubble image is F_cl；F_clAnd C_dlAre all positive integers.

The above method is characterized in that: after the judgment of the non-overlapping area and the overlapping area of the bubbles is completed, acquiring the number of the bubbles in the sixth step, wherein the method for acquiring the number of the bubbles in the first layer of bubble image, the method for acquiring the number of the bubbles in the second layer of bubble image and the method for acquiring the number of the bubbles in the l layer of bubble image by using the data processing controller are the same, and the method for acquiring the number of the bubbles in the l layer of bubble image by using the data processing controller is specifically as follows:

step 601, adopting the data processing controller to call a pixel coordinate extraction module to extract pixel coordinates of each pixel point on the mth overlapped bubble outline in the first-layer bubble image to obtain the pixel coordinates of each pixel point on the mth overlapped bubble outline in the first-layer bubble image; wherein, the pixel coordinate of the ith pixel point on the mth overlapped bubble outline in the ith layer of bubble image is P_lm,i(u_lm,i,v_lm,i) (ii) a The mth overlapped bubble profile in the ith layer of bubble image comprises N pixel points, i represents the serial number of the pixel points, i is obtained according to the mth overlapped bubble profile in the ith layer of bubble image in the anticlockwise or clockwise direction, m is a positive integer, and the value range of m is 1-C_dlI is a positive integer, and the value range of i is 1-N;

step 602, using the data processing controller according to a formula

Obtaining the centroid pixel coordinate P of the mth overlapped bubble outline in the ith layer of bubble image_lm,z(u_lm,z,v_lm,z)；

Step 603, adoptThe data processing controller is based on a formula

Obtaining the distance L from the ith pixel point to the centroid on the mth overlapped bubble profile in the ith layer of bubble image_lm,i；

Step 604, repeating step 603 for multiple times to obtain the distance from each pixel point on the mth overlapped bubble outline in the ith layer of bubble image to the centroid; adopting the data processing controller to call a curve drawing module, taking the serial numbers of pixel points as an abscissa, taking the distance from the ith pixel point to the centroid on the mth overlapped bubble outline in the ith layer of bubble image as an ordinate, and drawing a distance curve of each pixel point on the mth overlapped bubble outline in the ith layer of bubble image;

605, adopting the data processing controller to invoke a discrete Fourier transform module, and performing discrete Fourier transform on a distance curve of each pixel point on the mth overlapped bubble outline in the ith layer of bubble image to obtain a main value sequence of the mth overlapped bubble outline in the ith layer of bubble image; wherein, the length of the main value sequence of the mth overlapped bubble outline in the ith layer of bubble image is N;

step 606, using the data processing controller to respectively record two adjacent distance values in the main value sequence of the mth overlapped bubble profile in the ith layer of bubble image as y_jAnd y_j+1And is combined with y_jThe serial number of the corresponding pixel point is marked as x_j，y_j+1The serial number of the corresponding pixel point is marked as x_j+1(ii) a Wherein j is a natural number, and j is more than or equal to 0 and less than or equal to N-1;

step 607, the data processing controller calculates the formula

Carrying out interpolation processing on two adjacent distance values in the main value sequence of the mth overlapped bubble profile in the ith layer of bubble image to obtain a plurality of distance difference values; wherein, the number of distance difference value points in the main value sequence of the mth overlapped bubble outline in the first layer of bubble image is 3N-5N, x represents the interpolation position,

representing an interpolated distance;

step 608, repeating step 607 for multiple times, completing interpolation processing of two adjacent distance values in the main value sequence of the mth overlapped bubble profile in the ith layer of bubble image, and fitting the distance difference points by using a data processing controller to obtain a distance curve after interpolation processing;

step 609, the data processing controller calls a Gaussian smoothing module, and Gaussian smoothing is carried out on the distance curve after interpolation processing to obtain a smooth distance curve of the mth overlapped bubble profile in the first layer of bubble image;

step 6010, the data processing controller obtains the number of bubbles in the mth overlapped bubble profile in the lth bubble image according to the smoothed distance curve of the mth overlapped bubble profile in the lth bubble image; a maximum value exists between two adjacent minimum value points on a smooth distance curve of the mth overlapped bubble profile in the ith layer of bubble image, and a profile surrounded by pixel points corresponding to the two adjacent minimum value points on the smooth distance curve of the mth overlapped bubble profile in the ith layer of bubble image is a bubble profile;

step 6011, repeating the steps 601 to 6010 for multiple times to obtain C in the first-layer bubble image_dlThe number of bubbles in each overlapping bubble profile;

step 6012, the data processing controller invokes the addition module to set the number F of non-overlapping bubbles in the first layer of bubble image_clAdding the number of the bubbles in the outline of the overlapped bubbles in the first layer of bubble image to obtain the total number Z of the quartz crucible bubbles in the first layer of bubble image_lj(ii) a And the number of non-overlapping bubble outlines in the first-layer bubble image is the number of non-overlapping bubbles in the first-layer bubble image.

The above method is characterized in that: in step 203, the data processing controller is adopted to call a gray value balancing module to perform gray value balancing on the denoised quartz crucible bubble gray image, and the specific process is as follows:

step 2031, calling a pixel value extraction module by using the data processing controller, extracting pixel values of all pixel points in a bubble area in the de-noised quartz crucible bubble gray level image, and acquiring the pixel values of all pixel points in the bubble area in the de-noised quartz crucible bubble gray level image; the number of pixel points in a bubble area in the denoised quartz crucible bubble gray level image is m₁And the pixel value of the ith pixel point in the bubble area in the denoised quartz crucible bubble gray level image is Pl_i′I' and m₁Are all positive integers, and the value range of i' is more than or equal to 1 and less than or equal to m₁；

Step 2032, sorting the pixel values of all pixel points in the bubble area in the de-noised quartz crucible bubble gray scale image according to a sequence from small to large by adopting the data processing controller to obtain a minimum pixel value and a maximum pixel value, and assigning the maximum pixel value to Pl_maxAssigning the minimum pixel value to Pl_min；

Step 2033, using the data processing controller according to formula

Obtaining the average value of pixel values of all pixel points in the bubble area in the denoised quartz crucible bubble gray level image;

step 2034, assigning the average value of the pixel values of each pixel point in the bubble area in the de-noised quartz crucible bubble gray level image to each pixel point in the bubble area in the de-noised quartz crucible bubble gray level image by using the data processing controller, so as to update and balance the pixel values of each pixel point in the bubble area in the de-noised quartz crucible bubble gray level image and complete gray level balance of the de-noised quartz crucible bubble gray level image;

step 2035, repeating the steps 2031 to 2032 for a plurality of times until the gray value balance of a plurality of bubble areas in the de-noised quartz crucible bubble gray image is completed, and obtaining a quartz crucible bubble balance image.

The above method is characterized in that: in step 401, the data processing controller is adopted to call a binary image edge tracking filter algorithm module, edge extraction is carried out on a quartz crucible bubble binary image, and a bubble edge contour image is obtained, wherein the specific process is as follows:

step 4011, establishing a background image by using the data processing controller, wherein the size of the background image is the same as that of the quartz crucible bubble binary image, and the gray value of each pixel point in the background image is zero;

step 4012, scanning the quartz crucible bubble binary image from top to bottom by using the data processor, and recording a pixel coordinate (u) of a 255 pixel point of a first gray value when the data processor obtains the 255 pixel point of the first gray value₀,v₀) And the pixel coordinates (u) in the background image are determined₀,v₀) The pixel value is set to 255;

step 4013, using said data processor to calculate pixel coordinates (u)₀,v₀) The eight neighborhood pixels which are taken as the center are scanned in turn according to the anticlockwise direction, and the pixel coordinate (u) is used₀,v₀) When a pixel with the gray value of 255 exists in the eight neighborhood pixels which are taken as the center, the pixel coordinate (u) of the pixel with the second gray value of 255 is recorded₁,v₁) And the pixel coordinates (u) in the background image are determined₁,v₁) The pixel value is set to 255;

step 4014, repeat the method of step 4013 with the pixel coordinate (u)₁,v₁) The eight neighborhood pixels which are taken as the center are scanned in turn counterclockwise to obtain the pixel coordinate of the next pixel with the gray value of 255, and the pixel value of the coordinate, which is the same as the pixel coordinate of the next pixel with the gray value of 255, in the background image is set to be 255;

step 4015, calculating pixel coordinates (u)_r,v_r) Scanning the eight neighborhood pixels as the center in turn counterclockwise to obtain pixel coordinates of the r +1 th pixel with the gray value of 255, and calculating the pixel coordinates (u) in the background image_r+1,v_r+1) The pixel value is set to 255;

step 4016, repeating step 4015 for multiple times, continuing to scan and track the following contour point until the contour point is scanned and returns to the pixel coordinate (u)₀,v₀) Finishing the outline extraction of a bubble area;

and 4017, repeating the 4012 and the 4016 for multiple times until the whole quartz crucible bubble binary image is scanned, and contour extraction of a plurality of bubble regions is completed, so that a plurality of bubble contours are obtained on the background image, and a bubble edge contour image is obtained.

Compared with the prior art, the invention has the following advantages:

1. simple structure, reasonable in design and simple and convenient, the input cost is lower in the installation.

2. The electronic magnifier can be driven to move by the mechanical arm, so that the electronic magnifier can stretch into the inner wall of the quartz crucible to be measured, bubbles on the inner wall of the quartz crucible can be conveniently acquired to amplify images, the quartz crucibles with different outer diameters and heights can be used, and the application range is wide.

3. The adopted micro projector horizontally projects blue-violet light to the edge opening of the quartz crucible to be detected, the blue-violet light has the shortest wavelength and the highest frequency in all light, the energy of the blue-violet light is also the highest, the blue-violet light is not easy to disperse, and the blue-violet light background is used as compensation light, so that the contrast of a quartz crucible bubble image can be enhanced, the target and the background of the collected quartz crucible bubble amplified image have larger difference, and the extraction of the bubble profile is facilitated.

4. The Otsu algorithm module is used for carrying out image segmentation on the quartz crucible bubble equalization image for three times to obtain three layers of bubble images, the method is effectively suitable for the bubble images corresponding to the thicknesses of the quartz crucible transparent layer in the ranges of 0 mm-3 mm, 3 mm-8 mm and 8 mm-12 mm, and the accuracy of bubble layering is improved.

5. In the adopted bubble number obtaining process, the number of bubbles in a non-overlapping area and the number of bubbles in an overlapping area are obtained respectively, the overlapping area and the non-overlapping area can be distinguished for counting, the accuracy of bubble counting is effectively improved, and the method can adapt to various practical quartz crucibles.

6. The adopted quartz crucible bubble layer counting method has simple steps, convenient realization and simple and convenient operation, and ensures the accuracy of layer bubble image acquisition.

7. The adopted quartz crucible bubble layered counting method is simple and convenient to operate and good in using effect, firstly, the quartz crucible image is collected and transmitted, secondly, denoising and balancing of the quartz crucible amplified image are carried out, the quartz crucible bubble balanced image is obtained, and then, an Otsu algorithm module is adopted to carry out three-time image segmentation on the quartz crucible bubble balanced image, and the three-layer bubble image is obtained; and finally, edge extraction is carried out on the number of the bubbles in the three layers of bubble images respectively, so that the number of the bubbles in each layer of bubble image can be conveniently obtained, whether the number of the bubbles reaches the standard in the production process of the quartz crucible can be conveniently judged according to the number of the bubbles, the labor, the material and the financial resources are saved, the detection is accurate, the error rate is reduced, and the single crystal efficiency is improved.

In conclusion, the quartz crucible detection device is reasonable in design and low in cost, the bubble image of the transparent layer in the quartz crucible is obtained through the electronic magnifier, and the bubble image is layered, so that whether the production process of the quartz crucible meets the requirements or not is judged conveniently, labor, material resources and financial resources can be saved, the production benefit of manufacturers is improved, meanwhile, errors caused by human factors are avoided, the error rate in manual detection is reduced, the automation degree of quartz crucible detection is improved, the objectivity and the accuracy of a quartz crucible detection result are improved, and the accuracy of quartz crucible detection is improved.

The technical solution of the present invention is further described in detail by the accompanying drawings and embodiments.

Drawings

FIG. 1 is a schematic block diagram of the circuit of the quartz crucible bubble layer counting device of the present invention.

FIG. 2 is a flow chart of the bubble layer counting method of the quartz crucible of the invention.

FIG. 3 is an image of a first layer of bubbles according to the present invention.

FIG. 4 is a second layer bubble image of the present invention.

FIG. 5 is a third layer bubble image of the present invention.

Description of reference numerals:

1-a data processing controller; 2, a mechanical arm; 3-electronic magnifying glass;

4-display screen.

Detailed Description

The quartz crucible bubble layer counting device shown in fig. 1 comprises a data processing controller 1, a micro projector and an image acquisition device connected with the data processing controller 1, wherein the image acquisition device comprises a mechanical arm 2 and an electronic magnifier 3 clamped on the mechanical arm 2, the electronic magnifier 3 and the mechanical arm 2 are both connected with the data processing controller 1, and light projected by the micro projector is blue-violet light.

In this embodiment, the data processing controller 1 is a computer.

In this embodiment, the mechanical arm 2 is a six-degree-of-freedom mechanical arm, the magnification range of the electronic magnifier 3 is 1-500 times, the electronic magnifier 3 is connected with the data processing controller 1 through a USB transmission line, and the mechanical arm 2 is controlled by the data processing controller 1.

In this embodiment, the electronic magnifier 3 is used to collect and magnify the bubble image of the quartz crucible to obtain the magnified bubble image of the inner wall of the quartz crucible, because the diameter of the bubbles of the transparent layer on the inner wall of the quartz crucible is about several hundred micrometers, the bubbles of the quartz crucible obtained by a general camera are not convenient for image processing, and the obtained magnified bubble image is obtained

In the embodiment, the pixels of the electronic magnifier 3 are 30 thousands, the manual focusing is 10mm to 500mm, the USB interface is supplied with 5V direct current, and the magnification can be adjusted conveniently and manually or by a computer; in addition, the electronic magnifier is small, exquisite, firm and convenient to use; can be directly connected to a computer through a USB wire; the quartz crucible bubble image acquisition device can work under both strong light and weak light and has a certain automatic adjustment function, so that the bubble distribution of bubbles can be obtained by processing the transparent layer bubble image in the quartz crucible acquired by the electronic magnifier 3.

In this embodiment, arm 2 includes a plurality of TBS-K20 steering engines, and supply voltage 5V to 8.4V, but 360 turned angle, and accurate angle precision, the regulation of being convenient for to can remove electronic magnifier 3 to the region of waiting to gather, realize the accurate collection of quartz crucible stratum lucidum bubble image.

The bubble layer counting method for the quartz crucible as shown in FIG. 2 comprises the following steps:

step one, collecting and transmitting quartz crucible images:

step 101, the data processing controller 1 controls the mechanical arm 2 to rotate, and the mechanical arm 2 rotates to drive the electronic magnifier 3 to move, so that the vertical central line of the electronic magnifier 3 is perpendicular to the inner wall of the quartz crucible to be measured; a micro projector is arranged on the side edge of the quartz crucible to be measured, and horizontally projects blue-violet light to the edge opening of the quartz crucible to be measured;

step 102, manually adjusting the electronic magnifier 3 to focus until bubbles on the inner wall of the quartz crucible can be clearly observed by people;

103, the data processing controller 1 controls the electronic magnifier 3 to shoot bubble images of the inner wall of the quartz crucible to be detected, and sends the shot bubble amplified images of the inner wall of the quartz crucible to the data processing controller 1; wherein, the magnification of the electronic magnifier 3 is not less than 30 times;

step two, denoising and balancing the quartz crucible amplified image:

step 201, adopting the data processing controller 1 to call a gray processing module to perform gray processing on the amplified bubble image on the inner wall of the quartz crucible to obtain a gray image of the bubbles in the quartz crucible;

step 202, adopting the data processing controller 1 to call a weighted mean filtering module to filter the quartz crucible bubble gray level image to obtain a denoised quartz crucible bubble gray level image;

step 203, calling a gray value balancing module by using the data processing controller 1, and performing gray value balancing on the de-noised quartz crucible bubble gray image to obtain a quartz crucible bubble balancing image;

step three, layering quartz crucible bubble images:

step 301, calling an Otsu algorithm module by using the data processing controller 1 to perform primary image segmentation on the quartz crucible bubble equalization image to obtain a first layer bubble image; wherein the gray value threshold value of the primary image segmentation is 122-195;

step 302, calling an Otsu algorithm module by using the data processing controller 1, and performing secondary image segmentation on the quartz crucible bubble equalization image to obtain a second layer bubble image; wherein the gray value threshold value of the secondary image segmentation is 196-230;

step 303, calling an Otsu algorithm module by using the data processing controller 1, and performing image segmentation on the quartz crucible bubble equalization image for three times to obtain a third layer of bubble image; wherein the gray value threshold value of the three-time image segmentation is 231-255; wherein the bubble areas in the first layer of bubble image, the second layer of bubble image and the third layer of bubble image are white areas, and the background areas in the first layer of bubble image, the second layer of bubble image and the third layer of bubble image are black areas;

step four, acquiring bubble outlines of the quartz crucible bubble layered images: the method for acquiring the bubble outlines in the first layer of bubble image, the second layer of bubble image and the third layer of bubble image by adopting the data processing controller 1 is the same, and the method for acquiring the bubble outlines in the l-th layer of bubble image by adopting the data processing controller 1 is as follows:

step 401, calling a binary image edge tracking filtering algorithm module by using the data processing controller 1, and performing edge extraction on the l-th layer of bubble image to obtain an l-th layer of bubble edge contour image;

step 402, adopting the data processing controller 1 to call a morphological algorithm open operation, and performing open operation processing on the l-th layer bubble edge outline image to obtain an l-th layer bubble outline image after the open operation; wherein l is a positive integer, and the values of l are 1, 2 and 3.

In this embodiment, after the first-layer bubble profile image after the opening operation is obtained in the fourth step, the determination of the bubble non-overlapping area and the bubble overlapping area in the fifth step is performed, the data processing controller 1 is adopted to determine that the determination methods of the bubble non-overlapping area and the bubble overlapping area in the first-layer bubble image, the second-layer bubble image, and the third-layer bubble image are the same, and the data processing controller 1 is adopted to determine the bubble non-overlapping area and the bubble overlapping area in the l-th-layer bubble image, which is specifically as follows:

step 501, marking bubble outline connected areas on the first layer of bubble images by adopting a region growing method through the data processing controller 1 to obtain a plurality of bubble outline connected areas; wherein the number of the bubble profile communication areas is f_l，f₁Representing the number of connected areas of the bubble contours in the first layer bubble image, f₂Representing the number of connected areas of the bubble profile in the second layer bubble image, f₃Representing the number of bubble outline connected areas in the third layer of bubble images;

step 502, adopting the data processing controller 1 to call a minimum rectangle extraction module, and processing an nth bubble outline connected region to obtain a minimum external rectangle of the nth bubble outline connected region; wherein one side length of the minimum external rectangle is parallel to the row direction of the first layer of bubble image, the second layer of bubble image or the third layer of bubble image, n is a positive integer, and the value range of n is 1-f_l，f_lIs a positive integer not less than 1;

step 503, calling a pixel coordinate extraction module by using the data processing controller 1 to obtain pixel coordinates of an upper left corner pixel point, an upper right corner pixel point, a lower left corner pixel point and a lower right corner pixel point of the minimum circumscribed rectangle; adopting the data processing controller 1 to call a pixel coordinate conversion image coordinate module, and converting the pixel coordinates of the upper left corner pixel point, the upper right corner pixel point, the lower left corner pixel point and the lower right corner pixel point of the minimum external rectangle to obtain the image coordinate of the upper left corner pixel point of the minimum external rectangle

Image coordinates of upper right corner pixel

Image coordinates of lower left corner pixel points

And image coordinates of lower right corner pixel points

Step 504, the data processing controller 1 according to the formula

Obtaining the circularity E of the n-th bubble outline communication area_n；

Step 505, judging 0.5 by using the data processing controller 1<E_nIf it is not more than 1, when it is 0.5<E_nWhen the bubble outline is not more than 1, the bubble outline communication area is a non-overlapped bubble outline; otherwise, when 0.5<E_nWhen the bubble outline is not more than 1, the bubble outline communication area is an overlapped bubble outline;

step 506, repeating the steps 501 to 505 for multiple times until f in the first layer bubble image is completed_lJudging the connected area of the bubble outlines to obtain an overlapped bubble outline and a non-overlapped bubble outline in the first layer of bubble image; wherein the number of overlapped bubble outlines in the first layer bubble image is C_dlThe number of non-overlapping bubble contours in the first layer bubble image is F_cl；F_clAnd C_dlAre all positive integers.

In this embodiment, after the judgment of the bubble non-overlapping area and the bubble overlapping area in the fifth step is completed, the number of bubbles in the sixth step is obtained, and the methods for respectively obtaining the number of bubbles in the first layer of bubble image, the second layer of bubble image, and the third layer of bubble image by using the data processing controller 1 are all the same, and the method for obtaining the number of bubbles in the l-th layer of bubble image by using the data processing controller 1 is specifically as follows:

step 601, employing said dataThe processing controller 1 calls a pixel coordinate extraction module to extract pixel coordinates of all pixel points on the mth overlapped bubble outline in the first-layer bubble image to obtain the pixel coordinates of all pixel points on the mth overlapped bubble outline in the first-layer bubble image; wherein, the pixel coordinate of the ith pixel point on the mth overlapped bubble outline in the ith layer of bubble image is P_lm,i(u_lm,i,v_lm,i) (ii) a The mth overlapped bubble profile in the ith layer of bubble image comprises N pixel points, i represents the serial number of the pixel points, i is obtained according to the mth overlapped bubble profile in the ith layer of bubble image in the anticlockwise or clockwise direction, m is a positive integer, and the value range of m is 1-C_dlI is a positive integer, and the value range of i is 1-N;

step 602, using the data processing controller 1 according to a formula

Step 603, adopting the data processing controller 1 according to a formula

Step 604, repeating step 603 for multiple times to obtain the distance from each pixel point on the mth overlapped bubble outline in the ith layer of bubble image to the centroid; adopting the data processing controller 1 to call a curve drawing module, taking the serial numbers of pixel points as an abscissa, taking the distance from the ith pixel point to the centroid on the mth overlapped bubble outline in the ith layer of bubble image as an ordinate, and drawing a distance curve of each pixel point on the mth overlapped bubble outline in the ith layer of bubble image;

605, adopting the data processing controller 1 to invoke a discrete fourier transform module, and performing discrete fourier transform on a distance curve of each pixel point on the mth overlapped bubble outline in the ith layer of bubble image to obtain a main value sequence of the mth overlapped bubble outline in the ith layer of bubble image; wherein, the length of the main value sequence of the mth overlapped bubble outline in the ith layer of bubble image is N;

step 606, using the data processing controller 1 to respectively record two adjacent distance values in the main value sequence of the mth overlapped bubble profile in the ith layer of bubble image as y_jAnd y_j+1And is combined with y_jThe serial number of the corresponding pixel point is marked as x_j，y_j+1The serial number of the corresponding pixel point is marked as x_j+1(ii) a Wherein j is a natural number, and j is more than or equal to 0 and less than or equal to N-1;

step 607, the data processing controller 1 follows the formula

representing an interpolated distance;

step 608, repeating step 607 for multiple times, completing interpolation processing of two adjacent distance values in the main value sequence of the mth overlapped bubble profile in the ith layer of bubble image, and fitting the distance difference points by using the data processing controller 1 to obtain a distance curve after interpolation processing;

step 609, the data processing controller 1 calls a Gaussian smoothing module, and performs Gaussian smoothing on the interpolated distance curve to obtain a smoothed distance curve of the mth overlapped bubble profile in the first-layer bubble image;

step 6010, the data processing controller 1 obtains the number of bubbles in the mth overlapped bubble profile in the lth bubble image according to the smoothed distance curve of the mth overlapped bubble profile in the lth bubble image; a maximum value exists between two adjacent minimum value points on a smooth distance curve of the mth overlapped bubble profile in the ith layer of bubble image, and a profile surrounded by pixel points corresponding to the two adjacent minimum value points on the smooth distance curve of the mth overlapped bubble profile in the ith layer of bubble image is a bubble profile;

step 6012, the data processing controller 1 calls an addition module to determine the number F of non-overlapping bubbles in the first layer of bubble image_clAdding the number of the bubbles in the outline of the overlapped bubbles in the first layer of bubble image to obtain the total number Z of the quartz crucible bubbles in the first layer of bubble image_lj(ii) a And the number of non-overlapping bubble outlines in the first-layer bubble image is the number of non-overlapping bubbles in the first-layer bubble image.

In this embodiment, in step 203, the data processing controller 1 is adopted to invoke a gray value balancing module to perform gray value balancing on the denoised quartz crucible bubble gray image, and the specific process is as follows:

step 2031, calling a pixel value extraction module by using the data processing controller 1, extracting pixel values of all pixel points in a bubble area in the de-noised quartz crucible bubble gray level image, and obtaining the pixel values of all pixel points in the bubble area in the de-noised quartz crucible bubble gray level image; the number of pixel points in a bubble area in the denoised quartz crucible bubble gray level image is m₁And the pixel value of the ith pixel point in the bubble area in the denoised quartz crucible bubble gray level image is Pl_i′I' and m₁Are all positive integers, and the value range of i' is more than or equal to 1 and less than or equal to m₁；

Step 2032, sorting the pixel values of each pixel point in the bubble area in the de-noised quartz crucible bubble gray scale image by using the data processing controller 1 from small to large to obtain a minimum pixel value and a maximum pixel value, and assigning the maximum pixel value to Pl_maxAssigning the minimum pixel value to Pl_min；

Step 2033, using said data processing controller 1 according to formula

step 2034, assigning the average value of the pixel values of each pixel point in the bubble area in the de-noised quartz crucible bubble gray level image to each pixel point in the bubble area in the de-noised quartz crucible bubble gray level image by using the data processing controller 1, so as to update and balance the pixel values of each pixel point in the bubble area in the de-noised quartz crucible bubble gray level image and complete gray level balance of the de-noised quartz crucible bubble gray level image;

In this embodiment, in step 401, the data processing controller 1 is adopted to invoke a binary image edge tracking filter algorithm module, and perform edge extraction on a quartz crucible bubble binary image to obtain a bubble edge contour image, and the specific process is as follows:

step 4011, establishing a background image by using the data processing controller 1, wherein the size of the background image is the same as that of the quartz crucible bubble binary image, and the gray value of each pixel point in the background image is zero;

step 4012, scanning the quartz crucible bubble binary image from top to bottom by using the data processor 1, and recording a pixel coordinate (u) of a 255 pixel point of a first gray value when the data processor 1 obtains the 255 pixel point of the first gray value₀,v₀) And the pixel coordinates (u) in the background image are determined₀,v₀) The pixel value is set to 255;

step 4013, using said data processor 1 to calculate the pixel coordinate (u)₀,v₀) The eight neighborhood pixels which are taken as the center are scanned in turn according to the anticlockwise direction, and the pixel coordinate (u) is used₀,v₀) When a pixel with the gray value of 255 exists in the eight neighborhood pixels which are taken as the center, the pixel coordinate (u) of the pixel with the second gray value of 255 is recorded₁,v₁) And the pixel coordinates (u) in the background image are determined₁,v₁) The pixel value is set to 255;

In this embodiment, the micro projector horizontally projects blue-violet light to the edge opening of the quartz crucible to be measured, first, the blue-violet light has the shortest wavelength and the highest frequency among all lights, the energy of the blue-violet light is also the highest, the blue-violet light is not easy to disperse, and the blue-violet light background is used as compensation light, so that the contrast of a quartz crucible bubble image can be enhanced, a target of the collected quartz crucible bubble amplified image is greatly different from the background, and the extraction of the bubble profile is facilitated; secondly, the absorption rates of the transparent layers of the quartz crucible at different thicknesses to the blue-violet light are different, so that the gray values of the bubble images in the transparent layers of the quartz crucible at different thicknesses are different, and the segmentation and layering processing can be performed according to different gray value thresholds.

In this embodiment, the image filtering module generally includes a mean filtering module, a weighted mean filtering module and a median filtering module, and the weighted mean filtering module is selected to filter the quartz crucible bubble gray level image, because the weighted mean filtering module can effectively filter white noise in the quartz crucible bubble gray level image, and meanwhile, details in the quartz crucible bubble gray level image cannot be lost, so that the image filtering module is clearer.

In this embodiment, due to the background light, the gray value of each bubble edge in the bubble image after the denoising and equalizing step 201 and the step 202 is not in a stable range, and has a large floating value, which has a great influence on the subsequent bubble layering process, and causes the layered bubble image to be broken or repeatedly displayed, so that the gray value of the quartz crucible bubble gray image after the denoising step 202 is equalized.

In the embodiment, the opening operation is performed by adopting the morphological algorithm, firstly, because the bubble edge profile image obtained by extracting the edge of the bubble image on the l-th layer is obtained in the step 401, and because the transparent layer of the quartz crucible is transparent, the background color and the target color in the quartz crucible bubble equalization image are relatively close, so that more interference targets are introduced, therefore, the opening operation is introduced for performing the opening operation, the small interference objects in the bubble edge profile image are eliminated, the objects at the joint can be separated, the profile boundary is smoothed, meanwhile, the area of the target profile target cannot be changed, and the diameter size of the bubble can be conveniently and accurately analyzed subsequently; secondly, the edge extraction method is not convenient for obtaining the diameter of the bubble by fitting the bubble subsequently because the contour boundary in the crucible bubble contour image obtained by edge extraction is not smooth.

In the embodiment, the difference between the target area and the background area in the quartz crucible bubble equalization image subjected to image denoising and equalization is obvious, and the difference between the gray levels of the target area is large, so that the quartz crucible bubble equalization image is subjected to image segmentation by selecting a plurality of different gray value thresholds by adopting the gray levels to obtain three layers of bubble images, the purpose of bubble layering is achieved, and the processing process is simple and accurate.

In this embodiment, when the Otsu algorithm module is adopted to maximize the inter-class variance between the target region and the background region, the gray value is the threshold of the selected gray value, and after the distribution of the integral gray of the quartz crucible bubble equalized image and the characteristic relationship of the neighborhood of pixels and the like are considered comprehensively, the gray value of the maximum variance occurring in the pixel group subjected to gray classification is taken as the segmentation threshold of the integral image. In addition, through observing the thickness of the transparent layer of the quartz crucible for many times, the gray value threshold values corresponding to the thicknesses of the transparent layer of the quartz crucible in the ranges of 0 mm-3 mm, 3 mm-8 mm and 8 mm-12 mm are respectively determined to be 122-195, 196-230 and 231-255, so that the accuracy of bubble layering is improved, and whether the processing process of the quartz crucible is accurate or not is conveniently judged according to images of all layers.

In the embodiment, an eight-neighborhood-based contour tracking method is adopted, because the result can be obtained only by avoiding repeated scanning, and the tracking result is not correct due to the fact that the repeated scanning is difficult to determine, or a local area is repeatedly tracked by scanning, so that dead cycle occurs in execution; in addition, by adopting an eight-neighborhood-based contour tracing method, all target boundary information can be obtained through one-time scanning, and contour extraction of a plurality of target boundaries in a mixed connection mode is facilitated.

In this embodiment, the first layer bubble profile image obtained from the first layer bubble image obtained in step 301 through step four is shown in fig. 3, the second layer bubble profile image obtained from the second layer bubble image obtained in step 302 through step four is shown in fig. 4, and the third layer bubble profile image obtained from the third layer bubble image obtained in step 303 through step four is shown in fig. 5.

In this embodiment, it should be noted that the size of the image refers to rows × columns of pixel points.

In this embodiment, steps 605 to 607 are adopted to perform gaussian smooth interpolation processing on the distance curve of each pixel point on the mth overlapped bubble profile in the first layer of bubble image, on one hand, the distance curve is smoothed in order to remove burrs on the distance curve, so that a minimum value point and a maximum value point are conveniently obtained, and the number of bubbles in the overlapped bubble profile is conveniently obtained according to the minimum value point and the maximum value point;

The above description is only a preferred embodiment of the present invention, and is not intended to limit the present invention, and all simple modifications, changes and equivalent structural changes made to the above embodiment according to the technical spirit of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. A quartz crucible bubble layered counting method adopts a device comprising a data processing controller (1), a micro projector and an image acquisition device connected with the data processing controller (1), wherein the image acquisition device comprises a mechanical arm (2) and an electronic magnifier (3) clamped on the mechanical arm (2), the electronic magnifier (3) and the mechanical arm (2) are both connected with the data processing controller (1), and light projected by the micro projector is blue-violet light; the method is characterized by comprising the following steps:

step one, collecting and transmitting quartz crucible images:

step 101, a data processing controller (1) controls a mechanical arm (2) to rotate, the mechanical arm (2) rotates to drive an electronic magnifier (3) to move, and the vertical central line of the electronic magnifier (3) is perpendicular to the inner wall of a quartz crucible to be measured; a micro projector is arranged on the side edge of the quartz crucible to be measured, and horizontally projects blue-violet light to the edge opening of the quartz crucible to be measured;

102, manually adjusting the electronic magnifier (3) to focus until bubbles on the inner wall of the quartz crucible can be clearly observed by people;

103, the data processing controller (1) controls the electronic magnifier (3) to shoot bubble images of the inner wall of the quartz crucible to be detected, and sends the shot bubble amplified images of the inner wall of the quartz crucible to the data processing controller (1); wherein the magnification of the electronic magnifier (3) is not less than 30 times;

step two, denoising and balancing the quartz crucible amplified image:

step 201, adopting the data processing controller (1) to call a gray processing module to perform gray processing on the amplified bubble image on the inner wall of the quartz crucible to obtain a gray image of the quartz crucible bubble;

step 202, adopting the data processing controller (1) to invoke a weighted mean filtering module to filter the quartz crucible bubble gray level image to obtain a denoised quartz crucible bubble gray level image;

step 203, calling a gray value balancing module by using the data processing controller (1), and carrying out gray value balancing on the de-noised quartz crucible bubble gray image to obtain a quartz crucible bubble balancing image;

step three, layering quartz crucible bubble images:

step 301, calling an Otsu algorithm module by using the data processing controller (1) to perform image segmentation on the quartz crucible bubble equalization image for the first time to obtain a first layer of bubble image; wherein the gray value threshold value of the primary image segmentation is 122-195;

step 302, calling an Otsu algorithm module by using the data processing controller (1) to perform secondary image segmentation on the quartz crucible bubble equalization image to obtain a second layer bubble image; wherein the gray value threshold value of the secondary image segmentation is 196-230;

step 303, calling an Otsu algorithm module by using the data processing controller (1) to perform image segmentation on the quartz crucible bubble equalization image for three times to obtain a third layer of bubble image; wherein the gray value threshold value of the three-time image segmentation is 231-255; wherein the bubble areas in the first layer of bubble image, the second layer of bubble image and the third layer of bubble image are white areas, and the background areas in the first layer of bubble image, the second layer of bubble image and the third layer of bubble image are black areas;

step four, acquiring bubble outlines of the quartz crucible bubble layered images: the method for acquiring the bubble outlines in the first layer of bubble image, the second layer of bubble image and the third layer of bubble image by adopting the data processing controller (1) is the same, and the method for acquiring the bubble outlines in the l-th layer of bubble image by adopting the data processing controller (1) is as follows:

step 401, calling a binary image edge tracking filtering algorithm module by using the data processing controller (1) to perform edge extraction on the l-th layer of bubble image to obtain an l-th layer of bubble edge contour image;

step 402, adopting the data processing controller (1) to invoke open operation in a morphological algorithm, and performing open operation processing on the l-th layer bubble edge outline image to obtain the l-th layer bubble outline image after the open operation; wherein l is a positive integer, and the values of l are 1, 2 and 3.

2. The method for counting bubbles in a quartz crucible layer by layer as set forth in claim 1, wherein: the data processing controller (1) is a computer.

3. The method for counting bubbles in a quartz crucible layer by layer as set forth in claim 1, wherein: the mechanical arm (2) is a six-degree-of-freedom mechanical arm, the magnification range of the electronic magnifier (3) is 1-500 times, the electronic magnifier (3) is connected with the data processing controller (1) through a USB transmission line, and the mechanical arm (2) is controlled by the data processing controller (1).

4. The method for counting bubbles in a quartz crucible layer by layer as set forth in claim 1, wherein: after the first-layer bubble outline image after the opening operation is acquired in the fourth step, judging the non-overlapping area and the overlapping area of the bubbles in the fifth step, wherein the judging methods of the non-overlapping area and the overlapping area of the bubbles in the first-layer bubble image, the second-layer bubble image and the third-layer bubble image are the same by using the data processing controller (1), and the non-overlapping area and the overlapping area of the bubbles in the l-layer bubble image are judged by using the data processing controller (1), specifically as follows:

step 501, marking bubble outline connected areas on the first layer of bubble images by adopting a data processing controller (1) and a region growing method to obtain a plurality of bubble outline connected areas; wherein the number of the bubble profile communication areas is f_l，f₁Representing the number of connected areas of the bubble contours in the first layer bubble image, f₂Representing the number of connected areas of the bubble profile in the second layer bubble image, f₃Representing the number of bubble outline connected areas in the third layer of bubble images;

step 502, adopting the data processing controller (1) to invoke a minimum rectangle extraction module to process the nth bubble outline connected region to obtain a minimum circumscribed rectangle of the nth bubble outline connected region; wherein one side length of the minimum external rectangle is parallel to the row direction of the first layer of bubble image, the second layer of bubble image or the third layer of bubble image, n is a positive integer, and the value range of n is 1-f_l，f_lIs a positive integer not less than 1;

step 503, calling a pixel coordinate extraction module by using the data processing controller (1) to obtain pixel coordinates of an upper left corner pixel point, an upper right corner pixel point, a lower left corner pixel point and a lower right corner pixel point of the minimum circumscribed rectangle; adopting the data processing controller (1) to call a pixel coordinate conversion image coordinate module, and converting pixel coordinates of upper left corner pixel points, upper right corner pixel points, lower left corner pixel points and lower right corner pixel points of the minimum external rectangle to obtain image coordinates of the upper left corner pixel points of the minimum external rectangle

Image coordinates of upper right corner pixel

Image coordinates of lower left corner pixel points

And image coordinates of lower right corner pixel points

Step 504, the data processing controller (1) according to the formula

Obtaining the circularity E of the n-th bubble outline communication area_n；

505, judging that E is more than 0.5 by using the data processing controller (1)_nIf it is not more than 1, when E is more than 0.5_nWhen the bubble outline is not more than 1, the bubble outline communication area is a non-overlapped bubble outline; otherwise, when E is more than 0.5_nWhen the bubble outline is not more than 1, the bubble outline communication area is an overlapped bubble outline;

5. The method for counting bubbles in a quartz crucible layer by layer as set forth in claim 1, wherein: after judging the non-overlapping area and the overlapping area of the bubbles in the fifth step, acquiring the number of the bubbles in the sixth step, wherein the data processing controller (1) is adopted to acquire the number of the bubbles in the first layer of bubble image, the second layer of bubble image and the third layer of bubble image respectively, and the data processing controller (1) is adopted to acquire the number of the bubbles in the l layer of bubble image specifically as follows:

601, calling a pixel coordinate extraction module by using the data processing controller (1), and extracting pixel coordinates of each pixel point on the mth overlapped bubble outline in the first-layer bubble image to obtain the pixel coordinates of each pixel point on the mth overlapped bubble outline in the first-layer bubble image; wherein, the pixel coordinate of the ith pixel point on the mth overlapped bubble outline in the ith layer of bubble image is P_lm,i(u_lm,i,v_lm,i) (ii) a The mth overlapped bubble profile in the ith layer of bubble image comprises N pixel points, i represents the serial number of the pixel points, i is obtained according to the mth overlapped bubble profile in the ith layer of bubble image in the anticlockwise or clockwise direction, m is a positive integer, and the value range of m is 1-C_dlI is a positive integer, and the value range of i is 1-N;

step 602, adopting the data processing controller (1) according to a formula

Step 603, adopting the data processing controller (1) according to a formula

Step 604, repeating step 603 for multiple times to obtain the distance from each pixel point on the mth overlapped bubble outline in the ith layer of bubble image to the centroid; adopting the data processing controller (1) to call a curve drawing module, taking the serial numbers of pixel points as abscissa, taking the distance from the ith pixel point to the mass center on the mth overlapped bubble outline in the ith layer of bubble image as ordinate, and drawing a distance curve of each pixel point on the mth overlapped bubble outline in the ith layer of bubble image;

605, adopting the data processing controller (1) to invoke a discrete Fourier transform module, and performing discrete Fourier transform on a distance curve of each pixel point on the mth overlapped bubble outline in the ith layer of bubble image to obtain a main value sequence of the mth overlapped bubble outline in the ith layer of bubble image; wherein, the length of the main value sequence of the mth overlapped bubble outline in the ith layer of bubble image is the same as N;

step 606, using the data processing controller (1) to respectively record two adjacent distance values in the main value sequence of the mth overlapped bubble profile in the ith layer of bubble image as y_jAnd y_j+1And is combined with y_jThe serial number of the corresponding pixel point is marked as x_j，y_j+1The serial number of the corresponding pixel point is marked as x_j+1(ii) a Wherein j is a natural number, and j is more than or equal to 0 and less than or equal to N-1;

step 607, the data processing controller (1) generates a formula

representing an interpolated distance;

step 608, repeating step 607 for multiple times, completing interpolation processing of two adjacent distance values in the main value sequence of the mth overlapped bubble profile in the ith layer of bubble image, and fitting the distance difference points by using a data processing controller (1) to obtain a distance curve after interpolation processing;

step 609, the data processing controller (1) calls a Gaussian smoothing module, and Gaussian smoothing is carried out on the distance curve after interpolation processing to obtain a smooth distance curve of the mth overlapped bubble profile in the first layer of bubble image;

step 6010, the data processing controller (1) obtains the number of bubbles in the mth overlapped bubble profile in the ith layer of bubble image according to the smooth distance curve of the mth overlapped bubble profile in the ith layer of bubble image; a maximum value exists between two adjacent minimum value points on a smooth distance curve of the mth overlapped bubble profile in the ith layer of bubble image, and a profile surrounded by pixel points corresponding to the two adjacent minimum value points on the smooth distance curve of the mth overlapped bubble profile in the ith layer of bubble image is a bubble profile;

step 6012, the data processing controller (1) calls an addition module to determine the number F of non-overlapping bubbles in the first layer of bubble image_clAdding the number of the bubbles in the outline of the overlapped bubbles in the first layer of bubble image to obtain the total number Z of the quartz crucible bubbles in the first layer of bubble image_lj(ii) a And the number of non-overlapping bubble outlines in the first-layer bubble image is the number of non-overlapping bubbles in the first-layer bubble image.

6. The method for counting bubbles in a quartz crucible layer by layer as set forth in claim 1, wherein: in step 203, the data processing controller (1) is adopted to call a gray value balancing module to perform gray value balancing on the de-noised quartz crucible bubble gray image, and the specific process is as follows:

step 2031, calling a pixel value extraction module by using the data processing controller (1), extracting pixel values of all pixel points in a bubble area in the de-noised quartz crucible bubble gray level image, and obtaining the pixel values of all pixel points in the bubble area in the de-noised quartz crucible bubble gray level image; wherein, the denoised pixel points in the bubble area in the quartz crucible bubble gray level imageIs m₁And the pixel value of the ith pixel point in the bubble area in the denoised quartz crucible bubble gray level image is Pl_i′I' and m₁Are all positive integers, and the value range of i' is more than or equal to 1 and less than or equal to m₁；

Step 2032, sorting the pixel values of all pixel points in the bubble area in the de-noised quartz crucible bubble gray scale image by using the data processing controller (1) from small to large to obtain a minimum pixel value and a maximum pixel value, and assigning the maximum pixel value to Pl_maxAssigning the minimum pixel value to Pl_min；

Step 2033, using said data processing controller (1) according to formula

step 2034, assigning the average value of the pixel values of each pixel point in the bubble area in the de-noised quartz crucible bubble gray level image to each pixel point in the bubble area in the de-noised quartz crucible bubble gray level image by using the data processing controller (1), so as to update and balance the pixel values of each pixel point in the bubble area in the de-noised quartz crucible bubble gray level image, and thus, completing gray level balance of the de-noised quartz crucible bubble gray level image;

7. The method for counting bubbles in a quartz crucible layer by layer as set forth in claim 1, wherein: in step 401, the data processing controller (1) is adopted to call a binary image edge tracking filter algorithm module, edge extraction is carried out on the quartz crucible bubble binary image, and a bubble edge contour image is obtained, wherein the specific process is as follows:

step 4011, establishing a background image by using the data processing controller (1), wherein the size of the background image is the same as that of the quartz crucible bubble binary image, and the gray value of each pixel point in the background image is zero;

step 4012, scanning the quartz crucible bubble binary image from top to bottom by using the data processing controller (1), and recording a pixel coordinate (u) of a 255 pixel point with a first gray value when the data processing controller (1) acquires the first gray value of the 255 pixel point₀,v₀) And the pixel coordinates (u) in the background image are determined₀,v₀) The pixel value is set to 255;

step 4013, using said data processing controller (1) to coordinate (u) of pixel₀,v₀) The eight neighborhood pixels which are taken as the center are scanned in turn according to the anticlockwise direction, and the pixel coordinate (u) is used₀,v₀) When a pixel with the gray value of 255 exists in the eight neighborhood pixels which are taken as the center, the pixel coordinate (u) of the pixel with the second gray value of 255 is recorded₁,v₁) And the pixel coordinates (u) in the background image are determined₁,v₁) The pixel value is set to 255;