Spare liquid quality detection method

Abstract

The invention discloses a preparation liquid quality detection method, which belongs to the field of assisted reproduction, and comprises the following steps: collecting an original image of a culture solution preparation liquid, and reducing the width and the height in equal proportion; estimating the background light intensity of the reduced image based on a Gaussian blur function to obtain a background image; subtracting the background image from the shrunk image to obtain a foreground image; performing normalization and binarization processing on the foreground image; the contour breakpoints of the processed image are connected through closed operation, and a processed image is obtained; carrying out object contour searching on the processed image, extracting and representing an object contour in the image, and eliminating burrs and noisy points of a contour curve to obtain a final image; the contour area of the final image is calculated, liquid drops, bubbles and impurity contours are screened out according to the area, impurities and bubbles are distinguished according to roundness and gray values, the roundness and height of the contour of the liquid drops are calculated, the quality evaluation indexes of the existence of the impurities and the bubbles and the form quality of the liquid drops are given, the quality of the prepared liquid can be accurately and automatically evaluated, and the culture dish for completing the prepared liquid does not need to be manually screened.

Description

Technical Field

[0001] The present invention relates to the field of assisted reproduction, and in particular to a method for detecting the quality of a prepared liquid. Background Art

[0002] In the field of assisted reproduction, the embryo culture process requires the preparation of culture dishes for embryo culture in advance, a process called solution preparation. The solution preparation process involves titrating the culture solution and mineral oil in a certain arrangement in the culture dish to ensure that the culture solution droplets are covered by mineral oil to prevent the water in the culture solution from volatilizing and affecting the osmotic pressure of the culture solution.

[0003] However, during the liquid preparation process, operational errors may cause problems such as irregular droplet shapes and bubbles. Improper surface treatment of the culture dish or incorrect liquid dispensing method may cause the culture solution droplet height to be too low. Impurities may enter the culture dish during the liquid preparation process, causing droplet contamination.

[0004] Manual screening of qualified culture dishes for liquid preparation has high labor costs and low efficiency, and is affected by subjective factors, resulting in low repeatability and accuracy. Therefore, a method that can automatically evaluate the quality of liquid preparation is urgently needed to improve the efficiency and quality of liquid preparation. Summary of the invention

[0005] In order to overcome the deficiencies of the prior art, one of the objectives of the present invention is to provide a method for accurately and automatically evaluating the quality of a prepared solution without the need for manual screening of culture dishes for which the prepared solution has been completed.

[0006] One of the purposes of the present invention is achieved by the following technical solution:

[0007] A method for detecting the quality of a prepared solution comprises the following steps:

[0008] Collecting the original image of the culture solution preparation, and reducing the width and height of the original image in equal proportion;

[0009] The background light intensity of the reduced image is estimated based on the Gaussian blur function to obtain a background image;

[0010] Subtract the background image from the reduced image to obtain the foreground image;

[0011] Normalizing the foreground image, and binarizing the normalized foreground image to enhance the contrast and strengthen the features of the foreground image;

[0012] Connect the contour breakpoints of the binary processed image based on the morphological closing operation of dilation and erosion to obtain the processed image;

[0013] The object contour is found in the processed image. The contour tracing algorithm based on the Sobel edge detection algorithm and the chain code representation is used. By performing contour detection on the binary image, the boundary line in the image can be extracted. The geometric characteristics of the image are used to extract and represent the object contour in the image. The burrs and noise points of the contour curve are eliminated based on the Douglas-Peucker algorithm to obtain the final image.

[0014] The Shoelace polygon area calculation formula based on the vector cross product is used to calculate the contour area of ​​the final image. The droplet contours, bubbles and impurity contours are screened out according to the area size. Impurities and bubbles are distinguished according to the roundness and grayscale value. The droplet contour roundness and droplet height are calculated, and the presence of impurities and bubbles and droplet morphology quality evaluation indicators are given.

[0015] Furthermore, the proportional reduction of the width and height of the original image is specifically as follows: the size of the original image is (W, H), W is the width, H is the height, and the size of the compressed image is (W / k, H / k), where k is adjusted according to the image size and processing efficiency.

[0016] Furthermore, the background light intensity of the reduced image is estimated based on the Gaussian blur function to obtain the background image. Specifically, the Gaussian blur function is used to generate a weight matrix, and its formula is as follows:

[0017]

[0018] Among them, x, y are the relative positions of the pixel and the center point. The values ​​of x and y are related to the Gaussian kernel. The size of the Gaussian kernel (Size(bx, by)) is a two-dimensional matrix. bx and by are the width and height of the Gaussian kernel respectively. The value range of x is arrive The value range of y is arrive An integer between , σ is the standard deviation, σ determines the degree of blur, and σ is automatically calculated based on the Gaussian kernel size.

[0019] Furthermore, the foreground image is normalized as follows:

[0020]

[0021] Among them, src(x,y) is the pixel value of the fourth image, max_src and min_src are the minimum and maximum pixel values ​​of the foreground image, max_dst and min_dst are the minimum and maximum values ​​of the target range, and dst(x,y) is the pixel value of the normalized image.

[0022] Furthermore, the normalized foreground image is binarized as follows:

[0023]

[0024] Among them, dst(x,y) is the pixel value of the normalized image, and dst2(x,y) is the pixel value of the binary image.

[0025] Furthermore, the closing operation is specifically as follows:

[0026]

[0027] Among them, A is the processed image, B is the structural element, that is, the convolution kernel, and A·B is the closing operation of image A under the structural element B. represents the expansion operation, Represents an erosion operation.

[0028] Furthermore, the method of eliminating burrs and noises of the contour curve based on the Douglas-Peucker algorithm is specifically as follows:

[0029] The formula for the Douglas-Peucker algorithm is

[0030]

[0031] Among them, (x1, y1) and (x2, y2) are the coordinates of the two endpoints, and (x, y) is the coordinate of the point to be measured.

[0032] Furthermore, the calculation of the contour area of ​​the final image is specifically as follows:

[0033]

[0034] Where A is the area, (x i ,y i ) represents the coordinates of the i-th vertex of the polygon, (x i+1 ,y i+1 ) represents the coordinates of the next vertex.

[0035] Furthermore, in the method of distinguishing impurities and bubbles based on roundness and grayscale value, the roundness calculation formula is:

[0036]

[0037] Among them, Area is the area of ​​the contour and Perimeter is the perimeter of the contour.

[0038] Furthermore, in the calculation of the droplet profile roundness and droplet height, the droplet height formula is:

[0039]

[0040] Where V is the volume of the prepared liquid, A is the area of ​​the droplet contour, and h is the height of the droplet.

[0041] Compared with the prior art, the preparation liquid quality detection method of the present invention collects the original image of the culture solution preparation liquid, and proportionally reduces the width and height of the original image; estimates the background light intensity of the reduced image based on the Gaussian blur function to obtain a background image; subtracts the background image from the reduced image to obtain a foreground image; normalizes the foreground image, and binarizes the normalized foreground image to enhance the contrast and strengthen the features of the foreground image; connects the contour breakpoints of the binarized image based on the morphological closing operation of expansion and corrosion to obtain a processed image; searches for the object contour on the processed image, and uses contour tracking based on the Sobel edge detection algorithm and the chain code representation method The algorithm can extract the boundary lines in the image by performing contour detection on the binary image, extract and represent the contours of objects in the image by using the geometric characteristics of the image, and eliminate the burrs and noise of the contour curve based on the Douglas-Peucker algorithm to obtain the final image; the Shoelace polygon area calculation formula based on the vector cross product is used to calculate the contour area of ​​the final image, and the droplet contour, bubble and impurity contour are screened out according to the area size, and impurities and bubbles are distinguished according to the roundness and grayscale value. The droplet contour roundness and droplet height are calculated, and the presence of impurities and bubbles and droplet morphology quality evaluation indicators are given. The quality of the prepared solution can be accurately and automatically evaluated without the need for manual screening of the culture dishes for the prepared solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a flow chart of the method for testing the quality of the prepared solution of the present invention;

[0043] Figure 2 is the original image of the prepared solution;

[0044] Figure 3 is the reduced image;

[0045] Figure 4 is the background image;

[0046] Figure 5 is the foreground image;

[0047] Figure 6 is the normalized image;

[0048] Figure 7 is the image after binarization;

[0049] Figure 8 To process images;

[0050] Fig. 9 An object contour image is extracted and represented;

[0051] Fig.10 This is an image of a droplet with bubbles;

[0052] Fig.11 This is an image of a droplet with impurities. DETAILED DESCRIPTION

[0053] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0054] It should be noted that when a component is referred to as being "fixed to" another component, it may be directly on the other component or there may be another intermediate component through which it is fixed. When a component is considered to be "connected to" another component, it may be directly connected to the other component or there may be another intermediate component at the same time. When a component is considered to be "set on" another component, it may be directly set on the other component or there may be another intermediate component at the same time. The terms "vertical", "horizontal", "left", "right" and similar expressions used herein are for illustrative purposes only.

[0055] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0056] See also Figure 1 The present invention provides a method for detecting the quality of a prepared solution, comprising the following steps:

[0057] Collecting the original image of the culture solution preparation, and reducing the width and height of the original image in equal proportion;

[0058] The background light intensity of the reduced image is estimated based on the Gaussian blur function to obtain a background image;

[0059] Subtract the background image from the reduced image to obtain the foreground image;

[0060] Normalizing the foreground image, and binarizing the normalized foreground image to enhance the contrast and strengthen the features of the foreground image;

[0061] Connect the contour breakpoints of the binary processed image based on the morphological closing operation of dilation and erosion to obtain the processed image;

[0062] The object contour is found in the processed image. The contour tracing algorithm based on the Sobel edge detection algorithm and the chain code representation is used. By performing contour detection on the binary image, the boundary line in the image can be extracted. The geometric characteristics of the image are used to extract and represent the object contour in the image. The burrs and noise points of the contour curve are eliminated based on the Douglas-Peucker algorithm to obtain the final image.

[0063] The Shoelace polygon area calculation formula based on the vector cross product is used to calculate the contour area of ​​the final image. The droplet contours, bubbles and impurity contours are screened out according to the area size. Impurities and bubbles are distinguished according to the roundness and grayscale value. The droplet contour roundness and droplet height are calculated, and the presence of impurities and bubbles and droplet morphology quality evaluation indicators are given.

[0064] When the method for detecting the quality of prepared liquid of the present invention is implemented, the liquid is prepared first.

[0065] The specific process of liquid preparation is as follows:

[0066] The pipetting arm sucks up a volume V of culture fluid and a sufficient amount of mineral oil, the culture dish is moved to the sample loading position (x1, y1) through the stage, and the pipetting arm prepares the culture fluid to the specified position of the culture dish.

[0067] The specific steps for collecting the original image of the culture solution preparation are as follows: according to the position of the culture dish preparation solution, the stage transmission device corresponds to the shooting position (x2, y2), and the microscopic camera is used to focus and shoot the droplet of the culture solution to obtain the original image, such as Figure 2 As shown, in order to distinguish it from the subsequent images, the original image is called the first image. There is a positional relationship (Δx, Δy) between the loading position and the shooting position of the stage. The values ​​of Δx and Δy are determined by the actual transmission distance of the instrument. The positional relationship between the loading position and the shooting position is

[0068] x2=x1+Δx

[0069] y2=y1+Δy

[0070] The purpose of proportionally reducing the width and height of the original image is to improve the image processing efficiency. The specific method of proportionally reducing the width and height of the original image is as follows: compressing the captured droplet image to obtain a compressed image, such as Figure 3 As shown, the compressed image is called the second image. The compression process is: the width and height of the original image are reduced by k, the size of the first image is (W, H), W is the width, H is the height, and the size of the compressed image is (W / k, H / k). k is adjusted according to the image size and processing efficiency.

[0071] Due to the influence of factors such as external light intensity and system position deviation, the light intensity of the droplet image may be uneven, and the droplet image needs to be processed for uniform light intensity. The uniform light intensity processing includes generating a background image and obtaining a foreground image through the background image.

[0072] Specifically, the background image is obtained by estimating the background light intensity through Gaussian blur, such as Figure 4 As shown, it is called the third image.

[0073] The Gaussian blur function is used to generate a weight matrix (convolution kernel), and its formula is as follows:

[0074]

[0075] Where: x, y are the relative positions of the pixel and the center point. The values ​​of x and y are related to the Gaussian kernel. The size of the Gaussian kernel (Size(bx, by)) is a two-dimensional matrix, bx and by are the width and height of the Gaussian kernel respectively, and the value range of x is arrive The value range of y is arrive Integers between , a larger Gaussian kernel is needed to improve processing efficiency. The standard deviation σ determines the degree of blur and is automatically calculated based on the Gaussian kernel size.

[0076] Original image minus background image: Subtract the background image (third image) of the background light intensity from the compressed image (second image) to obtain the foreground image (fourth image), such as Figure 5 shown.

[0077] The foreground image is contrast enhanced, feature enhanced, normalized and binarized:

[0078] (1) The fourth image is normalized to obtain a normalized image, which is called the fifth image. Figure 6 As shown;

[0079]

[0080] Where, src(x,y) is the pixel value of the foreground image (the fourth image). max_src and min_src are the minimum and maximum values ​​of the pixel values ​​of the fourth image. max_dst and min_dst are the minimum and maximum values ​​of the target range. dst(x,y) is the pixel value of the fifth image after normalization.

[0081] (2) Binarization is performed on the fifth image. The image after binarization is called the sixth image. Figure 7 As shown;

[0082]

[0083] Wherein, dst(x, y) is the fifth image, and threshold is the set threshold.

[0084] In the sixth image, the contours of droplets, bubbles, impurities, etc. may have breakpoints. The closing operation in the morphological operation is used to connect the breakpoints to obtain a processed image, such as Figure 8 As shown, it is called the seventh image, and the closing operation is a combination of dilation and erosion.

[0085]

[0086] Where A is the sixth image. B is the structural element, i.e., the convolution kernel. A·B is the closing operation of image A under the structural element B. represents the expansion operation, Represents an erosion operation.

[0087] Find contours and eliminate burrs and noise

[0088] The object contour search for the processed image is as follows: the object contour search for the seventh image is performed, and the contour tracking algorithm based on the Sobel edge detection algorithm and the chain code representation method is used. By performing contour detection on the binary image, the boundary line in the image can be extracted, and the geometric characteristics of the image are used to extract and represent the object contour in the image, such as Fig. 9 shown.

[0089] The burrs and noise of the contour curve are eliminated based on the Douglas-Peucker algorithm to obtain the final image. The formula of the Douglas-Peucker algorithm is:

[0090]

[0091] Among them, (x1, y1) and (x2, y2) are the coordinates of the two endpoints, and (x, y) is the coordinate of the point to be measured.

[0092] According to the characteristics, droplets, bubbles and impurities are screened, and the relevant indicators for evaluating the quality of the prepared liquid, such as the presence of impurities and bubbles, and the quality of the droplet shape (roundness and height), are given. Specifically:

[0093] The shoelace polygon area calculation formula based on the vector cross product is used to calculate the contour area:

[0094]

[0095] Among them, (x i ,y i ) represents the coordinates of the i-th vertex of the polygon, (x i+1 ,y i+1 ) represents the coordinates of the next vertex.

[0096] First, the droplet outline, bubble and impurity outline are screened out according to the area size. Then, impurities and bubbles are distinguished according to the roundness and grayscale value. Fig.10 This is an image of a droplet with bubbles. The white circles are marked as identified bubbles. Fig.11 The image of a droplet with impurities is shown in Figure 1. The white rectangle is the identified impurity. The droplet contour roundness and droplet height are calculated, and the droplet quality evaluation related indicators such as the presence of impurities and bubbles, and the droplet morphology quality (roundness and height) are given.

[0097] Roundness calculation formula:

[0098]

[0099] Among them, Area is the area of ​​the contour and Perimeter is the perimeter of the contour.

[0100] Droplet height calculation formula:

[0101]

[0102] Wherein, V is the volume of the prepared solution in step 1, A is the area of ​​the droplet contour, and h is the height of the droplet.

[0103] The quality detection method of the prepared liquid of the present invention automatically gives evaluation indicators, greatly shortens the time of the quality evaluation of the prepared liquid, does not need to wait for manual microscope observation, reduces labor costs, can process a large number of samples instantly or in batches, and significantly improves work efficiency; manual observation is easily affected by subjective judgment, fatigue, distraction and other factors, while automated evaluation is based on preset algorithms and standards, reducing errors caused by human factors; image processing speed is fast, and the quality of the prepared liquid can be fed back in real time. And the processor performance requirements are relatively low, while ensuring the processing speed, the overall cost and energy consumption of the system are reduced; the quality of the prepared liquid is comprehensively evaluated through quantitative indicators such as droplet area, droplet height, and identification of bubbles and impurities, so that the results of the prepared liquid are more accurate and comprehensive.

[0104] The above embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this field, several modifications and improvements can be made without departing from the concept of the present invention, which are equivalent modifications and improvements made to the above embodiments based on the essential technology of the present invention, and all of them belong to the protection scope of the present invention.

Claims

1. A method for detecting the quality of a prepared solution, characterized in that: The following steps are involved: Collecting the original image of the culture solution preparation, and reducing the width and height of the original image in equal proportion; The background light intensity of the reduced image is estimated based on the Gaussian blur function to obtain a background image; Subtract the background image from the reduced image to obtain the foreground image; Normalizing the foreground image, and binarizing the normalized foreground image to enhance the contrast and strengthen the features of the foreground image; Connect the contour breakpoints of the binary processed image based on the morphological closing operation of dilation and erosion to obtain the processed image; The object contour is found in the processed image. The contour tracing algorithm based on the Sobel edge detection algorithm and the chain code representation is used. By performing contour detection on the binary image, the boundary line in the image can be extracted. The geometric characteristics of the image are used to extract and represent the object contour in the image. The burrs and noise points of the contour curve are eliminated based on the Douglas-Peucker algorithm to obtain the final image. The Shoelace polygon area calculation formula based on the vector cross product is used to calculate the contour area of ​​the final image. The droplet contours, bubbles and impurity contours are screened out according to the area size. Impurities and bubbles are distinguished according to the roundness and grayscale value. The droplet contour roundness and droplet height are calculated, and the presence of impurities and bubbles and droplet morphology quality evaluation indicators are given.

2. The method for detecting the quality of the prepared solution according to claim 1, characterized in that: The specific method of proportionally reducing the width and height of the original image is as follows: the size of the original image is (W, H), W is the width, H is the height, and the size of the compressed image is (W / k, H / k), where k is adjusted according to the image size and processing efficiency.

3. The method for detecting the quality of the prepared solution according to claim 1, characterized in that: The background light intensity of the reduced image is estimated based on the Gaussian blur function to obtain the background image. Specifically, the Gaussian blur function is used to generate a weight matrix, and its formula is as follows: Among them, x, y are the relative positions of the pixel and the center point. The values ​​of x and y are related to the Gaussian kernel. The size of the Gaussian kernel (Size(bx, by)) is a two-dimensional matrix. bx and by are the width and height of the Gaussian kernel respectively. The value range of x is arrive The value range of y is arrive An integer between , σ is the standard deviation, σ determines the degree of blur, and σ is automatically calculated based on the Gaussian kernel size.

4. The method for detecting the quality of the prepared solution according to claim 1, characterized in that: The specific normalization process of the foreground image is as follows: Among them, src(x,y) is the pixel value of the fourth image, max_src and min_src are the minimum and maximum pixel values ​​of the foreground image, max_dst and min_dst are the minimum and maximum values ​​of the target range, and dst(x,y) is the pixel value of the normalized image.

5. The method for detecting the quality of the prepared solution according to claim 1, characterized in that: The specific binarization process of the normalized foreground image is as follows: Among them, dst(x,y) is the pixel value of the normalized image, and dst2(x,y) is the pixel value of the binary image.

6. The method for detecting the quality of the prepared solution according to claim 1, characterized in that: The closing operation is as follows: Among them, A is the processed image, B is the structural element, that is, the convolution kernel, and A·B is the closing operation of image A under the structural element B. represents the expansion operation, Represents an erosion operation.

7. The method for detecting the quality of the prepared solution according to claim 1, characterized in that: The method of eliminating burrs and noise points of the contour curve based on the Douglas-Peucker algorithm is as follows: The formula for the Douglas-Peucker algorithm is Among them, (x1, y1) and (x2, y2) are the coordinates of the two endpoints, and (x, y) is the coordinate of the point to be measured.

8. The method for detecting the quality of the prepared solution according to claim 1, characterized in that: The calculation of the contour area of ​​the final image is specifically as follows: Where A is the area, (x i ,y i ) represents the coordinates of the i-th vertex of the polygon, (x i+1 ,y i+1 ) represents the coordinates of the next vertex.

9. The method for detecting the quality of the prepared solution according to claim 1, characterized in that: In the above-mentioned distinction between impurities and bubbles based on roundness and grayscale value, the roundness calculation formula is: Among them, Area is the area of ​​the contour and Perimeter is the perimeter of the contour.

10. The method for detecting the quality of the prepared solution according to claim 1, characterized in that: In the calculation of the droplet profile roundness and droplet height, the droplet height formula is: Where V is the volume of the prepared liquid, A is the area of ​​the droplet contour, and h is the height of the droplet.

Images