A method for analyzing particle aggregate images

Abstract

The present invention relates to the field of image analysis technology and discloses a method for analyzing particle aggregate images. The method comprises obtaining a silica sand sample and constructing a 12-bit superimposed image. The 12-bit superimposed image is then converted into a 16-bit projection image and denoised. The 12-bit superimposed image is then converted into an 8-bit grayscale slice image, and the height and width of the 8-bit grayscale slice image are obtained. The maximum size of the particles in the image is then obtained using prior knowledge. The partitioning parameters are calculated based on the maximum available computing memory of the computer. The image is then divided into n The method comprises the following steps: first, a sub-region is formed, and the relevant computing power and memory requirements are calculated; then, a local coordinate system is established in the sub-region, the particle morphological parameters are calculated, and incomplete and duplicate particles are eliminated; finally, the valid particle morphological parameters are sorted according to the pixel ID numbering principle to generate a sorted image. The present invention solves the problem that the existing technology causes inaccurate segmentation due to particle adhesion and overlap, which affects the accuracy of counting, measurement and morphological analysis.

Description

Technical Field

[0001] The present invention belongs to the technical field of image analysis, and in particular relates to a method for analyzing particle assembly images. Background Art

[0002] The precise analysis of particle aggregates is crucial in numerous fields, including materials science, geology, and biomedicine. For example, in materials science, the characteristics of particle aggregates directly influence the performance of a material. For example, the size, shape, and distribution of particles in ceramics determine their strength and toughness. In geological research, the characteristics of mineral particle aggregates in rocks can be used to infer the rock's origin and geological evolution history.

[0003] Traditionally, research on particle aggregates has relied heavily on manual observation and measurement, which not only consumes a lot of manpower and time, but also the results are easily affected by subjective factors, resulting in poor accuracy and repeatability. With the development of computer technology and image processing algorithms, digital image analysis technology has gradually been applied to the study of particle aggregates. Over the years, domestic and foreign scholars have conducted extensive and in-depth research on particle analysis technology based on image processing, and have achieved certain results, but there are still certain technical problems. On the one hand, during the image acquisition process, the image quality is uneven due to factors such as noise interference and uneven illumination, which brings difficulties to subsequent analysis. On the other hand, the existing image segmentation and feature extraction algorithms are difficult to accurately distinguish individual particles when processing images of aggregates with severe adhesion and overlap of particles, resulting in large errors in the results of particle counting, size measurement and morphological analysis, which cannot meet the needs of various fields.

[0004] Based on the above shortcomings, a method for analyzing large images of particle aggregates is proposed. By dividing the sample into intersecting areas and calculating the particle morphological information in each sub-area in the local coordinate system, an image segmentation method is used. Combined with digital image analysis, the accurate acquisition of particle morphological information is achieved, and work efficiency and analysis accuracy are improved. Summary of the Invention

[0005] In order to achieve the above objectives, the present invention provides a method for analyzing large images of particle aggregates to solve the technical problem in the prior art that it is difficult to accurately distinguish individual particles when processing images of adhered and overlapping particles, resulting in large errors in analysis results.

[0006] The technical solution adopted by the present invention is a method for analyzing an image of a particle assembly, comprising the following steps:

[0007] S1: Acquire silica sand sample and construct 12-bit superimposed image of silica sand sample;

[0008] S2: converting the 12-bit superimposed image of the silica sand sample into a 16-bit projection image of the silica sand sample, and performing denoising processing to obtain a noise-free 16-bit projection image;

[0009] S3: Convert the noise-free 16-bit projection image into an 8-bit grayscale slice image and obtain the height of the 8-bit grayscale slice image H ,width W ;

[0010] S4: Using prior knowledge, obtain the maximum size of particles in the 8-bit grayscale slice image ;

[0011] S5: Get the maximum available computing memory of the computer , the maximum size of the particles is obtained according to S4 and the maximum available computing memory on the computer , calculate the partitioning parameters, use the partitioning parameters to divide the 8-bit grayscale slice image into n sub-regions with overlapping areas, and calculate the computing power required for each sub-region Memory required for operations with the total repeated area ;

[0012] S6: Establish local coordinate systems in each of the n sub-regions, calculate the morphological parameters of the particles in each sub-region, check the particles in each sub-region in turn, delete incomplete particles and duplicate particles in each sub-region, and obtain valid particles;

[0013] S7: Sort the effective particle morphology parameters according to the pixel ID numbering principle to generate a sorted image.

[0014] The present invention is also characterized in that:

[0015] Furthermore, the specific steps of S1 are:

[0016] S1.1: Obtain silica sand sample;

[0017] S1.2: Perform a tomographic scan on the silica sand sample to obtain a 12-bit projection image of the silica sand sample;

[0018] S1.3: Repeat the tomography operation described in S1.2 to obtain several 12-bit projection images of the silica sand sample;

[0019] S1.4: Superimpose the 12-bit projection images of the silica sand sample obtained in S1.3 to obtain a 12-bit superimposed image of the silica sand sample.

[0020] Furthermore, in S2, the 16-bit projection image of the silica sand sample is denoised using a propagation-based phase contrast imaging algorithm.

[0021] Furthermore, the specific steps of S5 are:

[0022] S5.1: Obtain the maximum available computing memory on the computer The memory value occupied by each pixel in the 8-bit grayscale slice image m , using the maximum available computing memory on the computer The memory value occupied by each pixel in the 8-bit grayscale slice image m Calculate the maximum number of pixels allowed in each sub-region , its function is shown in formula (1):

[0023] (1)

[0024] in: The maximum number of pixels allowed for each sub-region, is the maximum available computing memory of the computer, m The memory value occupied by each pixel in the 8-bit grayscale slice image;

[0025] S5.2: Maximum size of particles obtained using S4 The maximum available computing memory of the computer obtained from S5.1 , calculate the side length of the sub-region L , its function is shown in formula (2):

[0026] (2)

[0027] in: The 8-bit grayscale slice image contains the maximum size of the particles. is the side length of the sub-region;

[0028] S5.3: Set the overlap ratio of adjacent sub-regions to ， The maximum diameter of the particles obtained using S4 , combined with the overlap ratio , set the overlap width between adjacent sub-regions O , the overlapping width between adjacent sub-regions O , the function is shown in formula (3):

[0029] (3)

[0030] in: is the overlapping width between adjacent sub-regions, is the overlapping ratio of adjacent sub-regions;

[0031] S5.4: Height of the 8-bit grayscale slice image obtained using S3 H ,width W and the overlapping width between adjacent sub-regions obtained in S5.3 O , calculate the number of sub-regions n , the number of sub-regionsn， The function is shown in formula (4):

[0032] (4)

[0033] in: is the height of the sub-region, is the width of the sub-region, is the number of sub-regions in the vertical direction, is the number of sub-regions in the horizontal direction, n is the number of sub-regions, O is the overlapping width between adjacent sub-regions;

[0034] S5.5: Sub-region side length obtained using S5.2 L , the overlapping width between adjacent sub-regions obtained in S5.3 O， S5.4 obtains the number of sub-regions, and divides the 8-bit grayscale slice image into n There are sub-regions with overlapping areas. The calculation formula for the area of the repeated area of each sub-region is shown in formula (5). The calculation formula for the total repeated area is shown in formula (6):

[0035] (5)

[0036] in: is the area of the repeated region for each subregion;

[0037] (6)

[0038] in: is the total repeated region area;

[0039] S5.6: Calculate the memory usage required for each sub-region , the memory usage required for each sub-region The function expression is as follows:

[0040] (7)

[0041] in: The memory usage required for each sub-region;

[0042] S5.7: Calculate the memory usage required for the total repeated region operation , the memory usage required for the total repeated region operation The function expression is shown in formula (8):

[0043] (8)

[0044] in: The memory usage required for the total repeated region operation;

[0045] S5.8: Calculation The total memory required for the subregions and the total repeated region , n The total memory required for the subregions and the total repeated region The function expression is as follows:

[0046] (9)

[0047] in: for n The total memory required for the subregions and the total repeated region;

[0048] S5.9: Judgment The total memory required for the subregions and the total repeated region Whether the maximum available operating memory of the computer is exceeded ,

[0049] like , then execute S6;

[0050] Otherwise, return to step S5.1 to recalculate the maximum number of pixels allowed in each sub-region .

[0051] Furthermore, the specific steps of S7 are:

[0052] S7.1: Set pixel sorting rules based on particle morphology parameters;

[0053] S7.2: Perform global calibration on the 8-bit grayscale slice image and assign a unique ID to each valid particle;

[0054] S7.3: Traverse the overlapping sub-regions in the image, record the pixel position corresponding to the particle ID in each sub-region, and generate a complete data set;

[0055] S7.4: Detect partially identified particles in the dataset, retain the complete instances and merge the pixels of other partial instances, and update the ID association;

[0056] S7.5: Sort the valid particle pixels according to the pixel sorting rule set in S7.1, and update the pixel positions in the 8-bit grayscale slice image;

[0057] S7.6: Convert the sorted pixel array into an image, generating a sorted image.

[0058] The beneficial effects of the present invention are as follows: the present invention divides the silica sand sample into intersecting areas and calculates the image segmentation method of the particle morphology information in each sub-area in a local coordinate system, thereby effectively solving the problem in the prior art that it is difficult to accurately distinguish individual particles when processing images of adhered and overlapping particles; the present invention eliminates image noise on light and detectors through a propagation-based phase contrast imaging algorithm (PPCI), converts a 12-bit stacked image into a 16-bit projection image and optimizes it, and then converts it into an 8-bit grayscale slice image, etc., thereby effectively improving the image quality, so that the analysis method of the present invention has stronger adaptability to images with uneven quality when acquired; the present invention divides the 8-bit grayscale slice image into n intersecting sub-areas by calculating the division parameters, and calculates the computing power required for each intersecting sub-area and the memory required for the total repeated area operation, thereby ensuring the smooth progress of the analysis task and realizing the rational use of computer computing resources, avoiding the situation where the analysis is interrupted or the analysis efficiency is low due to insufficient memory. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0060] Figure 1 These are images of LBS (Leighton Buzzard Sand) under a scanning electron microscope and its particle size distribution diagram, (a) image of LBS under a scanning electron microscope, (b) particle size distribution diagram of LBS under a scanning electron microscope;

[0061] Figure 2 This is a tomographic scan of the sand sample of the present invention;

[0062] Figure 3 The present invention eliminates the image noise on the light and detector through the phase retrieval algorithm and optimizes the projected image;

[0063] Figure 4 It is a 16-bit grayscale slice image reconstructed by the present invention using a sinogram;

[0064] Figure 5 The image is obtained by converting the 16-bit grayscale slice image into an 8-bit grayscale slice image according to the present invention;

[0065] Figure 6 It is a block diagram and control parameter diagram of image segmentation of the present invention;

[0066] Figure 7 It is a framework diagram of the steps of image segmentation of the present invention;

[0067] Figure 8 μCT reconstructed images of a large silica sand sample, (a) side view; (b) top view. DETAILED DESCRIPTION

[0068] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the embodiments described 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 making creative efforts are within the scope of protection of the present invention.

[0069] Example 1: The material of this example is Leighton Buzzard Sand (LBS), which is a kind of silica sand composed of hard and regular spherical particles. The image of LBS is as follows when observed under a scanning electron microscope. Figure 1 (a) shows that its particle size matches Figure 1 (b)

[0070] This example uses the BL13W1 particle beam at the Shanghai Synchrotron Radiation Facility (SSRF) to conduct synchrotron radiation computed tomography (SR-CT) experiments. The tube voltage was set to 45 keV, and the tube current was automatically adjusted using an optimization program. Image acquisition used a semi-fixed scanning method with 180-degree rotation of the specimen. Each CT imaging slice had a size of 2048 × 2048 pixels.

[0071] Step S1: obtaining a silica sand sample and constructing a 12-bit overlay image of the silica sand sample;

[0072] The specific steps of step S1 are:

[0073] S1.1: Obtain a silica sand sample. In this example, prepare an appropriate amount of Leighton Buzzard Sand (LBS) silica sand as a test sample.

[0074] S1.2: Perform a tomographic scan on the silica sand sample. In this embodiment, the silica sand sample is tomographically scanned using the BL13W1 particle beam synchrotron radiation computed tomography scanner at SSRF to obtain a 12-bit projection image of the silica sand sample. The tube voltage is set to 45 keV, and the tube current is automatically adjusted by the optimization program.

[0075] S1.3: Repeat the tomography operation in S1.2 to obtain several 12-bit projection images of the silica sand sample.

[0076] S1.4: Superimpose the 12-bit projection images of the silica sand samples obtained in S1.3 to obtain a 12-bit superimposed image of the silica sand sample. The scanned image is as follows: Figure 2 shown.

[0077] Step S2: converting the 12-bit superimposed image of the silica sand sample into a 16-bit projection image of the silica sand sample, and denoising the 16-bit projection image of the silica sand sample to obtain a noise-free 16-bit projection image;

[0078] The specific steps of step S2 are as follows: using image processing algorithm, converting the 12-bit superposition image of silica sand sample into 16-bit projection image, using propagation-based phase contrast imaging algorithm (PPCI, Propagation-based Phase Contrast Imaging) to eliminate the image noise on the light and detector, and optimizing the projection image. The optimized image is as follows Figure 3 shown.

[0079] Step S3: Convert the noise-free 16-bit projection image ( Figure 4 ) is converted to an 8-bit grayscale slice image ( Figure 5 ), get the height of the 8-bit grayscale slice image H ,width W；

[0080] The specific steps of step 3 are as follows: PITRE software is used to complete the process. PITRE (Phase-sensitive x-ray Image processing and Tomography REconstruction) software is an imaging data processing software developed by scientists at the Shanghai Synchrotron Radiation Facility for users. PITRE supports phase retrieval of propagation-based phase-contrast imaging / tomography (PPCI), extraction of apparent absorption, refraction, and scattering information of diffraction-enhanced imaging (DEI), and allows parallel beam tomography reconstruction of traditional absorption CT (Computed Tomography) data as well as PPCT phase retrieval and DEI-CT (Diffraction Enhanced Imaging Computed Tomography) extraction information.

[0081] Step S4: Using prior knowledge, obtain the maximum size of particles in the 8-bit grayscale slice image ,

[0082] The specific steps of step S4 are as follows: with the help of previous experience in processing similar silica sand images, research results in related fields, and prior knowledge of the properties of silica sand materials, combined with image processing algorithms, the 8-bit grayscale slice image is analyzed to determine the maximum size of particles contained in the image. ;

[0083] S5: Get the maximum available computing memory of the computer , the maximum size of the particles is obtained according to S4 and the maximum available computing memory on the computer , calculate the partition parameters of the 8-bit grayscale slice image, divide the 8-bit grayscale slice image into n intersecting sub-regions using the partition parameters, and calculate the computing power required for each intersecting sub-region Memory required for operations with the total repeated area ;

[0084] The specific steps of step S5 are:

[0085] S5.1: Obtain the maximum available computing memory on the computer The memory value occupied by each pixel in the 8-bit grayscale slice image m , using the maximum available computing memory on the computer The memory value occupied by each pixel in the 8-bit grayscale slice image m Calculate the maximum number of pixels allowed for each intersecting sub-region , its function is shown in formula (1):

[0086] (1);

[0087] in: The maximum number of pixels allowed for each sub-region, is the maximum available computing memory of the computer, m The memory value occupied by each pixel in the 8-bit grayscale slice image;

[0088] S5.2: Maximum size of particles obtained using S4 The maximum available computing memory of the computer obtained from S5.1 , calculate the side length of the sub-region L , its function is shown in formula (2):

[0089] (2);

[0090] in: The 8-bit grayscale slice image contains the maximum size of the particles. is the side length of the sub-region;

[0091] S5.3: Set the overlap ratio of adjacent sub-regions to ， The maximum diameter of the particles obtained using S4 , combined with the overlap ratio , set the overlap width between adjacent sub-regions O , the overlapping width between adjacent sub-regions O , the function is shown in formula (3):

[0092] (3);

[0093] in: is the overlapping width between adjacent sub-regions, is the overlapping ratio of adjacent sub-regions;

[0094] S5.4: Height of the 8-bit grayscale slice image obtained using S3 H ,width W and the overlapping width between adjacent sub-regions obtained in S5.3 O , calculate the number of sub-regions n , the number of sub-regions n， The function is shown in formula (4):

[0095] (4);

[0096] in: is the height of the sub-region, is the width of the sub-region, is the number of sub-regions in the vertical direction, is the number of sub-regions in the horizontal direction, n is the number of sub-regions, O is the overlapping width between adjacent sub-regions;

[0097] S5.5: Use the sub-region side length L obtained in S5.2 and the overlap width between adjacent sub-regions obtained in S5.3 O， S5.4 obtains the number of sub-regions, and divides the 8-bit grayscale slice image into n sub-regions with overlapping areas, such as Figure 6 As shown, Figure 6 Where P1, P2, P3, P4, and P5 are particles; S1, S2, ..., Sn are different divided regions; and OR is the intersection of different regions. The calculation formula for the area of the repeated region of each subregion is shown in formula (5), and the calculation formula for the total repeated region area is shown in formula (6):

[0098] (5);

[0099] in: is the area of the repeated region for each subregion;

[0100] (6);

[0101] in: is the total repeated region area;

[0102] S5.6: Calculate the memory usage required for each sub-region , the memory usage required for each sub-region The function expression is as follows:

[0103] (7);

[0104] in: The memory usage required for each sub-region;

[0105] S5.7: Calculate the memory usage required for the total repeated region operation , the memory usage required for the total repeated region operation The function expression is shown in formula (8):

[0106] (8);

[0107] in: The memory usage required for the total repeated region operation;

[0108] S5.8: Calculation The total memory required for the subregions and the total repeated region , n The total memory required for the subregions and the total repeated region The function expression is as follows:

[0109] (9);

[0110] in: for n The total memory required for the subregions and the total repeated region;

[0111] S5.9: Judgment The total memory required for the subregions and the total repeated region Whether the maximum available operating memory of the computer is exceeded ,

[0112] like , then execute S6;

[0113] Otherwise, return to step S5.1 and recalculate the maximum number of pixels allowed in each sub-region .

[0114] Step S6:n A local coordinate system is established in each intersecting sub-region, and the morphological parameters of the particles in each intersecting sub-region are calculated. The particles in each intersecting sub-region are checked one by one, and the incomplete particles and duplicate particles in each intersecting sub-region are deleted to obtain valid particles.

[0115] Step S6 is specifically as follows: a local coordinate system is established in each intersecting sub-region, and the morphological parameters of the particles in each sub-region, i.e., the size and shape of the particles, are calculated using an image processing algorithm. The particles in each sub-region are checked one by one, and whether the particles are complete or duplicate particles is determined based on the morphological parameters. Incomplete and duplicate particles are deleted to obtain valid particles, such as Figure 7 As shown, Figure 7 P1, P2, P3, P4, and P5 are particles; S1, S2, ... Sn are different divided areas.

[0116] S7: Sort the effective particle morphology parameters according to the pixel ID numbering principle to generate a sorted image.

[0117] Step S7 is specifically as follows: S7.1: set pixel sorting rules according to particle morphology parameters; S7.2: perform global calibration on the 8-bit grayscale slice image and assign a unique ID to each valid particle; S7.3: traverse the overlapping sub-regions in the image, record the pixel positions corresponding to the particle ID in each sub-region, and generate a complete data set; S7.4: detect partially identified particles in the data set, retain complete instances and merge the pixels of other partial instances, and update the ID association; S7.5: sort the valid particle pixels according to the pixel sorting rules set in S7.1, and update the pixel positions in the 8-bit grayscale slice image; S7.6: convert the sorted pixel array into an image to generate a sorted image.

[0118] Figure 8 is a μCT reconstructed image of a large silica sand sample, where Figure 8 (a) The side view shows the internal structural features of the vertical section of the sample. Figure 8 (b) Top view showing the planar distribution characteristics of the horizontal section of the sample;

[0119] Each embodiment in this specification is described in a related manner. Similar parts between the various embodiments can be referred to in conjunction with each other. Each embodiment focuses on the differences between the other embodiments. In particular, the system embodiment is generally similar to the method embodiment, so the description is relatively simple. For related parts, refer to the description of the method embodiment.

[0120] The above description is only a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention are included in the scope of protection of the present invention.

Claims

1. A method for analyzing an image of a particle assembly, characterized in that: The specific steps are as follows: S1: Acquire silica sand sample and construct 12-bit superimposed image of silica sand sample; S2: converting the 12-bit superimposed image of the silica sand sample into a 16-bit projection image of the silica sand sample, and performing denoising processing to obtain a noise-free 16-bit projection image; S3: Convert the noise-free 16-bit projection image into an 8-bit grayscale slice image, and obtain the height H and width W of the 8-bit grayscale slice image; S4: Using prior knowledge, obtain the maximum size D of particles in the 8-bit grayscale slice image max ; S5: Get the maximum available computing memory M of the computer available , according to S4, the maximum size D of the particles is obtained max and the maximum available operating memory M of the computer available , calculate the partitioning parameters, use the partitioning parameters to divide the 8-bit grayscale slice image into n sub-regions with overlapping areas, and calculate the computing power size C required for each sub-region sub The memory M required for the total repeated area operation overlap ; S6: Establish local coordinate systems in each of the n sub-regions, calculate the morphological parameters of the particles in each sub-region, check the particles in each sub-region in turn, delete incomplete particles and duplicate particles in each sub-region, and obtain valid particles; S7: Sort the effective particle morphology parameters according to the pixel ID numbering principle to generate a sorted image; The specific steps of S7 are: S7.1: Set pixel sorting rules based on particle morphology parameters; S7.2: Perform global calibration on the 8-bit grayscale slice image and assign a unique ID to each valid particle; S7.3: Traverse the overlapping sub-regions in the image, record the pixel position corresponding to the particle ID in each sub-region, and generate a complete data set; S7.4: Detect partially identified particles in the dataset, retain the complete instances and merge the pixels of other partial instances, and update the ID association; S7.5: Sort the valid particle pixels according to the pixel sorting rule set in S7.1, and update the pixel positions in the 8-bit grayscale slice image; S7.6: Convert the sorted pixel array into an image, generating a sorted image.

2. The method for analyzing particle assembly images according to claim 1, characterized in that: The specific steps of S1 are: S1.1: Obtain silica sand sample; S1.2: Perform a tomographic scan on the silica sand sample to obtain a 12-bit projection image of the silica sand sample; S1.3: Repeat the tomography operation described in S1.2 to obtain several 12-bit projection images of the silica sand sample; S1.4: Superimpose the 12-bit projection images of the silica sand sample obtained in S1.3 to obtain a 12-bit superimposed image of the silica sand sample.

3. The method for analyzing particle assembly images according to claim 1, characterized in that: In S2, the 16-bit projection image of the silica sand sample is denoised using a propagation-based phase contrast imaging algorithm.

4. The method for analyzing particle assembly images according to claim 1, wherein: The specific steps of S5 are: S5.1: Get the maximum available computing memory M on the computer available The memory value m occupied by each pixel in the 8-bit grayscale slice image is calculated using the maximum available computing memory M on the computer. available Calculate the maximum number of pixels N allowed in each sub-region with the memory value m occupied by each pixel in the 8-bit grayscale slice image pixels , its function is shown in formula (1): Where: N pixels The maximum number of pixels allowed for each sub-region, M available is the maximum available computing memory of the computer, and m is the memory value occupied by each pixel in the 8-bit grayscale slice image; S5.2: The maximum size D of the particles obtained using S4 max The maximum available computer memory M obtained from S5.1 available , calculate the side length L of the sub-region, and its function is shown in formula (2): Where: D max is the maximum size of the particle in the 8-bit grayscale slice image, and L is the side length of the sub-region; S5.3: Set the overlap ratio of adjacent sub-regions to a, and use the maximum diameter D of the particles obtained in S4 max , combined with the overlap ratio a, set the overlap width O between adjacent sub-regions. The overlap width O between adjacent sub-regions is expressed as a function as shown in formula (3): O≥a×D max (3) Where: O is the overlap width between adjacent sub-regions, a is the overlap ratio of adjacent sub-regions; S5.4: Calculate the number of subregions n using the height H and width W of the 8-bit grayscale slice image obtained in S3 and the overlap width O between adjacent subregions obtained in S5.

3. The function of the number of subregions n is shown in formula (4): Where: L hight is the height of the sub-region, L weight is the width of the sub-region, n h is the number of sub-regions in the vertical direction, n w is the number of sub-regions in the horizontal direction, n is the number of sub-regions, and O is the overlapping width between adjacent sub-regions; S5.5: Using the subregion side length L obtained in S5.2, the overlap width O between adjacent subregions obtained in S5.3, and the number of subregions obtained in S5.4, the 8-bit grayscale slice image is divided into n subregions with overlapping areas. The calculation formula for the area of the repeated area of each subregion is shown in formula (5), and the calculation formula for the total repeated area is shown in formula (6): HAS overlap =(L width ×O)+(L height ×O)-(O×O) (5) Among them: A overlap is the area of the repeated region for each subregion; A total-overlap =n×A overlap -(n h +n w -2)×(O×O) (6) Among them: A total-overlap is the total repeated region area; S5.6: Calculate the memory usage C required for each sub-region sub , the memory usage required for each sub-region C sub The function expression is as follows: C sub =L hight ×L weight ×m (7) Where: C sub The memory usage required for each sub-region; S5.7: Calculate the memory usage M required for the total repeated region operation overlap , the memory usage M required for the total repeated region operation overlap The function expression is shown in formula (8): M overlap =A total-overlap ×m (8) Where: M overlap The memory usage required for the total repeated region operation; S5.8: Calculate the total memory M required for n subregions and the total repeated region total , the total memory M required for the n sub-regions and the total repeated region total The function expression is as follows: M total =n×C sub +M overlap (9) Where: M total The total memory required for n sub-regions and the total repeated region; S5.9: Determine the total memory M required for n sub-regions and the total repeated region total Whether it exceeds the maximum available operating memory M of the computer available , If M total <M available , then execute S6; Otherwise, return to step S5.1 to recalculate the maximum allowed number of pixels N for each sub-region. pixels .

Images