Catalyst active phase identification method based on mathematical morphology and Gabor filtering

Abstract

The invention belongs to the technical field of image processing, and discloses a catalyst active phase identification method based on mathematical morphology and Gabor filtering, and the method comprises the following steps: S1, carrying out the sulfuration of a catalyst, transferring the sulfurized catalyst to a TEM, and carrying out the observation of the sulfurized catalyst to complete the image collection of an electron microscope device; s2, extracting texture features and edge features of the catalyst image by using a Gabor filter, and extracting stripes of the hydrogenation catalyst by using a morphological algorithm; s3, extracting the edge of the stripe by using a Sobe l operator to carry out skeleton line endpoint detection; and S4, extracting the coordinates of the whole curve in sequence according to the endpoints of the skeleton line, so as to realize the identification of the active phase of the hydrogenation catalyst. The catalyst activity phase identification method based on mathematical morphology and Gabor filtering is superior to a traditional catalyst activity determination method, has high precision, is highly consistent with a result of a traditional activity determination method, and has practical engineering application value.

Description

Technical Field

[0001] This invention belongs to the field of image processing technology, specifically relating to a method for identifying the active phase of a catalyst based on mathematical morphology and Gabor filtering. Background Technology

[0002] With the development and widespread application of information technologies such as personal computers, the Internet, and other communication technologies, human society has entered the information age, and the digitization of information has become the most effective method for information transmission, processing, and storage. Digital imaging has also made significant progress in the field of electron microscopy. Converting various images obtained by electron microscopy into digital signals in real time for processing and analysis is an inevitable trend in modern electron microscopy image processing. Transmission electron microscopy can clearly provide information on the active phase of hydrogenation catalysts, and the statistical results of the average number of crystal layers in the active phase are well correlated with the catalyst activity. After accurately identifying and segmenting electron microscopy images of hydrogenation catalysts, and grouping and layering them, the catalyst activity can be determined through the information on the distribution of catalyst layer numbers.

[0003] Traditional methods for determining catalyst activity include the conversion frequency method, reaction rate method, activation energy, and conversion rate. However, these methods often suffer from limitations such as difficulty in measurement, low required reaction conversion rates, and inaccuracies. Therefore, methods have emerged that convert electron microscopy images into digital signals for processing and analysis. Traditional methods for determining catalyst activity still have many shortcomings, and the mainstream method of segmenting electron microscopy images using image processing techniques also cannot meet the requirements for catalyst identification. Based on this, this application proposes a method for identifying the active phase of hydrogenation catalysts based on morphological trace detection. Summary of the Invention

[0004] To overcome the aforementioned technical problems, this invention provides a catalyst active phase identification method based on mathematical morphology and Gabor filtering. This method can better preserve catalyst information in images, resulting in high accuracy in catalyst identification and layer number statistics. It can efficiently and accurately determine the magnitude of catalyst activity, thus solving the limitations of traditional measurement methods, such as measurement difficulties, requirements for low reaction conversion rates, and inaccuracies.

[0005] The present invention adopts the following technical solution:

[0006] A catalyst active phase identification method based on mathematical morphology and Gabor filtering includes the following steps:

[0007] S1. Image acquisition using electron microscopy equipment; After the catalyst is sulfided, it is transferred to a TEM for observation to complete the image acquisition using electron microscopy equipment;

[0008] S2. Feature extraction: The texture and edge features of the catalyst image are extracted using a Gabor filter, and the stripes of the hydrogenation catalyst are extracted using a morphological algorithm.

[0009] S3. Skeleton line extraction; The Sobel operator is used to extract the edges of the stripes for skeleton line endpoint detection;

[0010] S4. Active phase stripe screening: Extract the coordinates of the entire curve in sequence according to the endpoints of the skeleton line to identify the active phase of the hydrogenation catalyst.

[0011] Compared with the prior art, the beneficial effects of the present invention are:

[0012] A catalyst active phase identification method based on mathematical morphology and Gabor filtering acquires catalyst layer distribution information from a large number of image samples, which effectively improves the robustness of the method and reduces errors caused by manual judgment. Experimental results show that this method has high accuracy and is highly consistent with the results of traditional activity assessment methods, demonstrating practical engineering application value. Attached Figure Description

[0013] Figure 1 A flowchart of the catalyst active phase identification method based on mathematical morphology and Gabor filtering provided by the present invention;

[0014] Figure 2 A dataset of hydrogenation catalyst images designed to verify the method provided by this invention;

[0015] Figure 3 The results of the classification experiment for the test set images. Detailed Implementation

[0016] The embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings. Throughout the drawings, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions. Unless otherwise specified, the raw materials and equipment used are commercially available or commonly used in the art. The methods in the embodiments, unless otherwise specified, are conventional methods in the art. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0017] like Figure 1 As shown, this invention provides a method for identifying the active phase of a catalyst based on mathematical morphology and Gabor filtering, comprising the following steps:

[0018] S1: Image acquisition using electron microscope equipment.

[0019] Before performing transmission electron microscopy (TEM) analysis, the catalyst must be pre-sulfurized. The sulfurization procedure is as follows: Take 1 g of 80-100 mesh catalyst, place it in a glass reactor, purge with N2 at 50 mL / min for about 10 min, then switch to H2 containing 15% H2S at a flow rate of 100 mL / min, raise the temperature at 6 °C / min to 360 ± 30 °C, then reduce the temperature to 1 °C / min, raise the temperature to 370 ± 30 °C, and hold for 4 h. Then cool down to 120 ± 30 °C at a cooling rate of 50 °C / min, purge with N2 for 0.5 h, and then, under N2 protection, pour the sulfurized catalyst into cyclohexane to prevent oxidation, grind for 40 min, and then load the catalyst and cyclohexane together into a sample vial.

[0020] Before dispersing the sample onto the sample holder, the suspension in the sample vial is first dispersed by ultrasonic waves. Then, an appropriate amount of sample is added to the microgrid using a dropper. After determining reasonable experimental parameters, the sample is quickly transferred to the TEM for observation after the solvent has evaporated.

[0021] S2: Feature extraction.

[0022] Gabor filters can effectively extract texture and edge features from images, while also reducing the impact of lighting variations and noise to some extent.

[0023] The Gabor transform, as an important time-frequency analysis method, is a convolution of a sine wave and a Gaussian kernel function in the spatial domain, and a translated Gaussian function in the frequency domain. The kernel function is defined as follows:

[0024]

[0025] Where x′ and y′ are determined by the following formula

[0026]

[0027] In equations (1) and (2), x and y represent pixel coordinates, λ represents the wavelength of the cosine function, θ represents the direction factor of the parallel stripes, σ represents the standard deviation of the Gaussian function, and γ represents the spatial aspect ratio. Control the phase shift of the cosine function.

[0028] S3: Skeleton line extraction.

[0029] The skeleton is extracted by successively removing edges from the input image contour, which has a certain width, until it is reduced to a width of only one pixel. The skeleton of a circle is its center, while the skeleton of a line and an isolated point is itself.

[0030] First, find an 8-point neighborhood centered on the boundary point, denoted as P1. The 8 points within this neighborhood, clockwise around the center point, are designated as P2, P3, ..., P9, where P2 is above P1. Complete the skeleton line extraction as follows:

[0031] (1). First, mark the boundary points that simultaneously satisfy the following conditions:

[0032] ①2≤N(P1)≤6;

[0033] ②S(P1)=1;

[0034] ③P2*P4*P6=0;

[0035] ④P4*P6*P8=0;

[0036] Where N(P1) is the number of non-zero leading points of P1, and S(P1) is the number of times the values ​​of these points change from 0 to 1 when ordered as P2, P3, ..., P9.

[0037] (2). Same as step (1), change condition ③ to P2*P4*P8=0; change condition ④ to P2*P6*P8=0; after all boundary points have been checked, remove all marked points.

[0038] (3). Steps (1) and (2) constitute an iteration until no more points satisfy the marking conditions. The remaining points form the region, which is the refined skeleton.

[0039] S4: Active Phase Stripe Screening

[0040] A threshold segmentation combined with the maximum connected unit method was used to extract the cuboid black block in the lower right corner of the electron microscope image, obtaining the corresponding pixel length in the x-direction. Dividing this by the actual length of 20 nm yielded the actual length of the unit pixel. Then, the Euclidean distances of n consecutive coordinates on the curve were calculated point by point, and the summation gave the curve length.

[0041] Define the shortest pixel length of the detection curve. min The contrast value of the curve to be detected c According to length min With value c The threshold is used to select m curves that meet the conditions. Where length min A higher value indicates that the detector will remove more short curves; c The value ranges from 0 to 1; the larger the value, the more inconspicuous curves will be removed.

[0042] The formulas for calculating the shortest pixel and contrast value are shown below:

[0043] Length min=min(sqrt((x1-x2)) 2 +(y1-y2) 2 (3)

[0044] The coordinates of the two endpoints of the curve are A(x1, y1) and B(x2, y2).

[0045] Value C =∑ δ δ(i,j) 2 P δ (i,j) (4)

[0046] Where δ(i,j)=|ij|, that is, the gray level difference between adjacent pixels; P δ(i,j) Let δ be the pixel distribution probability with a grayscale difference between adjacent pixels.

[0047] The catalyst active phase identification method based on mathematical morphology and Gabor filtering provided in this invention will be verified below.

[0048] exist Figure 2 In order to ensure the representativeness of the experimental results observed by TEM, 30 images each of different regions of the catalyst sample to be observed were taken, with 30 images taken in both UC-S and SC-S formats, ensuring that the captured area was greater than 30,000 nm. 2 .

[0049] Figure 3 The length is given min =24, value c Visualization of the algorithm when length is 0.7. Increasing length... min With value c This will lead to an increased false negative rate but a decreased false positive rate, while simultaneously improving precision and decreasing recall. Therefore, the F1 score is used to comprehensively evaluate precision and recall within a given length. min =24, value c The F1 score is highest and the various indicators are relatively balanced when the F1 score is 0.7. Figure 3 The length is given min =24, value c The algorithm visualization results at a ratio of 0.7 show that the evaluation index and visualization results demonstrate that the algorithm can effectively identify the active phase of the hydrogenation catalyst.

[0050] Although embodiments of the present invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and variations can be made to the above embodiments without departing from the principles and spirit of the present invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for catalyst active phase identification based on mathematical morphology and Gabor filtering, characterized in that, Includes the following steps: S1. Image acquisition using electron microscopy equipment; After the catalyst is sulfided, it is transferred to a TEM for observation to complete the image acquisition using electron microscopy equipment; S2. Feature Extraction: Gabor filters are used to extract texture and edge features from the catalyst image. And morphological algorithms were used to extract the striations of the hydrogenation catalyst; S3. Skeleton line extraction; The Sobel operator is used to extract the edges of the stripes for skeleton line endpoint detection; S4. Active phase stripe screening: Extract the coordinates of the entire curve in sequence according to the endpoints of the skeleton line to identify the active phase of the hydrogenation catalyst.

2. The method for identifying a catalyst active phase based on mathematical morphology and Gabor filtering according to claim 1, characterized in that, In S1, the vulcanization process is as follows: (1) Take 1g of 80-100 mesh catalyst, place it in a glass reactor, and purge with N2 at 50mL / min for about 10min; (2) Switch to H2 containing 15% H2S, with a flow rate of 100 mL / min and a heating rate of 6℃ / min, raise the temperature to 360±30℃, then change the heating rate to 1℃ / min, raise the temperature to 370±30℃, and maintain for 4 hours. (3) Cool down to 120±30℃ at a cooling rate of 50℃ / min, and then switch to N2 purging for 0.5h; (4) Under N2 protection, the sulfided catalyst is poured into cyclohexane to prevent oxidation, milled for 40 min, and the catalyst and cyclohexane are loaded into the sample bottle.

3. The catalyst active phase identification method based on mathematical morphology and Gabor filtering according to claim 1, characterized in that, In S3, the input image contour with a certain width is transformed into a frame with a width of only one pixel by successively removing the edges.

4. The catalyst active phase identification method based on mathematical morphology and Gabor filtering according to claim 1, characterized in that, In S3, first find 8 neighborhoods centered on the boundary point, denoted as P1. The 8 points in the neighborhood are denoted as P2, P3, ..., P9, clockwise around the center point, where P2 is above P1. The skeleton line extraction is completed according to the following steps: (1). First, mark the boundary points that simultaneously satisfy the following conditions: ①2≤N(P1)≤6; ②S(P1)=1; ③P2*P4*P6=0; ④P4*P6*P8=0; Where N(P1) is the number of non-zero leading points of P1, and S(P1) is the number of times the values ​​of these points change from 0 to 1 when ordered as P2, P3, ..., P9; (2). Same as step (1), change condition ③ to P2*P4*P8=0; change condition ④ to P2*P6*P8=0; after all boundary points have been checked, remove all marked points; (3). Steps (1) and (2) constitute an iteration until no more points satisfy the marking conditions. The remaining points form the region, which is the refined skeleton.

Images