A method for detecting activated sludge state based on settling image texture entropy change characteristics

Abstract

The application discloses a method for detecting activated sludge state based on settling image texture entropy change characteristics, which comprises the following steps: measuring and controlling sludge concentration of a sludge-water mixture sample; placing the sample in a transparent container to make the sludge in a suspended state; fixing the position and focal length of a camera to shoot a video of the activated sludge settling process; cutting a plurality of activated sludge settling images from the video of the activated sludge settling process to obtain an image set; cutting the image set at a proper position; performing pretreatment operation on the cut image set to calculate data of texture entropy change with time at a plurality of different heights; drawing the data into a settling image entropy change hologram; calculating the slope and slope direction of the matrix formed by the settling time, settling height and image texture entropy; drawing a polar coordinate scatter plot of the slope and slope direction data to calculate the normalized sedimentation separation time, normalized clarification degree and normalized compression degree; and analyzing the activated sludge state.

Description

Technical Field

[0001] This invention belongs to the field of water treatment technology, specifically relating to a method for detecting the state of activated sludge based on the texture entropy change characteristics of sedimentation images. Background Technology

[0002] Monitoring the physical properties of activated sludge, especially granular sludge, is crucial for overcoming the technical bottlenecks in stable operation control and for its widespread application. However, the currently detected indicators, such as particle size, settling velocity, SV5 / SV30, and microscopic images, do not provide significant early warning and control for granular sludge systems.

[0003] Settling velocity is a comprehensive parameter reflecting particle size and density. SV5 / SV30 is commonly used as an important parameter to reflect the overall granulation of activated sludge. This is very effective in the early stage of system operation. However, after the particles mature, due to the large heterogeneity of particles in the system, the substances with early warning signals are often in the minority and are easily washed away, which is insufficient to cause significant changes in SV5 / SV30. Therefore, it is impossible to characterize it by SV5 / SV30.

[0004] Particle size is an important indicator for measuring the microscale characteristics of sludge during sludge cultivation. However, a stable particle size does not necessarily mean stable porosity and internal material. Although combining its distribution can reflect small changes in particle size and sedimentation rate to some extent, this is still a static assessment of the system and cannot reflect the multiple factors that cause changes in particle size.

[0005] Roundness and fractal dimension obtained through microscopic image analysis can effectively measure the morphological characteristics of granular sludge, but the process is complex and requires a high level of expertise, which is detrimental to the promotion and application of granular sludge technology. In addition, the parameters currently commonly used mainly characterize the evolutionary characteristics of granular sludge, while there are relatively few indicators for measuring the degradation of granular sludge (such as disintegration and the proliferation of filamentous bacteria).

[0006] Patent application CN116342505A discloses a method and system for detecting the granulation degree of aerobic granular sludge. The method involves conducting a static sedimentation experiment on wastewater samples containing aerobic granular sludge to obtain sludge settling images at different times, and acquiring planar images reflecting the planar distribution of aerobic granular sludge in the unsettled wastewater samples. Two deep learning segmentation network models are constructed to identify the sludge settling images and planar images at different times, respectively, obtaining two different binary images containing binary classification results. A fitted image of the aerobic granular sludge contour and its area are obtained. The sludge settling scale at the target time is obtained. The maturity of the aerobic granular sludge is determined based on the area ratio Y of the aerobic granular sludge and the sludge settling stability index Z.

[0007] Although this technical solution solves the problem of accurate and real-time identification and analysis of the degree of granulation of aerobic activated sludge, the technical route is complex and requires a high level of expertise. It is designed to judge the degree of granulation of granular sludge, but does not have the function of characterizing its internal comprehensive factors. Summary of the Invention

[0008] To overcome the shortcomings of the prior art, the present invention aims to provide an activated sludge state detection method based on the texture entropy change characteristics of sedimentation images. This method extracts holographic data of image texture entropy changes during sludge sedimentation, and through analysis of the polar coordinate scatter distribution characteristics of the image texture entropy, it derives the different sedimentation behavior characteristics of the activated sludge system under different states. The present invention utilizes the image entropy change feature values ​​extracted from sedimentation images to perform qualitative analysis of the activated sludge state, making the results more accurate and objective. It supplements the deficiencies in methods for characterizing the morphological features of activated sludge degradation stages, and has the advantages of being simple, intuitive, accurate, efficient, and low-cost.

[0009] To achieve the above objectives, the technical solution adopted by this invention is as follows:

[0010] A method for detecting the state of activated sludge based on the texture entropy change characteristics of sedimentation images includes the following steps:

[0011] Step 1: Take a certain amount of sludge-water mixture from the aeration terminal as a sample, and measure and control the sludge concentration of the sample.

[0012] Step 2: Place the mud-water mixture sample obtained in Step 1 in a transparent container to ensure that the sludge is in a completely mixed state with the sludge suspended in the water.

[0013] Step 3: Use a camera with a fixed position and focal length to record a video of the activated sludge settling process in the mud-water mixture sample that is suspended in the transparent container obtained in Step 2.

[0014] Step 4: Extract images from the video of the activated sludge settling process captured in Step 3. Take multiple images of activated sludge settling at fixed time intervals to obtain an image set.

[0015] Step 5: Select appropriate locations from the image set obtained in Step 4 as Regions of Interest (ROIs) and crop them.

[0016] Step 6: Perform preprocessing operations on the image set cropped in Step 5, and calculate the data of texture entropy changes over time at multiple different heights;

[0017] Step 7: Draw the data obtained in Step 6 into a hologram of settlement image entropy change, and a matrix formed by settlement time, settlement height, and image texture entropy, and calculate its slope S and aspect A;

[0018] Step 8: Plot a polar coordinate scatter plot of the slope S and aspect A data calculated in Step 7. Select the same time period and divide it into a settling time cluster. The maximum projection length of the scatter points formed by each time cluster on the positive and negative half-axis of the X-axis and Y-axis is recorded as sorting degree, clarification degree and compressibility degree. Calculate the standardized sedimentation sorting time, standardized clarification degree and standardized compressibility degree.

[0019] Step 9: Determine the activated sludge state based on the polar coordinate scatter plot drawn in Step 8 and the calculated standardized sedimentation and sorting time, standardized clarification degree, and standardized compression degree.

[0020] In step 2, the mud-water mixture sample in the transparent container is stirred or aerated to make it suspended in the transparent container.

[0021] In step 3, the shooting conditions are complete darkness. Before shooting the video with the camera, a separate light is placed behind the light-transmitting container to control the light source and make the sludge image clear. The shooting time is not less than the minimum time for the sludge to complete sedimentation. The video should be shot for at least 5 minutes. During shooting, all shooting parameters should be kept consistent.

[0022] Step 4 uses open-source Python code to automatically capture images, starting from the 0th second of settling as the first image, and the final number of images captured is determined based on the completion of sludge settling.

[0023] The appropriate position in step 5 is the center of the transparent container. The region of interest (ROI) is selected as an area of ​​the same size and position where flocs exist, including the image along the longitudinal direction of the sludge. The image along the transverse direction is removed from the area affected by image distortion caused by glass refraction at the edge of the transparent container. The area is free of debris.

[0024] The preprocessing in step 6 includes converting the images in the image set to 8-bit grayscale, improving contrast, and calculating texture entropy, which is implemented by image processing software or open-source vision libraries.

[0025] The formulas for calculating slope S and aspect A in step 7 are as follows:

[0026]

[0027]

[0028] in, It is the rate of change of image entropy in the horizontal direction. It is the rate of change of elevation in the longitudinal direction; among which, , The solution is obtained within a 3×3 moving window and is calculated using the third-order inverse distance squared weighted difference method.

[0029] In step 9, the state of the activated sludge is determined as follows:

[0030] During the flocculent sludge stage, the standardized sedimentation and sorting time, the standardized compression degree, and the standardized clarification degree increase. When the standardized sedimentation and sorting time and the standardized compression degree begin to decrease, the sludge gradually changes from a flocculent state to a granular state until the standardized compression degree stabilizes. The sludge is basically granulated.

[0031] When the sludge exhibits a low standardized sedimentation and sorting time, a high standardized compression degree, and a high standardized clarification degree, it may be in the particle maturation stage. At this time, the particle morphology is stable and the structure is dense, resulting in excellent settling and clarification performance.

[0032] When the standardized sedimentation and sorting time, the standardized compression degree, and the standardized clarification degree increase, it may be the particle disintegration period. At this time, the edges of the granular sludge tend to be rough, the particle structure is loose, and there are a small amount of flocculent sludge in the system, which undergoes flocculation, and the clarification effect and sedimentation performance are weakened.

[0033] When there is a lower standardized sedimentation and sorting time, a lower standardized compression degree, and a higher standardized clarification degree, it may be the large particle sludge stage. At this time, although the particles have a rounded particle shape, the shape is more irregular and the structure is looser than that of mature granular sludge.

[0034] When there is a high standardization sedimentation and sorting time, a high standardization clarification degree, and a low standardization compression degree, it may be the growth period of filamentous bacteria. At this time, the sedimentation performance of granular sludge decreases and there are more floating filamentous substances in the supernatant, making it difficult to compress the sludge.

[0035] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0036] 1. Innovative Data Description: This invention utilizes image processing technology, combined with the texture entropy features of images generated by mud and water during sedimentation, to describe the state of granular sludge. It can provide characteristic information such as the microscopic morphology of particles and the dynamic changes between sludge particles during sedimentation, providing richer data support for the study of granular sludge sedimentation.

[0037] 2. High real-time performance: By adopting real-time image acquisition and processing technology, this invention can monitor the settling process of granular sludge in real time, improve the real-time performance of data acquisition, and provide support for timely adjustment of production process parameters.

[0038] 3. Early warning function: Based on image processing technology and combined with the settling law of granular sludge, this invention can detect abnormalities in advance, providing enterprises with an early warning function, which helps to prevent production accidents and improve production efficiency.

[0039] 4. Superior accuracy and stability: This invention uses image processing algorithms and technologies to process sludge settling images to obtain slope, texture entropy, sedimentation sorting degree, compression degree, and clarification degree to accurately determine the state of granular sludge. It does not rely on the experience of the staff and also improves the measurement accuracy and stability.

[0040] In summary, this invention, through innovative data descriptions such as sedimentation and sorting time, degree of compression, and degree of clarification, boasts high real-time performance, a high level of automation, early warning functions, and good accuracy and stability. It can analyze and judge the state of granular sludge based on the texture entropy change characteristics of sedimentation images, providing guidance for the stable control of granular sludge. Attached Figure Description

[0041] Figure 1 is an overall diagram of the reactor in an embodiment of the present invention; wherein, Figure 1(a) is a diagram of the operating conditions of the granular sludge reactor during the cultivation process in an embodiment of the present invention, and Figure 1(b) is a diagram of the changes in settling performance and the division of stages in the present invention.

[0042] Figure 2 shows the operating conditions and nutrient removal performance changes of the reactor system at different stages during 300 days of operation according to the embodiment of the present invention; wherein, Figure 2(a) is a COD concentration distribution diagram, and Figure 2(b) is an NH4+ concentration distribution diagram. 4+ -N concentration distribution diagram, Figure 2(c) shows NO 3- -N concentration distribution map, Figure 2(d) shows NO 2- -N concentration distribution map.

[0043] Figure 3 is a diagram showing the change in biomass of the reactor in an embodiment of the present invention; wherein, Figure 3(a) is a diagram of MLSS of sludge during the cultivation process, and Figure 3(b) is a curve showing the change in SS of effluent.

[0044] Figure 4 Microscopic images of sludge under typical conditions, taken after 1 day, 37 days, 82 days, 147 days, 180 days, and 289 days of cultivation.

[0045] Figure 5 shows the microstructure of sludge under typical conditions of the present invention; wherein, Figure 5(a) is the sludge particle size distribution diagram and Figure 5(b) is the fractal dimension diagram.

[0046] Figure 6 shows the temporal variation of texture entropy at a typical stage of the present invention.

[0047] Figure 7 shows the slope change diagram corresponding to the temporal change diagram of texture entropy in a typical stage of the present invention.

[0048] Figure 8 shows the changes in image texture entropy features of the present invention; wherein, Figure 8(a) shows the changes in precipitation standardization sorting time during the cultivation process, Figure 8(b) shows the changes in the degree of standardization clarification, and Figure 8(c) shows the changes in the degree of standardization compression.

[0049] Figure 9 This is a schematic diagram illustrating the shooting conditions in an embodiment of the present invention.

[0050] Figure 10 is a schematic diagram of a typical sedimentation separation process; among them, Figure 10(a) is a sedimentation diagram of granular sludge, and Figure 10(b) is a sedimentation diagram of flocculent sludge. Detailed Implementation

[0051] The technical solution of the present invention will be further described below through specific embodiments.

[0052] To address the shortcomings of existing analytical techniques, this invention proposes an image-based method that focuses on the sludge sedimentation process. It analyzes the changing characteristics of particles from the perspectives of sedimentation and sorting time, clarification degree, and compression degree, providing guidance for the stable regulation of activated sludge.

[0053] A method for detecting the state of activated sludge based on the texture entropy change characteristics of sedimentation images includes the following steps:

[0054] Step 1: Take 100ml of the sludge-water mixture at the aeration end as a sample, measure its sludge concentration, and control the sludge concentration within 6000mg / L-8000mg / L;

[0055] Step 2: After measuring the sludge concentration of the mud-water mixture sample obtained in Step 1, place it in a 100ml colorimetric tube. The mud-water mixture sample in the colorimetric tube is stirred or aerated to ensure that the sludge is in a completely mixed state in a suspended state. The force of the action should not be too strong when transferring the mixed sample to prevent the activated sludge from breaking.

[0056] Step 3: Place the colorimetric tube containing the suspended mud-water mixture sample from Step 2 into a completely sealed, opaque box. Place a separate light source behind the colorimetric tube to control the light source and ensure clear imaging of the sludge. The distance between the light source and the mud-water mixture sample should be 5 cm. Use a camera with a fixed position and focal length to record a video. The video duration should be no less than the minimum time required for the sludge to settle. The video should be at least 5 minutes long. During recording, keep all shooting parameters consistent so that the image can accurately reflect the texture information of the flocs.

[0057] Step 4: Extract images from the video of the activated sludge settling process captured in Step 3 at 1-second intervals. Extract multiple images of activated sludge settling. The extracted activated sludge image files should be in .tiff format to ensure the integrity of the image data to the greatest extent possible, thus obtaining an image set.

[0058] Step 5: Select the middle position of the colorimeter tube as the region of interest (ROI) from the image set obtained in Step 4 and crop it. The selected region of interest (ROI) should be a region of the same size and position, and this region should contain flocs, without any foreign objects or interference from refraction caused by glass bending. The image length * width is 1038 * 126 pixels.

[0059] Step 6: Import the image set cropped in Step 5 into ImageJ software. Perform preprocessing operations such as 8-bit grayscale conversion and contrast enhancement. When enhancing contrast, select 0.5% for Saturated pixels and choose options such as Normalize and Equalize histogram. Select a Region of Interest (ROI) for every 1 ml range of the image, and select a total of 100 ROIs vertically. Calculate GLCM for each ROI, select Entropy for the feature value, set the direction to 90°, and set the step size to 1 pixel. Save the automatically calculated results in .csv format files, generating a total of 100 .csv files showing the change of image texture entropy over time at different heights. Merge the .csv files showing the change of image texture entropy over time at different heights into a single .csv file.

[0060] Step 7: Using the .csv file obtained in Step 6, plot a hologram of settlement image entropy change with settlement time as the x-axis, settlement height as the y-axis, and image texture entropy as the z-axis. Calculate the slope and aspect of the matrix formed by settlement time, settlement height, and image texture entropy. The slope S and aspect A at a point in the settlement entropy change hologram are functions of the settlement entropy function E = f(x, y) in the horizontal and vertical directions, respectively.

[0061]

[0062]

[0063] in, It is the rate of change of image entropy in the horizontal direction. This is the rate of change of elevation in the longitudinal direction. Among them, , The solution is obtained within a 3×3 moving window, using the third-order inverse distance squared weighted difference method. The final calculated slope is a counterclockwise rotation starting from the positive x-axis (0°) and ranging from 0° to 360°.

[0064] Step 8: Plot a polar coordinate scatter plot of the slope and aspect data calculated in Step 7. Divide the data into a settling time cluster every 25 seconds. The maximum projection length of the scatter points formed by each time cluster on the positive and negative half-axes of the X and Y axes is recorded as sorting degree, clarification degree, and densification degree. Read the data once every 1 second and calculate the time when the inflection point of sorting degree, clarification degree, and densification degree appears during the settling process. Divide these values ​​by the reactor settling time for this stage to obtain the normalized sorting time, normalized clarification degree, and normalized densification degree.

[0065] Step 9: Determine the activated sludge state based on the polar coordinate scatter plot drawn in Step 8 and the calculated standardized sedimentation and sorting time, standardized clarification degree, and standardized compression degree.

[0066] During the flocculent sludge stage, the standardized sedimentation and sorting time, the standardized compression degree, and the standardized clarification degree increase. When the standardized sedimentation and sorting time and the standardized compression degree begin to decrease, the sludge gradually changes from a flocculent state to a granular state until the standardized compression degree stabilizes. The sludge is basically granulated.

[0067] When the sludge exhibits a low standardized sedimentation and sorting time, a high standardized compression degree, and a high standardized clarification degree, it may be in the particle maturation stage. At this time, the particle morphology is stable and the structure is dense, resulting in excellent settling and clarification performance.

[0068] When the standardized sedimentation and sorting time, the standardized compression degree and the standardized compression degree increase, it may be the particle disintegration period. At this time, the edges of the granular sludge tend to be rough, the particle structure is loose, and there are a small amount of flocculent sludge in the system. Flocculation occurs, and the clarification effect and sedimentation performance are weakened.

[0069] When there is a lower standardized sedimentation and sorting time, a lower standardized compression degree, and a higher standardized clarification degree, it may be the large particle sludge stage. At this time, although the particles have a rounded particle shape, the shape is more irregular and the structure is looser than that of mature granular sludge.

[0070] When there is a high standardization sedimentation and sorting time, a high standardization clarification degree, and a low standardization compression degree, it may be the growth period of filamentous bacteria. At this time, the sedimentation performance of granular sludge decreases and there are more floating filamentous substances in the supernatant, making it difficult to compress the sludge.

[0071] Example

[0072] Step 1: Sludge from a wastewater treatment plant in Xi'an was inoculated into an AGS reactor built in the laboratory. The reactor had an effective volume of 12.37 L, a working height of 700 mm, and an inner diameter of 150 mm. Aeration was achieved using a bottom-mounted microporous aerator, with a rotor flow meter controlling the aeration rate. A bottom-inlet water intake method was used. The reactor's operation was automatically controlled by a time-programmed controller. The start-up and shutdown times of the influent, aeration, settling, and drainage processes were adjusted according to experimental needs. The entire experiment was conducted at room temperature (23 ± 3℃). The operating conditions of the activated sludge reactor during the cultivation process are shown in Figure 1. The operating conditions and nutrient removal performance of the reactor system during 300 days of operation at different stages are shown in Figure 2. The MLSS and effluent SS variation curves of the sludge during the cultivation process are shown in Figure 3. Microscopic photographs of the sludge in a typical state are shown in Figure 4. Figure 4 As shown in Figure 5, the particle size distribution and fractal dimension of sludge under typical conditions are as follows;

[0073] Step 2: Take a sample of the sludge-water mixture from the aeration end of the activated sludge reactor at the same time point throughout the day. Place 100ml of the sample in a 100ml colorimetric tube, ensuring the sample is in a completely mixed, suspended state. Place the sample in a completely sealed, light-proof space, and use a photographic device such as... Figure 9 As shown;

[0074] Step 3: When the sample is completely mixed, immediately start recording the sample with the same shooting parameters on the camera. The recording time should be no less than 5 minutes. Stop recording when the sedimentation is complete to obtain a complete video of sludge sedimentation.

[0075] Step 4: Import the obtained sample sludge settling video into the image capture software, set the time interval to 1 second, the captured image file format to .tiff, and store the captured image set in the same folder;

[0076] Step 5: Import the folder obtained in Step 4 into ImageJ software as Image Sequence. Use the rectangular marquee tool to select the center of the colorimeter tube as the region of interest (ROI). The selected ROIs should be the same size and position, free from debris and interference from refraction caused by glass bending. Crop the image with a length * width of 1038 * 126 pixels and a height ROI range of 0-100ml. After selection, click Duplicate to crop.

[0077] Step 6: Convert the image set obtained in Step 5 to 8-bit grayscale, then perform contrast enhancement and contrast balancing operations. Use the Enhance Contrast function, set the Saturated pixels to 0.5%, and select options such as Normalize, Equalize histogram, and Process all slices. After setting, click OK to complete the settings. Select a Region of Interest (ROI) for every 1 ml range of the image, and select a total of 100 ROIs vertically. Calculate GLCM for each ROI, select Entropy as the feature value, set the direction to 90°, and set the step size to 1 pixel. Save the automatically calculated results in .csv format files, generating a total of 100 .csv files showing the change of image texture entropy over time at different heights. Merge the .csv files showing the change of image texture entropy over time at different heights into a single .csv file, where the column index includes settling time, height, and image texture entropy.

[0078] Step 7: Using the .csv file obtained in Step 6, plot a hologram of sedimentation image entropy change with settling time as the x-axis, settling height as the y-axis, and image texture entropy as the z-axis. The holograms of sedimentation image entropy change for sludge in different states are shown in Figures 6(a)-6(e). Calculate the slope and aspect of the matrix formed by settling time, settling height, and image texture entropy. The slope S and aspect A at a point in the sedimentation entropy change hologram are functions of the sedimentation entropy function E = f(x, y) in the horizontal and vertical directions, respectively, i.e.:

[0079]

[0080]

[0081] in, It is the rate of change of image entropy in the horizontal direction. This is the rate of change of elevation in the longitudinal direction. Among them, , The solution is performed within a 3×3 moving window using a third-order inverse distance squared weighted difference method. The final calculated slope aspect is obtained by rotating counterclockwise from the positive x-axis (0°) to a range of 0°~360°.

[0082] Step 8: Plot a polar coordinate scatter plot of the slope and aspect data calculated in Step 7, dividing the data into a settling time cluster every 25 seconds. The changes in the scatter plot of slope and aspect of sludge in different states are shown in Figures 7(a)-7(e). The maximum projection length of the scatter points formed by each time cluster on the positive and negative half-axes of the X and Y axes is denoted as sorting degree, clarification degree, and densification degree. Data is read every 1 second, and the time when sorting degree, clarification degree, and densification degree occur during the settling process are calculated. Dividing these values ​​by the settling time of the reactor at this stage yields the normalized sorting time, normalized clarification degree, and normalized densification degree. A typical sedimentation separation process diagram is shown below. Figure 9 As shown.

[0083] Step 9: Use the polar coordinate scatter plot drawn in Step 8, along with the calculated standardized sedimentation and sorting time, standardized clarification degree, and standardized compression degree, to analyze the activated sludge state.

[0084] During the flocculent sludge stage, the standardized sedimentation and sorting time, the standardized compression degree, and the standardized clarification degree increase. When the standardized sedimentation and sorting time and the standardized compression degree begin to decrease, the sludge gradually changes from a flocculent state to a granular state until the standardized compression degree stabilizes. The sludge is basically granulated.

[0085] When the sludge exhibits a low standardized sedimentation and sorting time, a high standardized compression degree, and a high standardized clarification degree, it may be in the particle maturation stage. At this time, the particle morphology is stable and the structure is dense, resulting in excellent settling and clarification performance.

[0086] When the standardized sedimentation and sorting time, the standardized clarification degree, and the standardized compression degree increase, it may be the particle disintegration period. At this time, the edges of the granular sludge tend to be rough, the particle structure is loose, and there are a small amount of flocculent sludge in the system, which undergoes flocculation, and the clarification effect and sedimentation performance are weakened.

[0087] When there is a lower standardized sedimentation and sorting time, a lower standardized compression degree, and a higher standardized clarification degree, it may be the large particle sludge stage. At this time, although the particles have a rounded shape, their shape is more irregular and the structure is looser than that of mature granular sludge.

[0088] When there is a high standardization sedimentation and sorting time, a high standardization clarification degree, and a low standardization compression degree, it may be the growth period of filamentous bacteria. At this time, the sedimentation performance of granular sludge decreases and there are more floating filamentous substances in the supernatant, making it difficult to compress the sludge.

[0089] Example Analysis:

[0090] Based on the changes in the settling performance of activated sludge, this example divides the cultivation time into 5 stages: cultivation period, particle maturation period (during which the sludge is flocculent), disintegration period, large particle sludge period, and filamentous bacteria growth period; these are referred to as stages P1-P5 respectively. The cultivation process is divided into several stages: the initial stage (day 1 to day 43) during which the average particle size of the activated sludge increases slightly from 65µm to 134µm, and the sludge begins to granulate from a flocculent structure; the maturation stage (days 44 to 119) during which the particle size remains stable at around 300µm, and the particle structure is dense; the disintegration stage (days 120 to 159) during which the average particle size gradually decreases to 283µm, mainly due to the breakup of large particles and a significant increase in the proportion of small particles, resulting in a looser particle structure; the large-particle sludge stage (days 160 to 228) during which the particle size gradually increases from 333µm to 640µm, and the particles re-granulate, with a larger average particle size and a looser structure compared to the mature sludge; and the filamentous growth stage (days 229 to 300) during which the particle structure is surrounded by a large amount of filamentous material, leading to poorer settling performance.

[0091] When sludge exists in flocculent form, its morphology gradually becomes granular, flocculation weakens or disappears, and settling performance is greatly improved. During settling, most of the process is clarification, and the netting and sweeping gradually disappear. Most of the settling process is compression. In flocculent sludge settling, most of the process is compression, and the voids are small. With the formation of granules, flocculation is improved, and the flocculation capacity is greatly enhanced. Flocs are more likely to flocculate together, improving the void effect and enhancing the capacity. In flocculent sludge systems, there is no particle sorting. With granulation, the particle size and morphology vary greatly, and the sorting effect is obvious, with the sludge washing effect occurring. The degree of compression decreases significantly, and the sedimentation and sorting time, degree of compression, and degree of clarification all decrease significantly.

[0092] During the granular maturity stage, the granules are stable in morphology and have a dense structure, exhibiting excellent settling and clarification properties. Granular sludge undergoes clarification during sedimentation and exhibits minimal flocculation. At this stage, the system consists mostly of mature granular sludge, with good homogeneity in sludge particle size and morphology. This results in a low sedimentation and sorting time, with minimal sorting. During sedimentation, clarification occurs simultaneously with sedimentation, with almost no compression. Clarification is completed when sedimentation is finished, with a clarification degree close to 1, and a high degree of compression and clarification.

[0093] In the initial stage of particle disintegration, the particle structure becomes loose, initially affecting the degree of compression, which increases significantly at this point, while the sorting time and clarification remain unchanged. Under the impact of hydraulic action, flocculent sludge begins to detach from the loose particle structure. At this stage, the particle size distribution is relatively uneven, the sorting time increases, and the smaller flocculent sludge has poor settling performance, thus increasing the degree of clarification. The degree of compression remains unchanged and is still at a high level. In the later stage of particle disintegration, these flocculent sludges begin to granulate under hydraulic action, the degree of clarification begins to decrease slightly, and the degree of compression decreases, but the particle size distribution remains uneven, and the sorting time remains unchanged and remains at a relatively high value. After the particle disintegration period, there is no longer a large amount of flocculent sludge, the sludge settling performance recovers, the homogeneity of the particle size distribution increases, and the sorting time begins to decrease.

[0094] During the large granular sludge stage, although the granules have a rounded shape, their shape is more irregular and the structure is looser than that of mature granular sludge. The particle size and morphology of the granular sludge in the system are less homogeneous. Compared with mature granular sludge, it has a higher sorting effect, a longer sedimentation and sorting time, and a higher degree of compression. Because it is still a granular sludge system, it still has a high degree of clarification and a low degree of compression.

[0095] During the growth stage of filamentous bacteria, the settling performance of granular sludge decreases, and compared with mature granular sludge, it has a longer sorting time. Due to the presence of more floating filamentous matter in the supernatant, the supernatant is difficult to clarify, resulting in a lower degree of clarification and difficulty in clarification.

[0096] As shown in Figure 1, the settling time is relatively long and unstable only during the floc growth period, and remains stable in the later stage. The SV5 / SV30 value is large and unstable during the particle formation period, and is maintained at about 1 during the later cultivation process. The value increases and fluctuates during the disintegration period.

[0097] As shown in Figure 2, granular sludge maintains good COD removal capacity throughout its life cycle, while filamentous bacteria maintain good NH4+ removal capacity during their growth phase. 4 + The removal ability of -N is not ideal.

[0098] As shown in Figure 3, during the cultivation of granular sludge, the MLSS of the sludge gradually accumulates with the growth process, reaching its maximum value during the large granular sludge stage, and then declines during the filamentous bacteria growth stage with a disordered trend. The SS in the effluent fluctuates during the formation stage.

[0099] like Figure 4 As shown, the microscopic characteristics of sludge under typical conditions - microscopic images show the morphological changes of granular sludge at each stage from cultivation to disintegration.

[0100] Figure 5 shows the microscopic morphology of sludge in a typical state. Figure 5(a) shows the particle size distribution of sludge during the sludge cultivation process: in the early stage of particle cultivation, the particle size is mostly below 100 μm, and in the mature stage, the particle size increases to 300-500 μm. In the early stage of disintegration, the particle size range is wide and the distribution is uniform. The distribution range of large particle sludge is also wide, but the proportion of particles >500 μm is relatively large. Figure 5(b) shows the fractal dimension diagram, which shows an upward trend, and the particle structure gradually becomes more complex.

[0101] Figure 6 shows the temporal variation of texture entropy in a typical stage (Figure 6(a)-Figure 6(e)); Figure 7 shows the slope variation corresponding to the temporal variation of texture entropy in a typical stage (Figure 7(a)-Figure 7(e)).

[0102] As shown in Figure 8, in the early stage of cultivation, there is no particle sorting effect in the flocculent sludge system. With granulation, the standardized sedimentation and sorting time, standardized compression degree, and standardized clarification degree decrease significantly. In the early stage of disintegration, the standardized sorting time and standardized clarification degree remain unchanged. In the later stage of particle disintegration, the standardized clarification degree begins to decrease slightly, the standardized compression degree decreases, and the standardized sorting time remains unchanged, still maintaining a relatively high value. In the large particle sludge stage, compared with the mature granular sludge, it has a higher standardized sedimentation and sorting time and a higher standardized compression degree. In the filamentous bacteria growth stage, it has a higher standardized sorting time, a lower standardized clarification degree, and is difficult to clarify.

[0103] like Figure 9 The diagram shown is a schematic diagram of the shooting conditions in an embodiment of the present invention.

[0104] Figure 10 shows the settling diagram of the activated sludge of the present invention, and the settling diagram of granular sludge has a better characterization effect than that of flocculent sludge.

[0105] Compared with the prior art, the present invention has the following advantages:

[0106] 1. Real-time monitoring: Image texture analysis technology can be used to monitor the microscopic features of the granular sludge system in real time, providing more timely and accurate information.

[0107] 2. Early warning: By identifying abnormal fine particles, early warnings can be provided before problems occur, which helps to avoid instability and failure of granular sludge systems.

[0108] 3. Abundant data: Traditional physical property monitoring methods can only provide information on static attributes, while image texture analysis methods can provide richer characteristic information of granular sludge systems, thus providing a more comprehensive understanding of the system's operating status.

[0109] 4. High level of automation: The operation steps in this invention, such as image capture, image processing, and texture entropy calculation, can be automated with the help of certain equipment, which effectively improves the level of automated acquisition and analysis of granular sludge settling data, and can be integrated into online monitoring equipment, which is conducive to realizing the automation and unmanned operation of water plants.

[0110] In summary, the stability monitoring method for aerobic granular sludge systems based on image texture analysis proposed in this invention has significant beneficial effects. It can provide wastewater treatment plants with more stable and efficient operation, and also provides a new means for the widespread application and promotion of activated sludge, especially aerobic granular sludge technology.

[0111] This invention is not limited to the above embodiments. Based on the technical solutions disclosed in this invention, those skilled in the art can make some substitutions and modifications to some of the technical features without creative effort, and all such substitutions and modifications are within the protection scope of this invention.

Claims

1. A method for detecting the state of activated sludge based on the texture entropy change characteristics of sedimentation images, characterized in that: Includes the following steps: Step 1: Take a certain amount of sludge-water mixture from the aeration terminal as a sample, and measure and control the sludge concentration of the sample. Step 2: Place the mud-water mixture sample obtained in Step 1 in a transparent container to ensure that the sludge is in a completely mixed state with the sludge suspended in the water. Step 3: Use a camera with a fixed position and focal length to record a video of the activated sludge settling process in the mud-water mixture sample that is suspended in the transparent container obtained in Step 2. Step 4: Extract images from the video of the activated sludge settling process captured in Step 3. Take multiple images of activated sludge settling at fixed time intervals to obtain an image set. Step 5: Select appropriate locations from the image set obtained in Step 4 as Regions of Interest (ROIs) and crop them. Step 6: Perform preprocessing operations on the image set cropped in Step 5, and calculate the data of texture entropy changes over time at multiple different heights; Step 7: Draw the data obtained in Step 6 into a hologram of settlement image entropy change, and a matrix formed by settlement time, settlement height, and image texture entropy, and calculate its slope S and aspect A; The formulas for calculating the slope S and aspect A are as follows: in, It is the rate of change of image entropy in the horizontal direction. It is the rate of change of elevation in the longitudinal direction; among which, , The solution is performed within a 3×3 moving window, using the third-order inverse distance squared weighted difference method. The final calculated slope is a counterclockwise rotation starting from the positive x-axis (0°) and ranging from 0° to 360°. Step 8: Plot a polar coordinate scatter plot of the slope and aspect data calculated in Step 7. Divide the data into a settling time cluster every 25 seconds. The maximum projection length of the scatter points formed by each time cluster on the positive and negative half-axes of the X and Y axes is recorded as sorting degree, clarification degree, and compressibility degree. Read the data once every 1 second and calculate the time when the inflection point of sorting degree, clarification degree, and compressibility degree appears during the settling process. Divide these values ​​by the reactor settling time of this stage to obtain the standardized sedimentation sorting time, standardized clarification degree, and standardized compressibility degree. Step 9: Determine the activated sludge state based on the polar coordinate scatter plot drawn in Step 8 and the calculated standardized sedimentation and sorting time, standardized clarification degree, and standardized compression degree.

2. The activated sludge state detection method based on sedimentation image texture entropy change characteristics according to claim 1, characterized in that: In step 2, the mud-water mixture sample in the transparent container is stirred or aerated to make it suspended in the transparent container.

3. The activated sludge state detection method based on sedimentation image texture entropy change characteristics according to claim 1, characterized in that: In step 3, the shooting conditions are complete darkness. Before shooting the video with the camera, a separate light is placed behind the light-transmitting container to control the light source and make the sludge image clear. The shooting time is not less than the minimum time for the sludge to complete sedimentation. The video should be shot for at least 5 minutes. During shooting, all shooting parameters should be kept consistent.

4. The activated sludge state detection method based on sedimentation image texture entropy change characteristics according to claim 1, characterized in that: Step 4 uses open-source Python code to automatically capture images, starting from the 0th second of settling as the first image, and the final number of images captured is determined based on the completion of sludge settling.

5. The activated sludge state detection method based on sedimentation image texture entropy change characteristics according to claim 1, characterized in that: The appropriate position in step 5 is the center of the transparent container. The region of interest (ROI) is selected as an area of ​​the same size and position where flocs exist, including the image along the longitudinal direction of the sludge. The image along the transverse direction is removed from the area affected by image distortion caused by glass refraction at the edge of the transparent container. The area is free of debris.

6. The activated sludge state detection method based on sedimentation image texture entropy change characteristics according to claim 1, characterized in that: The preprocessing in step 6 includes converting the images in the image set to 8-bit grayscale, improving contrast, and calculating texture entropy, which is implemented by image processing software or open-source vision libraries.

7. The activated sludge state detection method based on sedimentation image texture entropy change characteristics according to claim 1, characterized in that: In step 9, the state of the activated sludge is determined as follows: During the flocculent sludge stage, the standardized sedimentation and sorting time, the standardized compression degree, and the standardized clarification degree increase. When the standardized sedimentation and sorting time and the standardized compression degree begin to decrease, the sludge gradually changes from a flocculent state to a granular state until the standardized compression degree stabilizes. The sludge is basically granulated. When the sludge exhibits a low standardized sedimentation and sorting time, a high standardized compression degree, and a high standardized clarification degree, it may be in the particle maturation stage. At this time, the particle morphology is stable and the structure is dense, resulting in excellent settling and clarification performance. When the standardized sedimentation and sorting time, the standardized compression degree, and the standardized clarification degree increase, it may be the particle disintegration period. At this time, the edges of the granular sludge tend to be rough, the particle structure is loose, and there are a small amount of flocculent sludge in the system, which undergoes flocculation, and the clarification effect and sedimentation performance are weakened. When there is a lower standardized sedimentation and sorting time, a lower standardized compression degree, and a higher standardized clarification degree, it may be the large particle sludge stage. At this time, although the particles have a rounded particle shape, the shape is more irregular and the structure is looser than that of mature granular sludge. When there is a high standardization sedimentation and sorting time, a high standardization clarification degree, and a low standardization compression degree, it may be the growth period of filamentous bacteria. At this time, the sedimentation performance of granular sludge decreases and there are more floating filamentous substances in the supernatant, making it difficult to compress the sludge.

Images