An intelligent pipeline defect identification system and method based on pipeline network microorganisms

Abstract

The present invention discloses a pipeline defect intelligent identification system and method based on pipeline network microorganisms, which belongs to the field of pipeline monitoring technology. The system includes: a marker microorganism delivery module, which delivers marker microorganisms in the pipeline; a water sample collection module, which collects water samples at designated positions in the pipeline; an electrostatic field separation module, which separates particles of different sizes and properties in the water sample at each water sample collection position through electrostatic field separation technology; photographs and image segmentation of the separated particles to obtain a target area image; a microorganism identification module, which is used to extract features from each target area image, identify the type of microorganism, and obtain the marker microorganism concentration at each sampling position; a defect identification module, which identifies whether there are defects in the pipeline to be monitored, as well as the location and severity of the defects. The present invention uses the concentration change of marker microorganisms in the pipeline to indicate defects, and has simple operation, low cost, short detection cycle, and can be fully monitored in real time.

Description

Technical Field

[0001] The present invention belongs to the technical field of pipeline monitoring, and in particular relates to a pipeline defect intelligent identification system and method based on pipeline network microorganisms. Background Art

[0002] As an important infrastructure, pipelines undertake the key task of fluid transportation and play an irreplaceable role in the transportation of resources such as oil, natural gas, and water. However, due to long-term use, external environmental influences, and the corrosion of internal fluids, various defects such as corrosion, cracks, deformation, and deposition may occur inside and outside the pipelines. These defects will not only affect the normal operation of the pipelines and reduce the transportation efficiency, but may also cause serious safety accidents and threaten people's lives and property.

[0003] Traditional pipeline defect detection methods mainly include ultrasonic testing, X-ray testing, magnetic particle testing, etc. Although these methods have high detection accuracy, they have limitations such as complex operation, high cost, and long detection cycle. Especially for large and complex pipeline systems, traditional detection methods are difficult to achieve comprehensive and real-time monitoring. In addition, these methods often require professional inspection personnel and equipment, and the detection difficulty and cost will increase significantly for remote areas or difficult-to-access pipeline sections. Therefore, it is particularly important to develop an efficient, low-cost detection method that can comprehensively monitor the pipeline status. Summary of the invention

[0004] The present invention proposes a pipeline defect intelligent identification system and method based on pipeline network microorganisms. The system and method use the concentration change of marker microorganisms in the pipeline to indicate pipeline defects. By placing marker microorganisms at fixed points and in fixed quantities, and using automated detection means to monitor their concentration distribution in the pipeline, combined with image processing and artificial intelligence recognition technology, intelligent identification of pipeline defects is achieved, solving the technical problems of existing pipeline defect detection, which are complex operation, high cost, long detection cycle, and difficulty in achieving comprehensive real-time monitoring.

[0005] The present invention provides the following technical solutions:

[0006] In a first aspect, a pipeline defect intelligent identification system based on pipeline network microorganisms is provided, comprising:

[0007] A marker microorganism delivery module is used to deliver marker microorganisms at fixed points and in fixed quantities in the pipeline to be monitored;

[0008] A water sample collection module is used to collect water samples from a designated location of the pipeline to be tested after a preset period of time after the marker microorganisms are placed;

[0009] An electrostatic field separation module, which is used for each water sample collection position to separate particles of different sizes and properties in the water sample through electrostatic field separation technology. The particles include marker microorganisms, other microorganisms, and impurities;

[0010] An imaging module, which is used to take pictures and perform image segmentation on the separated particles to obtain a target area image containing the morphology of microorganisms or the parts required by the set requirements;

[0011] A microorganism recognition module, which is used to extract features from each target area image and identify the types of microorganisms to obtain the concentration of marker microorganisms at each sampling position;

[0012] A defect recognition module, which is used to identify whether there are defects in the pipeline to be monitored, as well as the location and severity of the defects based on the preset concentration gradient threshold of the marker microorganisms and the concentration of the marker microorganisms at different positions of the pipeline to be monitored.

[0013] Optionally, the electrostatic field separation module adopts a multi-stage electrostatic separator structure, and the electrostatic parameters of the electrostatic separator at each stage can be adjusted to regulate the intensity and direction of the electrostatic field;

[0014] The separation of particles of different sizes and properties in the water sample through electrostatic field separation technology is specifically as follows: when the water sample is collected and passes through the electrostatic separator, the electrostatic separator applies different electric field forces to all microorganisms and impurities in the water sample, and the separation is achieved by using the charge difference of different microorganisms and impurities in the water sample.

[0015] Optionally, the imaging module includes a high-resolution camera and a segmentation unit; the high-resolution camera takes pictures of the separated particles to obtain the original captured image; the segmentation unit performs segmentation on the original captured image through a fully automatic segmentation method to segment out the target area image containing the morphology of microorganisms or the parts required by the set requirements;

[0016] The fully automatic segmentation method specifically includes:

[0017] Input the original captured image; convert the original captured image into a grayscale image; enhance the contrast through point operation and perform Gaussian filtering for denoising; use the Canny edge detection operator to detect the boundary of each microorganism; close the edge image through morphological processing and obtain the final segmentation result.

[0018] Optionally, the microorganism recognition module includes a feature extraction unit; the feature extraction unit is used to extract image features from each target area image;

[0019] The extracted image features include geometric features, internal structure histogram features, Fourier descriptors, gray-level co-occurrence matrices, and rotation-invariant local binary patterns;

[0020] The geometric features include: the area of the microorganism, the perimeter of the boundary contour of the microorganism, the roundness of the microorganism, the major axis length and minor axis length of the ellipse having the same normalized second-order central moment as the microorganism region;

[0021] The internal structure histogram features are used to describe the internal spatial structure of the image and reflect the distribution of pixel intensities in the image;

[0022] The Fourier descriptor distinguishes different contours through frequency domain information to identify objects and is used to reflect the contour features of the image;

[0023] The gray-level co-occurrence matrix is used to reflect the spatial correlation of the gray values of any two pixel points in the image;

[0024] The rotation-invariant local binary pattern is a robust feature operator that is invariant to any monotonic transformation of a grayscale image and has gray-scale invariance and rotation invariance, and is used to reflect the texture features of the image.

[0025] Optionally, the microorganism recognition module further includes an identification unit and a marked microorganism concentration acquisition unit; the identification unit optimizes the parameters of the SVM model using the particle swarm optimization algorithm, and classifies the features of the extracted target region image using the optimized and trained SVM model to identify the type of each microorganism; the marked microorganism concentration acquisition unit counts the marked microorganisms for each sampling position, and obtains the concentration of the marked microorganisms according to the number of marked microorganisms and the volume of the collected water sample.

[0026] Optionally, the process of optimizing the parameters of the SVM model using the particle swarm optimization algorithm is as follows:

[0027] According to the preset particle swarm size, randomly initialize the position and velocity of each particle; each particle represents a combination of SVM model parameters;

[0028] Use the SVM parameter combination represented by each particle to train the SVM model, and calculate the classification accuracy of the SVM model using the test set in the training process, and use the classification accuracy of the SVM model as the fitness value of the particle;

[0029] Update the individual historical optimal position and global optimal position of the particle swarm according to the fitness value of the particle;

[0030] Update the velocity and position of each particle according to the updated individual optimal position and global optimal position of the particle swarm;

[0031] Repeatedly update the individual optimal position and the global optimal position of the particles until the termination condition is met. Output the global optimal position and the corresponding global optimal fitness value. The parameter combination of the SVM model corresponding to the particle with the global optimal position is the optimal parameter combination of the SVM model. The optimal parameter combination of the SVM model includes the penalty factor C and the kernel parameter 。

[0032] Optionally, the objective function of the SVM model during training is:

[0033]

[0034] The constraint conditions are:

[0035]

[0036]

[0037] where and are the Lagrange multipliers of training samples i and j respectively, , N is the number of samples used for model training, and are the feature data of training samples i and j respectively; and are the class labels of training samples i and j respectively, is the RBF kernel function; C is the penalty factor;

[0038] The decision function of the SVM model is:

[0039]

[0040]

[0041] where b is the constant bias term, x is the feature data of the target region image to be classified, is the kernel parameter.

[0042] Optionally, the defect identification module is used to identify whether there are defects in the pipeline to be monitored, as well as the location and severity of the defects, based on the preset concentration gradient threshold of the marker microorganism and the concentration of the marker microorganism at different positions of the pipeline to be monitored. The specific process is as follows:

[0043] Set the first concentration threshold range, the second concentration threshold range, and the third concentration threshold range; the first concentration threshold range is between the second concentration threshold ranges; the second concentration threshold range is between the third concentration threshold ranges;

[0044] If the concentration of the marker microorganism at the current position is within the first concentration threshold range, there is no defect at the current position of the pipeline to be monitored;

[0045] If the concentration of the marker microorganism at the current position exceeds the upper limit of the first concentration threshold range, there are defects such as leakage, corrosion or other defects that cause the increase of the microorganism concentration at the current position of the pipeline to be monitored; if the concentration of the marker microorganism at the current position is lower than the lower limit of the first concentration threshold range, there are defects such as poor water flow, sediment accumulation or other defects that cause the decrease of the microorganism concentration at the current position of the pipeline to be monitored;

[0046] If the concentration of the marker microorganism at the current position is outside the first concentration threshold range and within the second concentration threshold range, the severity of the defect is primary; if the concentration of the marker microorganism at the current position is outside the second concentration threshold range and within the third concentration threshold range, the severity of the defect is intermediate; if the concentration of the marker microorganism at the current position is outside the third concentration threshold range, the severity of the defect is severe.

[0047] Optionally, the marker microorganism injection module is used to inject the marker microorganism at fixed points and in fixed quantities in the pipeline to be monitored. The injection positions of the marker microorganism are at the connection points, branches and turning points of the pipeline to be monitored, and the injection positions of the marker microorganism are given unique coordinate values;

[0048] The water sample collection module is used to collect water samples from the specified positions of the pipeline to be detected after a preset time period after injecting the marker. The specified positions for collecting water samples are the inlet, outlet, connection points, branches and turning points of the pipeline to be monitored.

[0049] In a second aspect, a method for intelligent identification of pipeline defects based on network microorganisms is provided, including:

[0050] Inject the marker microorganism at fixed points and in fixed quantities in the pipeline to be monitored;

[0051] After injecting the marker microorganism, collect water samples from the specified positions of the pipeline to be detected after a preset time period;

[0052] For each water sample collection position, separate the particles of different sizes and properties in the water sample thereof by the electrostatic field separation technology. The particles include marker microorganisms, other microorganisms and impurities;

[0053] Take pictures and perform image segmentation on the separated particles to obtain a target area image including the microorganism morphology or the set required part;

[0054] Extract features from each target area image, identify the microorganism species, so as to obtain the concentration of the marker microorganism at each sampling position;

[0055] Based on the preset concentration gradient threshold of marker microorganisms and the concentrations of marker microorganisms at different positions of the pipeline to be monitored, identify whether there are defects in the pipeline to be monitored, as well as the location and severity of the defects.

[0056] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0057] The present invention utilizes the concentration change and distribution characteristics of marker microorganisms in the pipeline to realize the intelligent identification of pipeline defects. It is not only simple to operate, but also has high accuracy and sensitivity, and can timely detect tiny defects inside the pipeline, providing an important basis for the maintenance and safe operation of the pipeline. In addition, by combining automated monitoring means and image processing technology, the present invention can realize the comprehensive real-time monitoring of the pipeline state, improve the efficiency and accuracy of pipeline maintenance, reduce errors and omissions caused by manual monitoring, and reduce maintenance costs. Description of the Drawings

[0058] Figure 1 is the structural framework diagram of the intelligent pipeline defect identification system based on pipe network microorganisms of the present invention;

[0059] Figure 2 is the working flow chart of the identification unit of the intelligent pipeline defect identification system based on pipe network microorganisms of the present invention;

[0060] Figure 3 is the flow chart of the intelligent pipeline defect identification method based on pipe network microorganisms of the present invention. Detailed Embodiments

[0061] The present invention will be further described below with reference to the drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and cannot be used to limit the protection scope of the present invention. It should be noted that the term "including" in the specification and claims of the present invention and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or are inherent to these processes, methods, products or devices.

[0062] Embodiment 1

[0063] As Figure 1 shown, an intelligent pipeline defect identification system based on pipe network microorganisms is provided, including: a marker microorganism injection module, a water sample collection module, an electrostatic field separation module, an imaging module, a microorganism identification module, and a defect identification module.

[0064] Optionally, the labeled microorganism injection module, water sample collection module, and electrostatic field separation module are deployed at the terminal; the imaging module, microorganism recognition module, and defect recognition module are deployed at the server side.

[0065] I. Labeled Microorganism Injection Module

[0066] It is used to inject labeled microorganisms at fixed points and in fixed quantities into the pipeline to be monitored.

[0067] When injecting labeled microorganisms, the fixed-point injection of labeled microorganisms is achieved based on the positioning method of the pipeline drawing and the geographic information system, and the coordinates of the injection point are recorded in the geographic information system to indicate the position of the injection point. The injection point is selected at the connection point, branch point, or turning point of the pipeline to be monitored.

[0068] For a single injection of labeled microorganisms, the concentration of the labeled microorganism suspension is adjusted according to the size and water flow velocity of the pipeline, and a flow meter is used to ensure the accuracy of the number of microorganisms injected each time. The labeled microorganisms are usually genetically modified and have special biological characteristics, enabling them to effectively grow, reproduce, and distribute in the pipeline. Additionally, they can exhibit strong unidirectional movement characteristics under an electric field. The selection of labeled microorganisms can be determined by expert experience. Optionally, Escherichia coli, Pseudomonas, or yeast can be selected as the labeled microorganism. Usually, only one type of labeled microorganism is used for defect monitoring.

[0069] It should be noted that the pipeline drawing to be monitored details the structure, size, material, and connection relationship of the pipeline network. Through the digital pipeline drawing, the positions and attributes of each part of the pipeline network can be accurately determined. The geographic information system (GIS) is a computer system used to store, edit, query, analyze, and display geographic data. Integrating the pipeline drawing with GIS can achieve precise positioning and visualization management of the pipeline network. In GIS, information such as the layout, orientation, and surrounding environment of the pipeline network can be intuitively seen, providing an accurate spatial reference for fixed-point injection.

[0070] In GIS, each injection point is assigned a unique coordinate value. Through the coordinate value, the position of each injection point can be accurately located to ensure the accuracy of the injection. In addition to coordinate recording, the position of the injection point can also be indicated in GIS by means of annotation, icon, or color, which helps the operator quickly identify and locate the injection point, improving the injection efficiency.

[0071] The connection points, branches, and bends of pipelines are areas prone to defects in the pipeline network. Placing marker microorganisms at these locations can more effectively monitor the health of the pipelines. Welding defects or sealing problems may exist at connection points, erosion or sedimentation may occur at branches due to changes in the water flow direction, and eddies or erosion may occur at bends due to changes in the water flow velocity.

[0072] Single - time placement of marker microorganisms can simplify the operation process, reduce the number of placements and costs. Through precise concentration adjustment and dosing control, it can ensure that the number of microorganisms placed each time is accurate and meets the monitoring requirements. A flowmeter is an instrument used to measure the flow rate of fluids. During the placement of microorganisms, using a flowmeter can precisely control the dosing amount to ensure that the number of microorganisms placed each time is accurate.

[0073] Genetic modification technology can endow marker microorganisms with special biological characteristics, enabling them to grow, reproduce, and distribute more effectively in the pipeline, thereby improving the sensitivity and accuracy of monitoring. Exhibiting strong unidirectional movement characteristics under an electric field is a special behavior of marker microorganisms, enabling them to move directionally along the electric field direction in the pipeline, making them easier to detect and track.

[0074] II. Water sample collection module

[0075] It is used to collect water samples from a designated location of the pipeline to be detected after a preset time period after placing the marker microorganisms.

[0076] The water sample collection module usually includes a water sample extraction device and a water sample storage container. The water sample extraction device is used to extract water samples from the pipeline, and the water sample storage container is used to store the extracted water samples. The water sample extraction device uses a sterile container to collect water samples and automatically collects water samples at regular intervals from the inlet, outlet, connection points, branches, and bends of the pipeline. The water sample storage container uses a sterile container to store water samples and ensures the integrity of the water samples through refrigeration treatment.

[0077] It should be noted that the water sample extraction device includes an extraction mechanism and a sterile container. The extraction mechanism includes a pump, pipelines, and valves and is used to extract water samples from the pipeline. The sterile container is used to directly receive the extracted water samples and needs to undergo strict sterilization treatment before use to ensure that it is sterile inside and to ensure that the water samples are not contaminated by the outside world during the collection process.

[0078] The water sample storage container has a sterile storage function and a refrigeration treatment function. The sterile storage function is to store using a sterile container to prevent the water samples from being contaminated during storage. The refrigeration treatment function can reduce the activity of microorganisms, extend the preservation time of the water samples, help maintain the physical and chemical properties of the water samples stable, and ensure that the water samples are not affected by the external environment during storage and transportation and maintain their integrity.

[0079] It should be noted that the microbial injection and water sample collection of this application are fully automated based on pipeline drawings and geographical information; Requirements for the injection time of markers: The injection of marker microorganisms is carried out regularly according to the usage conditions and maintenance plans of the pipeline. For example, it can be injected once a month or once a quarter, and the specific frequency depends on the usage frequency of the pipeline to be monitored and the expected probability of defect occurrence. Requirements for the water sample collection time: The water sample needs to be collected within a certain period of time after the injection of marker microorganisms to ensure that the distribution of marker microorganisms in the pipeline can be detected. The specific time interval can be determined according to the growth and diffusion characteristics of the marker microorganisms.

[0080] III. Electrostatic field separation module

[0081] For each water sample collection location, the electrostatic field separation technology is used to separate particles of different sizes and properties in the water sample, and the particles include marker microorganisms, other microorganisms, and impurities.

[0082] The electrostatic separation module adopts a multi-stage structure design, which can achieve the function of multiple separations. Each stage of the electrostatic separator can adjust the parameters of the electrostatic field to accurately separate microorganisms and impurities of different types and sizes. It should be noted that marker microorganisms, other microorganisms, and impurities will show different behavioral characteristics in the electric field, such as charge distribution, polarization degree, and migration speed, which determine the response mode and separation effect of microorganisms and impurities in the electric field. Based on the differences in behavioral characteristics, by constructing a specific electric field environment, microorganisms and impurities are subjected to different forces in the electric field, thereby achieving effective separation.

[0083] The electrostatic field separation technology is based on the different charge properties of particles under the action of an external electric field, and realizes separation through the generated electrostatic force. When particles enter the electrostatic field, they will be subjected to electric field forces in different directions according to their own charge properties, thereby generating offset movement. This offset movement causes different particles to be subjected to different forces in the electric field, and thus realizes separation.

[0084] The multi-stage structure design has three characteristics: step-by-step separation, adjustable parameters, and high energy efficiency. The multi-stage structure design enables particles to be gradually separated in multiple electric fields. Each stage of the electric field can separate particles of different sizes and properties, thereby achieving a more refined separation effect; in each stage of the electric field, the electric field strength and direction can be adjusted according to actual needs to meet the separation requirements of different particles; the multi-stage structure design can improve the separation efficiency and reduce energy consumption at the same time. By optimizing the parameters of each stage of the electric field, higher separation efficiency can be achieved while reducing energy consumption.

[0085] IV. Imaging module

[0086] Take pictures and perform image segmentation on the separated particles to obtain the target area image containing the microbial morphology or the set required part.

[0087] The imaging module includes a high-resolution camera and a segmentation unit. The high-resolution camera takes pictures of the separated particles to obtain the original captured image; the segmentation unit segments the original captured image by a fully automatic segmentation method to segment out the target area image containing the microbial morphology or the set required part.

[0088] The structure of the high-resolution camera can refer to the prior art and is usually equipped with adjustable light control and autofocus functions, capable of capturing the fine structure and dynamic changes of microorganisms. Optionally, the high-resolution camera is a Canon EOS 80D camera, with the ISO speed of the camera being 800 and the exposure time being 1 / 200 s. These parameters ensure high clarity and low noise of the image, with an image resolution of 4000×6000, providing rich detail information, and an image size of 600×900, which helps to reduce the calculation amount and improve the subsequent processing rate.

[0089] Microbial morphology: The target microbial area can be the complete morphology of a marker microorganism, including its cell structure, shape, etc. For example, if it is necessary to identify the type and concentration of the marker microorganism, extracting the whole morphology can provide more comprehensive information.

[0090] Set required part: The target microbial area can also be the set required part of the marker microorganism, such as cell membrane, cell nucleus, specific organelles, etc. For example, if it is necessary to detect the specific physiological state or metabolic activity of the marker microorganism, extracting the set required part may be more helpful to achieve this goal.

[0091] The fully automatic segmentation method specifically includes:

[0092] Step a: Input the original captured image;

[0093] Step b: Convert the original captured image into a grayscale image;

[0094] Step c: Enhance the contrast through point operation and perform Gaussian filtering for denoising; the specific steps of enhancing the contrast through point operation can refer to the prior art.

[0095] Step d: Use the Canny edge detection operator to detect the boundary of each microorganism.

[0096] The steps of Canny edge detection include gradient calculation: Given the grayscale value function f(x, y) in the microbial image, calculate the gradient of the image at each pixel point (x, y) 、the magnitude of the gradient and the gradient direction at this coordinate, and the specific formula is:

[0097]

[0098]

[0099]

[0100] Among them, and are the gradient components in the x-axis direction and y-axis direction respectively, and the gradient magnitude represents the intensity of the gradient, represents the direction of the gradient.

[0101] The steps of Canny edge detection include non-maximum suppression: in the four gradient directions of 0°, 45°, 90°, and 135°, non-maximum suppression is performed on the gradient magnitude. For each pixel point (x, y), within its 8-neighborhood, two intersection points are found according to the gradient direction. The intersection points are the two endpoints where the straight line in the gradient direction intersects the neighborhood. Since the image is discrete, the intersection points may not exist at integer coordinates, so interpolation is required to estimate the gradient magnitude at the intersection points. If the gradient magnitude of the center point is the maximum among these three points, then this point is retained; otherwise, it is set to 0.

[0102] The steps of Canny edge detection include double-threshold detection: set two thresholds: a low threshold and a high threshold. If the gradient magnitude of a pixel point is less than the low threshold, it is discarded; if the gradient magnitude is greater than the high threshold, this point is retained as a strong edge point; if the gradient magnitude is between the two, then check the pixel points in its 8-neighborhood. If the gradient magnitude of any point in the neighborhood is greater than the high threshold, this point is retained as a weak edge point; otherwise, it is discarded.

[0103] Step e: Close the edge image through morphological processing and obtain the final segmentation result.

[0104] V. Microbial recognition module

[0105] Used to extract features from each target region image, identify the types of microorganisms, and obtain the concentration of indicator microorganisms at each sampling location.

[0106] The microbial recognition module includes a feature extraction unit, an identification unit, and an indicator microorganism concentration acquisition unit; the feature extraction unit is used to extract image features from each target region image; the identification unit uses the particle swarm optimization algorithm to optimize the parameters of the SVM model, and uses the optimized and trained SVM model to classify the features of the extracted target region images to identify the types of each microorganism; the indicator microorganism concentration acquisition unit, for each sampling location, counts its indicator microorganisms, and obtains the concentration of indicator microorganisms according to the number of indicator microorganisms and the volume of the collected water sample.

[0107] The image features extracted by the feature extraction unit include: geometric features, internal structure histogram features, Fourier descriptors, gray-level co-occurrence matrices, and rotation-invariant local binary patterns.

[0108] The geometric features include: the area of the microorganism, the perimeter of the boundary contour of the microorganism, the roundness of the microorganism, the major axis length and the minor axis length of the ellipse having the same normalized second-order central moment as the microorganism region.

[0109] In this embodiment, the area of the microorganism is A, and the total number of pixels in the microorganism region in the segmented image is calculated; the perimeter of the boundary contour of the microorganism is P, which is to extract the boundary of the microorganism using Canny edge detection and then calculate the pixel length of the boundary; the roundness of the microorganism is calculated using the formula 4πA / is calculated, where A is the area and P is the perimeter; the major axis length and the minor axis length are the major axis and the minor axis of the ellipse having the same normalized second-order central moment as the microorganism region.

[0110] The internal structure histogram feature is used to describe the internal spatial structure of the image and reflect the distribution of pixel intensities in the image. In this embodiment, the steps for calculating the internal structure histogram include (steps G1 - G3):

[0111] Step G1: Sampling point marking: Mark K sampling points equidistantly distributed on the boundary contour of the microorganism;

[0112] Step G2: Internal structure angle calculation; For any three sampling points, calculate the inscribed angle between them;

[0113] Step G3: Histogram statistics; Count all the obtained inscribed angles within a certain angle range to obtain the ISH (Internal Structure Histogram) feature.

[0114] The Fourier descriptor distinguishes different contours through frequency domain information and then identifies objects, and is used to reflect the contour features of the image. In this embodiment, the steps for calculating the Fourier descriptor include (steps N1 - N4):

[0115] Step N1: Sampling point marking: Mark M sampling points equidistantly distributed on the boundary contour of the microorganism;

[0116] Step N2: Euclidean distance calculation; Calculate the Euclidean distance from each sampling point to the center point and arrange them in order as { , ,..., };

[0117] Step N3: Discrete Fourier Transform: Perform a discrete Fourier transform on the distance sequence to extract frequency-domain components. The specific formula is:

[0118]

[0119] where, represents the s-th complex coefficient, s = 0, 1, ..., M - 1; j represents the unit imaginary part; represents the Euclidean distance from the t-th sampling point to the center point.

[0120] Step N4: Standardization: Perform internal standardization on the Fourier coefficients to reduce the influence of noise, translation, and rotation. The specific formula is: , is the eigenvector corresponding to the s-th complex coefficient; Since the microbial boundary contour can be represented as a real function, so, has conjugate symmetry, and finally a set of -dimensional eigenvectors ( , ..., ) are obtained.

[0121] The gray-level co-occurrence matrix is used to reflect the spatial correlation of the gray values of any two pixel points in the image. In this embodiment, the steps for calculating the gray-level co-occurrence matrix include (Steps O1 - O4):

[0122] Step O1: Gray-level compression: Compress and quantize the gray levels of the original image into 16 levels;

[0123] Step O2: Calculate the co-occurrence matrix: Calculate the co-occurrence matrix in four directions: 0°, 45°, 90°, and 135°;

[0124] Step O3: Normalization processing: Perform normalization processing on the co-occurrence matrix;

[0125] Step O4: Statistic calculation: Calculate six statistics, namely energy, entropy, contrast, inertia moment, correlation degree, and inverse difference moment, to obtain a 24-dimensional eigenvector.

[0126] The rotation-invariant local binary pattern is a robust feature operator that is invariant to any monotonic transformation of grayscale images and has grayscale invariance and rotation invariance, and is used to reflect the texture features of images. In this embodiment, the steps of the rotation-invariant local binary pattern include (Steps K1 - K5):

[0127] Step K1: Block processing: Divide the grayscale image into blocks;

[0128] Step K2: Sampling and encoding: Perform encoding sampling in the circular neighborhood of each pixel point within each block;

[0129] Step K3: Circular shift operation: Perform a circular shift operation on the sampled LBP (Local Binary Pattern) values to obtain several different values;

[0130] Step K4: Take the minimum value: Take the smallest one as the final LBP value;

[0131] Step K5: Histogram statistics: Perform histogram statistics on the LBP values of each block to obtain rotation-invariant LBP (Local Binary Pattern) features.

[0132] The recognition unit uses the particle swarm optimization algorithm to optimize the parameters of the SVM model, and uses the optimized and trained SVM model to classify the features of the extracted target region image to identify the species of each microorganism.

[0133] Specifically, the particle swarm optimization algorithm is used to optimize the parameters of the SVM model. The specific process is (Steps U1 - U5):

[0134] Step U1: According to the preset particle swarm size, randomly initialize the position and velocity of each particle; each particle represents a combination of SVM model parameters;

[0135] Step U2: Use the SVM parameter combination represented by each particle to train the SVM model, and calculate the classification accuracy rate of the SVM model using the test set in the training process. Take the classification accuracy rate of the SVM model as the fitness value of the particle;

[0136] Step U3: According to the fitness value of the particle, update the individual historical optimal position and the global optimal position of the particle swarm;

[0137] Step U4: Update the velocity and position of each particle according to the updated individual optimal position and global optimal position of the particle swarm;

[0138] Step U5: Repeat updating the individual optimal position and global optimal position of the particle until the termination condition is met. Output the global optimal position and the corresponding global optimal fitness value. The combination of SVM model parameters corresponding to the particle at the global optimal position is the optimal parameter combination of the SVM model; the optimal parameter combination of the SVM model includes the penalty factor C and the kernel parameter .

[0139] The classification accuracy rate of the SVM model is to use the SVM model parameters C of the current particle and Build an SVM model and evaluate the recognition accuracy rate of the test set.

[0140] In this embodiment, the particle swarm optimization algorithm searches for the optimal solution by randomly initializing particles and then iteratively updating the velocities and positions of the particles. Each particle moves along the directions of the individual optimum and the global optimum. During the search process in the multi-dimensional space, the formulas for each particle to update its respective velocity and position are as follows:

[0141]

[0142]

[0143] where id represents particle i in the d-dimensional space, k is the iteration number, w is the inertia weight, and are two acceleration constants, and are two random numbers uniformly distributed in (0, 1), and represent the individual optimum position and the global optimum position respectively; is the velocity of particle i in the d-dimensional space at the k-th iteration; is the position of particle i in the d-dimensional space at the k-th iteration;

[0144] In this embodiment, the objective function of the SVM model during training is:

[0145]

[0146] The constraint conditions are:

[0147]

[0148]

[0149] where and are the Lagrange multipliers of training samples i and j respectively, , N is the number of samples used for model training, and are the feature data of training samples i and j respectively; and are the class labels of training samples i and j respectively, is the RBF kernel function; C is the penalty factor;

[0150] The decision function of the SVM model is:

[0151]

[0152]

[0153] where b is a constant bias term, and x is the feature data of the target region image to be classified. is the kernel parameter.

[0154] The optimal parameter combinations C and of the SVM model obtained through the objective function, constraints, and particle swarm algorithm are used to obtain the decision function of the optimal SVM model.

[0155] The objective function in the SVM model is used to guide the training process of the model and find the optimal Lagrange multipliers; the decision function then uses these optimal Lagrange multipliers to classify new samples.

[0156] That is, as Figure 2 shown, the specific process steps for the recognition unit to perform recognition are as follows:

[0157] Step P1: Input the feature data set of the microorganism.

[0158] Step P2: Set the parameters of the PSO algorithm, including the total number of particles in the particle swarm, the maximum number of iterations, the inertia weight, the local and global learning factors, and the limit ranges of the parameters C and γ to be optimized.

[0159] Step P3: Initialize the particle swarm, randomly initialize the velocity and position of the particles, and set the iteration count to zero.

[0160] Step P4: Loop through steps P6 to P9, and increment the iteration count by 1 for each loop.

[0161] Step P5: Use the SVM parameters C and γ optimized by PSO to establish an SVM classification model for the microorganism image features, and evaluate the recognition accuracy of the test set.

[0162] Step P6: Calculate the fitness of each particle.

[0163] Step P7: Calculate the individual and global optimized fitness of the particles, and update the velocity and position of each particle.

[0164] Step P8: Determine whether the loop terminates. If the minimum error between two generations or the maximum number of iterations is reached, terminate the loop; otherwise, return to step P5.

[0165] Step P9: Output the optimal parameters C and γ.

[0166] Step P10: Establish an optimized SVM model.

[0167] Step P11: Output the microorganism classification recognition result.

[0168] After determining the number of marked microorganisms, the sample volume of the collected water sample can be determined: by intercepting a specific length of fluid in the pipeline and measuring its volume, and finally the concentration of the marked microorganisms can be calculated, that is, dividing the number of marked microorganisms by the sample volume to obtain the concentration of the marked microorganisms.

[0169] VI. Defect Identification Module

[0170] Based on the preset concentration gradient threshold of the marked microorganisms and the concentration of the marked microorganisms at different positions of the pipeline to be monitored, identify whether there are defects in the pipeline to be monitored, as well as the location and severity of the defects.

[0171] The specific process is as follows: set the first concentration threshold range, the second concentration threshold range, and the third concentration threshold range; the first concentration threshold range is between the second concentration threshold ranges; the second concentration threshold range is between the third concentration threshold ranges.

[0172] Optionally, the upper limit values of the first concentration threshold range, the second concentration threshold range, and the third concentration threshold range are 2 times, 3 times, and 4 times the marked microorganism dosing concentration; the lower limit values of the first concentration threshold range, the second concentration threshold range, and the third concentration threshold range are 0.5 times, 0.2 times, and 0.01 times the marked microorganism dosing concentration. The upper and lower threshold values set in the first concentration threshold range, the second concentration threshold range, and the third concentration threshold range are obtained based on in-depth research and analysis of the relationship between pipeline defects and changes in microorganism concentration. Although specific multiples are initially set as thresholds (such as upper threshold values of 2 times, 3 times, 4 times, and lower threshold values of 0.5 times, 0.2 times, 0.01 times), in actual applications, the thresholds are not fixed. With changes in pipeline operating conditions, succession of microorganism populations, and fluctuations in environmental factors, the thresholds should be dynamically adjusted according to actual situations to ensure the accuracy and reliability of the judgment results.

[0173] If the concentration of the marked microorganisms at the current position is within the first concentration threshold range, there is no defect at the current position of the pipeline to be monitored.

[0174] If the concentration of the marked microorganisms at the current position exceeds the upper limit value of the first concentration threshold range, there are defects such as leakage, corrosion, or other defects that cause an increase in microorganism concentration at the current position of the pipeline to be monitored; these defects damage the integrity of the pipeline, allowing external microorganisms to enter the pipeline interior and multiply in large numbers.

[0175] If the concentration of the marked microorganisms at the current position is lower than the lower limit value of the first concentration threshold range, there are defects such as poor water flow, sediment accumulation, or other defects that cause a decrease in microorganism concentration at the current position of the pipeline to be monitored. These defects need to be taken seriously, and necessary measures should be taken for investigation and treatment.

[0176] If the concentration of the marker microorganism at the current position is outside the first concentration threshold range and within the second concentration threshold range, the severity of the defect is primary; if the concentration of the marker microorganism at the current position is outside the second concentration threshold range and within the third concentration threshold range, the severity of the defect is intermediate; if the concentration of the marker microorganism at the current position is outside the third concentration threshold range, the severity of the defect is severe.

[0177] Embodiment 2

[0178] As Figure 3 shown, an intelligent pipeline defect identification method based on pipeline network microorganisms includes the following steps:

[0179] Step S1: Regularly and quantitatively release marker microorganisms at fixed points in the pipeline to be monitored.

[0180] When releasing marker microorganisms, factors such as the pipeline layout, water flow velocity, selection of release points, and release amount need to be considered. A reasonable release strategy can ensure the uniform distribution of marker microorganisms in the pipeline and improve the accuracy and reliability of detection results.

[0181] Step S2: After releasing the marker microorganisms, collect water samples from designated positions of the pipeline to be monitored after a preset time period.

[0182] When collecting water samples, professional sampling tools and methods need to be used to ensure that the collected water samples are representative and can truly reflect the distribution of microorganisms inside the pipeline. Attention should also be paid to avoiding contamination during the sampling process.

[0183] Step S3: For each water sample collection position, separate particles of different sizes and properties in the water sample through electrostatic field separation technology. The particles include marker microorganisms, other microorganisms, and impurities.

[0184] Through electrostatic field separation technology, microorganisms in the water sample can be effectively separated from other impurities, providing clear microorganism samples for subsequent image processing and feature extraction.

[0185] Step S4: Take pictures and perform image segmentation on the separated particles to obtain target area images containing microorganism morphology or specified required parts.

[0186] The photographing device should have the characteristics of high resolution and high sensitivity, and be able to clearly capture the morphology and features of microorganisms. The device should also have a stable light source and shooting environment to ensure the stable and reliable quality of the captured images.

[0187] Image segmentation is a key technology in image processing, aiming to effectively separate the target area from the background area in the image. For the captured images, edge detection methods are used for segmentation.

[0188] Step S5: Extract features from each target region image to identify the types of microorganisms, so as to obtain the concentration of marker microorganisms at each sampling location.

[0189] The extracted features refer to the information that can reflect the essential attributes of the target object in the image, such as shape, texture, and color. For microorganism images, shape features, texture features, and color features can be selected as the basis for classification and identification.

[0190] Step S6: Based on the preset concentration gradient threshold of the marker microorganisms and the concentration of the marker microorganisms at different positions of the pipeline to be monitored, identify whether there are defects in the pipeline to be monitored, as well as the location and severity of the defects.

[0191] According to the classification results of the SVM model, the concentration of the marker microorganisms can be obtained. According to the concentration, it can be determined whether there are defects in the pipeline, as well as the location and severity of the defects.

[0192] For a more specific process of the above method, reference can be made to the corresponding content disclosed in the foregoing embodiments, and details will not be repeated here.

[0193] In this specification, the various embodiments are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The same or similar parts among the various embodiments can be referred to each other. For the method disclosed in the embodiments, since it corresponds to the system disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the description of the system part.

[0194] Those skilled in the art can clearly understand that the technology in the embodiments of the present invention can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solutions in the embodiments of the present invention, in essence, or the part that contributes to the prior art can be embodied in the form of a software product. The computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments of the present invention.

[0195] The above is only the preferred embodiment of the present invention. The protection scope of the present invention is not limited to the above embodiments. All technical solutions within the idea of the present invention belong to the protection scope of the present invention. It should be noted that for those of ordinary skill in the art, several improvements and refinements made without departing from the principle of the present invention should be regarded as within the protection scope of the present invention.

Claims

1. An intelligent pipeline defect identification system based on pipeline network microorganisms, characterized in that: include: A marker microorganism delivery module is used to deliver marker microorganisms at fixed points and in fixed quantities in the pipeline to be monitored; A water sample collection module is used to collect water samples from a designated location of the pipeline to be tested after a preset period of time after the marker microorganisms are placed; An electrostatic field separation module is used to separate particles of different sizes and properties in each water sample collection location by electrostatic field separation technology, wherein the particles include marker microorganisms, other microorganisms and impurities; An imaging module is used to take pictures and segment the separated particles to obtain a target area image containing the microbial morphology or a set required part; The microorganism identification module is used to extract features from each target area image and identify the type of microorganisms to obtain the concentration of the marker microorganisms at each sampling location; A defect recognition module is used to identify whether there is a defect in the pipeline to be monitored, as well as the location and severity of the defect based on a preset concentration gradient threshold of the marker microorganism and the concentration of the marker microorganism at different locations of the pipeline to be monitored; The electrostatic field separation module adopts a multi-stage electrostatic separator structure, and each stage of the electrostatic separator can adjust the parameters of the electrostatic field to adjust the intensity and direction of the electrostatic field; The electrostatic field separation technology is used to separate particles of different sizes and properties in the water sample. Specifically, when the collected water sample passes through the electrostatic separator, the electrostatic separator applies different electric field forces to all microorganisms and impurities in the water sample, and separates them by utilizing the charge differences of different microorganisms and impurities in the water sample; The marker microorganism delivery module is used to deliver the marker microorganisms at fixed points and in fixed quantities in the pipeline to be monitored. The delivery locations of the marker microorganisms are located at the connection points, branches and turns of the pipeline to be monitored, and the delivery locations of the marker microorganisms are assigned unique coordinate values; The water sample collection module is used to collect water samples from designated positions of the pipeline to be detected after the marker is placed, after a preset time period. The designated positions for collecting water samples are the inlet, outlet, connection point, branch and turn of the pipeline to be monitored.

2. The pipeline defect intelligent identification system based on pipeline network microorganisms according to claim 1 is characterized in that: The imaging module includes a high-resolution camera and a segmentation unit; the high-resolution camera takes pictures of the separated particles to obtain an original photographed image; the segmentation unit segments the original photographed image by a fully automatic segmentation method to segment out a target area image containing the microbial morphology or a set required part; The fully automatic segmentation method specifically comprises: Input the original captured image; convert the original captured image into a grayscale image; enhance the contrast through point operation and perform Gaussian filtering to remove noise; use the Canny edge detection operator to detect the boundary of each microorganism; close the edge image through morphological processing and obtain the final segmentation result.

3. The pipeline defect intelligent identification system based on pipeline network microorganisms according to claim 1 is characterized in that: The microorganism identification module includes a feature extraction unit; The feature extraction unit is used to extract image features from each target area image; The extracted image features include geometric features, internal structure histogram features, Fourier descriptors, gray-level co-occurrence matrices, and rotation-invariant local binary patterns; The geometric features include: the area of ​​the microorganism, the perimeter of the microorganism boundary contour, the roundness of the microorganism, and the major axis length and minor axis length of an ellipse having the same normalized second-order central moment as the microorganism area; The internal structure histogram feature is used to describe the internal spatial structure of the image and reflect the distribution of pixel intensity in the image; The Fourier descriptor is used to distinguish different contours through frequency domain information, and then identify objects, and is used to reflect the contour characteristics of the image; The gray level co-occurrence matrix is ​​used to reflect the spatial correlation of the gray values ​​of any two pixels in the image; The rotation-invariant local binary pattern is a robust feature operator that is invariant to any monotonic transformation of a grayscale image, has grayscale invariance and rotation invariance, and is used to reflect the texture characteristics of an image.

4. The pipeline defect intelligent identification system based on pipeline network microorganisms according to claim 1 is characterized in that: The microorganism identification module also includes an identification unit and a marker microorganism concentration acquisition unit; the identification unit uses a particle swarm optimization algorithm to optimize the parameters of the SVM model, and uses the optimized and trained SVM model to classify the features of the extracted target area image to identify the type of each microorganism; the marker microorganism concentration acquisition unit counts the marker microorganisms at each sampling position, and obtains the concentration of the marker microorganism based on the number of marker microorganisms and the volume of the collected water sample.

5. The pipeline defect intelligent identification system based on pipeline network microorganisms according to claim 4 is characterized in that: The particle swarm optimization algorithm is used to optimize the parameters of the SVM model. The specific process is as follows: According to the preset particle swarm size, the position and velocity of each particle are randomly initialized; each particle represents a parameter combination of the SVM model; The SVM model is trained using the SVM parameter combination represented by each particle, and the classification accuracy of the SVM model is calculated using the test set of the training process. The classification accuracy of the SVM model is used as the fitness value of the particle. According to the fitness value of the particle, update the individual historical optimal position and global optimal position of the particle swarm; Update the speed and position of each particle according to the updated individual optimal position and global optimal position of the particle swarm; Repeatedly update the individual optimal position and global optimal position of the particle until the termination condition is met, output the global optimal position and the corresponding global optimal fitness value, the parameter combination of the SVM model corresponding to the particle at the global optimal position is the optimal parameter combination of the SVM model; the optimal parameter combination of the SVM model includes the penalty factor C and the kernel parameter .

6. The pipeline defect intelligent identification system based on pipeline network microorganisms according to claim 4 is characterized in that: The objective function of the SVM model during training is for: ； The constraints are: ； ； in, and are the Lagrange multipliers of training samples i and j respectively, , N is the number of samples used for model training, and are the feature data of training samples i and j respectively; and are the category labels of training samples i and j respectively, is the RBF kernel function; C is the penalty factor; The decision function of the SVM model for: ； ； Among them, b is a constant bias term, x is the feature data of the target area image to be classified, is the kernel parameter.

7. The pipeline defect intelligent identification system based on pipeline network microorganisms according to claim 1 is characterized in that: The defect recognition module is used to identify whether there is a defect in the pipeline to be monitored, as well as the location and severity of the defect based on the preset concentration gradient threshold of the marker microorganism and the concentration of the marker microorganism at different positions of the pipeline to be monitored. The specific process is as follows: A first concentration threshold range, a second concentration threshold range and a third concentration threshold range are set; the first concentration threshold range is located between the second concentration threshold range; the second concentration threshold range is located between the third concentration threshold range; If the concentration of the marker microorganism at the current position is within the first concentration threshold range, then the current position of the pipeline to be monitored has no defects; If the marker microorganism concentration at the current position exceeds the upper limit of the first concentration threshold range, then the current position of the pipeline to be monitored has leakage, corrosion or other defects that cause the microorganism concentration to increase; if the marker microorganism concentration at the current position is lower than the lower limit of the first concentration threshold range, then the current position of the pipeline to be monitored has poor water flow, sediment accumulation or other defects that cause the microorganism concentration to decrease; If the marker microorganism concentration at the current location is outside the first concentration threshold range and within the second concentration threshold range, the severity of the defect is primary; if the marker microorganism concentration at the current location is outside the second concentration threshold range and within the third concentration threshold range, the severity of the defect is intermediate; if the marker microorganism concentration at the current location is outside the third concentration threshold range, the severity of the defect is severe.

8. A pipeline defect intelligent identification method based on pipeline network microorganisms, characterized in that: include: Place marker microorganisms at fixed points and in fixed quantities in the pipeline to be monitored; After the marker microorganisms are placed, water samples are collected from designated locations of the pipeline to be tested after a preset period of time; For each water sample collection location, particles of different sizes and properties in the water sample are separated by electrostatic field separation technology, and the particles include marker microorganisms, other microorganisms and impurities; The separated particles are photographed and image segmented to obtain target area images containing microbial morphology or set required parts; Perform feature extraction on each target area image and identify the microbial species to obtain the marker microbial concentration at each sampling location; Based on the preset concentration gradient threshold of the marker microorganisms and the concentration of the marker microorganisms at different locations of the pipeline to be monitored, identify whether there are defects in the pipeline to be monitored, as well as the location and severity of the defects; The electrostatic field separation technology is used to separate particles of different sizes and properties in the water sample. Specifically, when the collected water sample passes through the electrostatic separator, the electrostatic separator applies different electric field forces to all microorganisms and impurities in the water sample, and separates them by utilizing the charge differences of different microorganisms and impurities in the water sample; A multi-stage electrostatic separator structure is adopted, and each stage of the electrostatic separator can adjust the parameters of the electrostatic field to adjust the intensity and direction of the electrostatic field; The placement of the marker microorganisms is located at the connection point, branch and bend of the pipeline to be monitored, and the placement of the marker microorganisms is assigned a unique coordinate value; The designated locations for collecting water samples are the inlets, outlets, connection points, branches and bends of the pipeline to be monitored.

Images