Pipeline deformation detection data pseudo-color imaging method and device and storage medium

Abstract

The application discloses a pipeline deformation detection data pseudo-color imaging method and device and a storage medium. The method comprises the following steps: acquiring original pipeline deformation detection data collected by multiple acquisition channels for a target pipeline; performing layer-by-layer singular value decomposition on the original pipeline deformation detection data; determining an optimal decomposition layer in each decomposition layer based on approximate pipeline deformation detection data, accumulating the detail pipeline deformation detection data of each acquisition channel of the optimal decomposition layer to obtain total detail pipeline deformation detection data; determining a feature position of the total detail pipeline deformation detection data greater than a preset threshold, replacing the approximate pipeline deformation detection data of the optimal decomposition layer at the feature position with original pipeline deformation detection data at the corresponding position to obtain denoising pipeline deformation detection data; performing baseline correction, data repair and data enhancement operations on the denoising pipeline deformation detection data, and generating a pipeline pseudo-color image based on the processed pipeline deformation detection data in each acquisition channel.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of pipeline detection, in particular to a pseudo-color imaging method and device for pipeline deformation detection data and a storage medium. BACKGROUND

[0002] Long-distance oil and gas pipelines are key infrastructure for energy supply. With the expansion of pipeline network scale and the increase of service life, the aging of pipelines significantly increases the risk of safety hazards such as leakage, explosion, geological disaster damage and structural deformation, and efficient detection and maintenance means are urgently needed to ensure the safe and stable operation of pipelines. Pipeline internal detection is the core technology for identifying internal defects and structural abnormalities. Among them, geometric deformation detection is mainly used for new pipeline acceptance, periodic passability inspection of in-service pipelines, and detection and dredging when the flow rate decreases abnormally. This detection collects pipeline profile data through intelligent detectors, which is the direct basis for evaluating the structural integrity of the pipeline. When the pipeline is locally deformed, ellipticity is a key indicator for measuring the safety state, which directly affects the pressure-bearing capacity and failure risk of the pipeline. It is crucial to achieve clear and accurate visualization of geometric deformation. Converting detection data into a two-dimensional pseudo-color image can significantly improve pipeline component recognition efficiency and reduce defect identification errors, providing intuitive evidence for fault location and repair.

[0003] Currently, when performing pipeline deformation detection, pipeline profile data is usually directly generated into a two-dimensional pseudo-color image. However, the complex working conditions of pipelines often result in pipeline profile data containing a large amount of noise, which can lead to inaccurate two-dimensional pseudo-color images and reduce the detection accuracy of pipeline deformation. SUMMARY

[0004] The present application provides a pseudo-color imaging method and device for pipeline deformation detection data and a storage medium, which can improve the generation accuracy of pipeline pseudo-color images and thus improve the detection accuracy of pipeline deformation.

[0005] According to a first aspect of the present application, a pseudo-color imaging method for pipeline deformation detection data is provided, comprising:

[0006] In response to the pseudo-color imaging signal of the target pipeline, the original pipeline deformation detection data collected by the multiple acquisition channels for the target pipeline are obtained;

[0007] For each acquisition channel, the original pipeline deformation detection data is subjected to layer-by-layer singular value decomposition to obtain the approximate pipeline deformation detection data and the detail pipeline deformation detection data of each decomposition layer;

[0008] determine an optimal decomposition layer satisfying a decomposition requirement in each decomposition layer based on the approximate pipeline deformation detection data of each decomposition layer, and accumulate the detail pipeline deformation detection data of each acquisition channel of the optimal decomposition layer to obtain total detail pipeline deformation detection data;

[0009] determine a feature position corresponding to the total detail pipeline deformation detection data greater than a preset threshold, and replace the approximate pipeline deformation detection data of the optimal decomposition layer at the feature position in each acquisition channel with original pipeline deformation detection data at the corresponding feature position to obtain noise reduction pipeline deformation detection data corresponding to the original pipeline deformation detection data in each acquisition channel;

[0010] perform at least one operation of baseline correction, data repair, and data enhancement on the noise reduction pipeline deformation detection data in each acquisition channel to obtain processed pipeline deformation detection data in each acquisition channel, and generate a pipeline pseudo-color image of the target pipeline based on the processed pipeline deformation detection data in each acquisition channel.

[0011] According to a second aspect of the present application, a pipeline deformation detection data pseudo-color imaging device is provided, comprising:

[0012] An acquisition unit is configured to acquire original pipeline deformation detection data collected by multiple acquisition channels for a target pipeline in response to a pseudo-color imaging signal of the target pipeline.

[0013] A decomposition unit is configured to perform layer-by-layer singular value decomposition on the original pipeline deformation detection data for each acquisition channel to obtain approximate pipeline deformation detection data and detail pipeline deformation detection data of each decomposition layer.

[0014] A determination unit is configured to determine an optimal decomposition layer satisfying a decomposition requirement in each decomposition layer based on the approximate pipeline deformation detection data of each decomposition layer, and accumulate the detail pipeline deformation detection data of each acquisition channel of the optimal decomposition layer to obtain total detail pipeline deformation detection data.

[0015] A replacement unit is configured to determine a feature position corresponding to the total detail pipeline deformation detection data greater than a preset threshold, and replace the approximate pipeline deformation detection data of the optimal decomposition layer at the feature position in each acquisition channel with original pipeline deformation detection data at the corresponding feature position to obtain noise reduction pipeline deformation detection data corresponding to the original pipeline deformation detection data in each acquisition channel.

[0016] The image generation unit is configured to perform at least one of baseline correction, data repair and data enhancement on the denoised pipeline deformation detection data in each acquisition channel to obtain processed pipeline deformation detection data in each acquisition channel, and generate a pipeline pseudo-color image of the target pipeline based on the processed pipeline deformation detection data in each acquisition channel.

[0017] According to a third aspect of the present application, a computer readable storage medium is provided, which stores a computer program, and the computer program is executed by a processor to implement the above pseudo-color imaging method of pipeline deformation detection data.

[0018] According to a fourth aspect of the present application, a computer device is provided, which comprises a memory, a processor and a computer program stored in the memory and executable on the processor, and the processor implements the above pseudo-color imaging method of pipeline deformation detection data when executing the program.

[0019] The pseudo-color imaging method, device and storage medium of pipeline deformation detection data provided by the present application can improve the generation accuracy of the pipeline pseudo-color image by comprehensively analyzing the pipeline deformation detection data of multiple channels to generate the pipeline pseudo-color image, can improve the data denoising effect by determining the optimal decomposition layer to avoid the problem of insufficient or excessive smoothing of the pipeline deformation detection data caused by the fixed decomposition layer number, can avoid the loss of high-frequency features in the pipeline deformation detection data by considering the detail pipeline deformation detection data of the singular value decomposition layer to ensure the comprehensiveness of the data required for generating the pseudo-color image, and can improve the generation accuracy of the pseudo-color image, can improve the quality of the pipeline deformation detection data by performing baseline correction, data repair and data enhancement on the pipeline deformation detection data, can improve the image quality of the pseudo-color image generated based on the pipeline deformation detection data, and can improve the detection accuracy of the pipeline deformation detection based on the pseudo-color image. BRIEF DESCRIPTION OF DRAWINGS

[0020] The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiments of the application and together with the description serve to explain the application. In the drawings:

[0021] Figure 1 A flow chart of a pseudo-color imaging method for pipeline deformation detection data provided by an embodiment of the application is shown;

[0022] Figure 2 A flow chart of another pseudo-color imaging method for pipeline deformation detection data provided by an embodiment of the application is shown;

[0023] Figure 3 A schematic diagram of three-dimensional mapping images before and after baseline correction provided by an embodiment of the application is shown;

[0024] Figure 4 An effect diagram of pipeline deformation detection data before and after interpolation provided by an embodiment of the application is shown;

[0025] Figure 5 A schematic diagram of a convolution calculation process for pipeline deformation detection data provided by an embodiment of the application is shown;

[0026] Figure 6 A comparison diagram of pipeline deformation detection data before and after data enhancement provided by an embodiment of the application is shown;

[0027] Figure 7 A structural schematic diagram of a pseudo-color imaging device for pipeline deformation detection data provided by an embodiment of the application is shown;

[0028] Figure 8 A structural schematic diagram of another pseudo-color imaging device for pipeline deformation detection data provided by an embodiment of the application is shown;

[0029] Figure 9 An entity structural schematic diagram of a computer device provided by an embodiment of the application is shown. DETAILED DESCRIPTION

[0030] The application will be described in detail below with reference to the drawings and embodiments. It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict.

[0031] At present, when pipeline deformation detection is performed, the pipeline contour data is directly generated as a two-dimensional pseudo-color image. Due to the complex working conditions of the pipeline, the pipeline contour data often contains a large amount of noise, so that the generated two-dimensional pseudo-color image is not accurate enough, thereby reducing the detection accuracy of the pipeline deformation.

[0032] To solve the above problems, the embodiment of the present application provides a pseudo-color imaging method for pipeline deformation detection data, as shown in Figure 1 The method comprises the following steps:

[0033] 101, in response to the pseudo-color imaging signal of the target pipeline, acquiring the original pipeline deformation detection data collected by the multi-acquisition channel for the target pipeline.

[0034] For the embodiment of the present application, the target pipeline can be a long-distance pipeline for oil or natural gas, or any pipeline in any scene; the original pipeline deformation detection data is the profile data of the pipeline, such as the pipe diameter, wall thickness, bending radius, position change of the pipeline in the radial direction, bending curvature, position and type of the weld, etc.; the multi-acquisition channel refers to simultaneously or in parallel acquiring the data of different dimensions, different positions or different types of pipeline profile by multiple independent data acquisition units or sensors. It includes multi-channel in spatial dimension and multi-channel in time dimension. The multi-channel in spatial dimension refers to deploying multiple sensors at the same cross section or axial position of the pipeline to measure different parameters (such as radial displacement, strain, temperature, etc.) respectively, for example, 8 laser displacement sensors are uniformly arranged in the circumferential direction of the pipeline to simultaneously measure the radial deformation of 8 points. The multi-channel in time dimension refers to synchronously collecting the data of multiple positions by multiplexing technology or distributed sensor network to avoid time delay error caused by serial acquisition, for example, using a fiber grating sensor array, arranging a measuring point every 1 meter along the axial direction of the pipeline, and synchronously transmitting all the data of the measuring points to the acquisition system. The embodiment of the present application generates a pipeline pseudo-color image by comprehensively considering the pipeline deformation detection data collected by multiple acquisition channels, which can ensure that the generated pseudo-color image contains all the information of the pipeline, thereby improving the generation quality of the pseudo-color image and further improving the detection accuracy of pipeline deformation detection based on the pseudo-color image.

[0035] 102, for each acquisition channel, performing layer-by-layer singular value decomposition on the original pipeline deformation detection data to obtain the approximate pipeline deformation detection data and the detail pipeline deformation detection data of each decomposition layer.

[0036] Among them, the approximate pipeline deformation detection data retains the main trend and low frequency component of the pipeline deformation detection data, and the detail pipeline deformation detection data retains the high frequency detail information of the pipeline deformation detection data.

[0037] For the embodiment of the present application, taking a certain acquisition channel as an example, a data matrix corresponding to the original pipeline deformation detection data corresponding to the acquisition channel is determined, wherein the rows in the data matrix represent different detection positions, and the columns represent different detection times or detection methods. First, singular value decomposition is performed on the data matrix at the first decomposition layer, and the formulas are as follows:

[0038] D=UΣV T

[0039] wherein D is a data matrix, U is a left singular vector matrix, Σ is a diagonal matrix, V T is a right singular vector matrix. The largest left singular vector u1 is determined in the singular vector matrix U, and the largest right singular vector v1 is determined in the singular vector matrix V The approximate pipeline deformation detection data A1 of the first decomposition layer is determined according to the following formula:

[0040]

[0041] wherein σ1 is the largest singular value in the data matrix. The difference between the original pipeline deformation detection data and the approximate pipeline deformation detection data A1 is taken as the detail pipeline deformation detection data of the first decomposition layer. Further, the detail pipeline deformation detection data of the first decomposition layer is taken as the new original pipeline deformation detection data, and the new original pipeline deformation detection data is continuously decomposed layer by layer according to the above manner, thereby obtaining the approximate pipeline deformation detection data and the detail pipeline deformation detection data of multiple decomposition layers according to the above manner.

[0042] 103. Based on the approximate pipeline deformation detection data of each decomposition layer, the optimal decomposition layer satisfying the decomposition requirement is determined in each decomposition layer, and the detail pipeline deformation detection data of each acquisition channel of the optimal decomposition layer is accumulated to obtain the total detail pipeline deformation detection data.

[0043] For the embodiment of the present application, if the decomposition layer number k is set too low, the high-frequency noise in the pipeline deformation detection data cannot be sufficiently filtered, resulting in a significant reduction in the noise reduction effect, and the noise still remains in the signal. On the contrary, if the decomposition layer number k is set too high, the pipeline deformation detection data will be over-smoothed, not only filtering out the noise, but also eliminating important high-frequency detail features in the pipeline deformation detection data, such as subtle changes at the weld, etc. Therefore, the optimal decomposition layer number needs to be determined based on the approximate pipeline deformation detection data of each decomposition layer, and the original pipeline deformation detection data is processed based on the approximate pipeline deformation detection data and the detail pipeline deformation detection data of the optimal decomposition layer, thereby improving the noise reduction effect of the pipeline deformation detection data and further improving the generation quality of the pipeline pseudo-color image.

[0044] 104. The feature position corresponding to the total detail pipeline deformation detection data greater than the preset threshold value is determined, and the approximate pipeline deformation detection data of the optimal decomposition layer at the feature position in each acquisition channel is replaced by the original pipeline deformation detection data at the corresponding feature position to obtain the noise reduction pipeline deformation detection data corresponding to the original pipeline deformation detection data in each acquisition channel.

[0045] The preset threshold is set according to actual requirements. For the embodiment of the present application, the pipeline deformation detection data features are affected by the process and service life of the weld, and the features may not be obvious. In addition, the original pipeline deformation detection data may contain a large amount of irregular noise, and the useful information of the pipeline in the detail pipeline deformation detection data is confused in the noise signal, so that the signal features are not easy to be screened. Therefore, the detail pipeline deformation detection data of each acquisition channel of the optimal decomposition layer is accumulated to obtain total detail pipeline deformation detection data, amplify the pipeline feature information in the detail signal, increase the difference between the pipeline feature signal and the noise signal, and reduce the influence of noise on signal analysis and recognition. Since some feature information of the pipeline assembly also exists in the smaller singular values, which means that the detail pipeline deformation detection data also contains related information of the pipeline assembly. After multi-layer multi-resolution singular value decomposition, in order to avoid the problem of loss of pipeline assembly feature information, that is, to reduce the phenomenon of attenuation of the pipeline feature signal with the increase of the number of decomposition layers, the pipeline feature reconstruction method based on threshold method is adopted in the embodiment of the present application. The total detail pipeline deformation detection data is analyzed, a suitable pipeline feature signal threshold range is selected, the feature position of the total detail pipeline deformation detection data greater than the preset threshold is identified, the data of the position in the approximate pipeline deformation detection data corresponding to each channel is replaced with the data of the corresponding position in the original pipeline deformation detection data, so that the reconstruction of the approximate pipeline deformation detection data is realized, and the reconstructed approximate pipeline deformation detection data is the denoising pipeline deformation detection data. The embodiment of the present application can avoid the problem of data loss by analyzing the detail pipeline deformation detection data, so as to improve the generation quality of the pseudo-color image. The embodiment of the present application analyzes the approximate pipeline deformation detection data and the detail pipeline deformation detection data, and proposes a reconstruction method for the approximate pipeline deformation detection data. The method realizes the denoising of the original pipeline deformation detection data, retains the signal amplitude of the pipeline feature, greatly relieves the signal amplitude attenuation caused by denoising, and lays a high-quality data foundation for subsequent image processing.

[0046] 105. Perform at least one operation of baseline correction, data repair, and data enhancement on the denoising pipeline deformation detection data in each acquisition channel to obtain processed pipeline deformation detection data in each acquisition channel, and generate a pipeline pseudo-color image of the target pipeline based on the processed pipeline deformation detection data in each acquisition channel.

[0047] For the embodiment of the present application, after the original pipeline deformation detection data of each acquisition channel is denoised, the denoised pipeline deformation detection data can also be subjected to baseline correction, data repair, data enhancement and the like, to obtain the processed pipeline deformation detection data in each acquisition channel. Then, the processed pipeline deformation detection data in each acquisition channel is fused to obtain fused pipeline deformation detection data. Then, the fused pipeline deformation detection data is converted into a grayscale image and then into a color image, which helps to highlight the features in the image, so that it is more observed and analyzed. Each pixel value of the grayscale image is mapped to a pseudo-color mapping table, so that the image is converted from a grayscale level to a color level. Specifically, the fused pipeline deformation detection data needs to be subjected to grayscale processing first. The grayscale value of each pixel in the grayscale image represents the brightness of the position. Unlike color images, grayscale images do not contain color information and only represent brightness changes in grayscale levels. In a grayscale image, there are 256 grayscale levels, with a change range of 0 to 255. 0 represents pure black, 255 represents pure white, and intermediate values represent different degrees of gray change. In the process of generating corresponding grayscale values using the fused pipeline deformation detection data, a grayscale value is selected as the reference for the grayscale level change, which is usually set to the average value. Let the reference value of the current grayscale be g, the deformation detection data be y i , and the average value of all data be The grayscale reference value is corresponded to the average value of the data, so that when the data is enlarged or reduced by a certain proportion k based on the average value, the corresponding grayscale value will also be enlarged or reduced by the same proportion.

[0048] The proportion coefficient The calculation formula of the grayscale corresponding to the fused pipeline deformation detection data is:

[0049]

[0050] The calculated grayscale value G i is an integer, and in actual calculation, it also needs to be taken as evidence. In addition, when k>1 or k<-1, it exceeds the grayscale range, and the grayscale value is replaced by the maximum or minimum value of the grayscale level. The above completes the generation of the preliminary grayscale image. In addition, the fused pipeline deformation detection data is obtained by using the sliding window method to intercept a part of the data in one sampling for grayscale value calculation, and the pipeline features in the intercepted data are enlarged. Colors can also be mixed by three primary colors in a certain proportion. The three primary colors are red (R), green (G), and blue (B) three color lights, so the image using this color representation method is also called an RGB image. According to the principle of three primary colors, any color light F can be represented by the addition of different components of R, G, and B:

[0051] F=r(R)+g(G)+b(B)

[0052] Adjusting the spatial coordinate values in the RGB color space can change the combination of mixed colors. Any given color exists in the RGB color space. In image processing, the Jet mapping rule is an important color mapping technique used to convert grayscale images into color images, thereby more intuitively displaying the details and features of the image. Specifically, the Jet mapping rule maps grayscale values to three color components: red, green, and blue, which together determine the color and brightness of each pixel in the image. For example, the Jet mapping table is shown in Table 1.

[0053] Table 1 Jet mapping rule

[0054]

[0055] where i is the corresponding grayscale value, and the three color components can be obtained for the grayscale value from 0 to 255. The Jet mapping rule aims to provide a color mapping method that is both intuitive and expressive. It uses the change of color to emphasize different areas and features in the grayscale image, making the originally monotonous image colorful. Through this mapping method, we can more easily identify and distinguish the signal characteristics in the pipeline data, thereby better understanding and analyzing the content contained in the data. The embodiments of the present application significantly improve the quality of the pseudo-color image through a series of processing steps. First, the baseline correction technique eliminates the differences between data in different channels, ensuring the consistency of data in each channel, making the generated image more flat; second, the data is repaired, effectively filling the missing original data caused by the small number of channels, further improving the integrity of the data; on this basis, the data is enhanced so that the pipeline components and defects are more obvious in the image, facilitating subsequent analysis and identification; finally, the Jet pseudo-color mapping technique further improves the visual effect of the image. Through these processes, the generated pseudo-color image can clearly present the characteristics of the pipeline components and defects, providing high-quality original images for subsequent pipeline feature recognition, and can be used as a comparison for other data imaging methods, making data analysis more accurate and reducing the probability of errors.

[0056] According to the pipeline deformation detection data pseudo-color imaging method provided by the application, compared with the current pipeline deformation detection data pseudo-color imaging method, the original pipeline deformation detection data of the target pipeline is collected by using multiple collection channels, the original pipeline deformation detection data under each collection channel is decomposed layer by layer, the optimal decomposition layer is determined in each decomposition layer based on the decomposition result, the approximate pipeline deformation detection data of the corresponding layer is denoised based on the detail pipeline deformation detection data of the optimal decomposition layer, the denoised pipeline deformation detection data corresponding to the original pipeline deformation detection data in each collection channel is obtained, then the baseline correction, data repair and data enhancement of the denoised pipeline deformation detection data are continued, and finally the pseudo-color image of the pipeline is generated based on the processed pipeline deformation detection data. Therefore, the pipeline pseudo-color image is generated by comprehensively analyzing the pipeline deformation detection data of multiple channels, the generation accuracy of the pipeline pseudo-color image is improved, the optimal decomposition layer is determined, the problem of insufficient or excessive smoothing of the pipeline deformation detection data caused by the fixed decomposition layer number is avoided, so that the denoising effect of the data is improved, the loss of the high-frequency characteristics in the pipeline deformation detection data is avoided by considering the detail pipeline deformation detection data of the singular value decomposition layer, so that the comprehensiveness of the data required for generating the pseudo-color image is ensured, and the generation accuracy of the pseudo-color image is improved, the baseline correction, data repair and data enhancement of the pipeline deformation detection data are performed, the quality of the pipeline deformation detection data is improved, the image quality of the pseudo-color image generated based on the pipeline deformation detection data is improved, and the detection accuracy of the pipeline deformation detection based on the pseudo-color image is improved.

[0057] Further, in order to better illustrate the above-mentioned pipeline deformation detection data pseudo-color imaging process, as a refinement and expansion of the above-mentioned embodiment, the embodiment of the application provides another pipeline deformation detection data pseudo-color imaging method, as shown in Figure 2 The method comprises the following steps:

[0058] 201, in response to the pseudo-color imaging signal of the target pipeline, the original pipeline deformation detection data collected by the multiple collection channels for the target pipeline is acquired.

[0059] Specifically, a multi-channel deformation detector is used, which can distribute multiple sensing arms or sensors along the circumference of the pipeline to simultaneously collect deformation data of multiple positions of the pipeline, thereby improving the detection efficiency and accuracy.

[0060] 202, for each collection channel, the original pipeline deformation detection data is decomposed layer by layer, and the approximate pipeline deformation detection data and the detail pipeline deformation detection data of each decomposition layer are obtained.

[0061] Specifically, multi-channel data is extracted according to time windows (e.g., 1 second) to construct a data matrix. Then, initial singular value decomposition is performed on the data matrix to extract approximate pipe deformation detection data and detailed pipe deformation detection data of the first decomposition layer. Then, singular value decomposition is performed on the approximate pipe deformation detection data to extract approximate pipe deformation detection data and detailed pipe deformation detection data of the second decomposition layer. Singular value decomposition is performed on the approximate pipe deformation detection data of each decomposition layer in the above manner to obtain approximate pipe deformation detection data and detailed pipe deformation detection data of multiple decomposition layers.

[0062] 203. Calculate the fuzzy entropy of the approximate pipe deformation detection data for each decomposition layer, and calculate the correlation coefficient between the approximate pipe deformation detection data and the original pipe deformation detection data for each decomposition layer.

[0063] In this embodiment of the invention, to determine the optimal decomposition layer, it is first necessary to determine the fuzzy entropy of the approximate pipe deformation detection data of each decomposition layer. Based on this, step 203 specifically includes: taking any decomposition layer in each decomposition layer as a target decomposition layer, constructing a pipe deformation detection feature vector from the approximate pipe deformation detection data in the target decomposition layer, and reconstructing the pipe deformation detection feature vector into multiple sub-pipe deformation detection feature vector sequences Y based on preset embedding dimensions m and m+1. i m , Y i m+1 , Among them, Y i m These are m consecutive values ​​starting from the i-th vector in the pipeline deformation detection feature vector. Y represents m consecutive values ​​starting from the j-th vector in the pipeline deformation detection feature vector. i m+1 For the (m+1) consecutive values ​​starting from the i-th vector in the pipeline deformation detection feature vector, The m+1 consecutive values ​​starting from the j-th vector in the pipeline deformation detection feature vector are used to determine the sub-pipeline deformation detection feature vector sequence Y. i m and the sub-pipe deformation detection feature vector sequence Sequence distance between And determine the feature vector sequence Y for deformation detection of the sub-pipe. i m+1 and the feature vector sequence of the sub-pipe deformation detection Sequence distance between And based on the sequence distance Determine the sequence similarity between the corresponding pairs of pipeline deformation detection feature vector sequences. wherein, n is a fuzzy energy coefficient, r is a width of a similarity tolerance boundary, exp() is a similarity calculation function; based on a sequence similarity between each pair of sub-pipeline deformation detection feature vector sequences determining a fuzzy entropy FuzzyEn(m, n, r, N) of the approximate pipeline deformation detection data of the target decomposition layer, wherein FuzzyEn(m, n, r, N) = lnφ m (n, r) - lnφ m+1 (n, r), N is a total number of vectors in the pipeline deformation detection feature vector.

[0064] Specifically, the pipeline deformation detection feature vector is reconstructed, and a vector sequence Y is composed according to the following formula i m

[0065] , Y i m+1 , Y i m+1 :

[0066] Y i m = {x(i), x(i+1),..., x(i+m-1)} - x0(i)

[0067]

[0068] Y i m+1 = {x(i), x(i+1),..., x(i+m-1), x(i+m)} - x0(i)

[0069]

[0070] wherein x(i), x(i+1),..., x(i+m-1) are respectively vector elements in the pipeline deformation detection feature vector, i is a vector identifier in the pipeline deformation detection feature vector, x(j), x(j+1),..., x(j+m-1) are respectively vector elements in the pipeline deformation detection feature vector, j is a vector identifier in the pipeline deformation detection feature vector, and x0(i) and x0(j) are baselines:

[0071]

[0072] Further, the sequence distance between sequence Y i m and sequence is determined is the maximum absolute difference of the corresponding scalar components:

[0073]

[0074] Further, the sequence distance between the sequence Y i m+1 and the sequence is determined according to the above manner. Then, based on the sequence distance, the sequence similarity between the corresponding sub-pipeline deformation detection feature vector sequences is determined, and finally the fuzzy entropy of the approximate pipeline deformation detection data of each decomposition layer is determined based on the sequence similarity. At the same time, for each decomposition layer, the correlation coefficient between the approximate pipeline deformation detection data and the original pipeline deformation detection data also needs to be determined, based on which, the method comprises: determining the covariance Cov(X f ,Y) between the approximate pipeline deformation detection data X f and the original pipeline deformation detection data Y of each decomposition layer f respectively, and determining the approximate data variance Var(X f ) of the approximate pipeline deformation detection data X f and the original data variance Var(Y) of the original pipeline deformation detection data; based on the covariance Cov(X f ,Y), the approximate data variance Var(X f ) and the original data variance Var(Y) corresponding to each decomposition layer f, the correlation coefficient r(X f ,Y) between the approximate pipeline deformation detection data X f and the original pipeline deformation detection data Y of each decomposition layer f is determined, wherein,

[0075] Specifically, the covariance between the approximate pipeline deformation detection data and the original pipeline deformation detection data of this layer, and the variance of the approximate pipeline deformation detection data and the variance of the original pipeline deformation detection data of this layer are determined, and finally the correlation coefficient between the approximate pipeline deformation detection data and the original pipeline deformation detection data of each layer is determined based on the above covariance and variance.

[0076] 204, respectively, the fuzzy entropy and the correlation coefficient are normalized to obtain the normalized fuzzy entropy and the normalized correlation coefficient.

[0077] Specifically, the normalized fuzzy entropy FuzzyEn new

[0078]

[0079] wherein, FuFuzzyEn is the fuzzy entropy, FuzzyEn min is the minimum fuzzy entropy in the fuzzy entropy corresponding to each decomposition layer, FuzzyEn max is the maximum fuzzy entropy in the fuzzy entropy corresponding to each decomposition layer. Further, the normalized correlation coefficient r new is determined according to the following formula:

[0080]

[0081] wherein, r is the correlation coefficient, r min is the minimum correlation coefficient in the correlation coefficient corresponding to each decomposition layer, r max is the maximum correlation coefficient in the correlation coefficient corresponding to each decomposition layer.

[0082] 205. Based on the normalized fuzzy entropy and the normalized correlation coefficient under each decomposition layer, a decomposition effect evaluation value of each decomposition layer is determined, and based on the decomposition effect evaluation value, an optimal decomposition layer is determined in each decomposition layer.

[0083] Specifically, for each decomposition layer, the decomposition effect evaluation value DRE is calculated according to the following formula:

[0084] DRE = r new -FuzzyEn new

[0085] Finally, the decomposition layer corresponding to the maximum decomposition effect evaluation value is taken as the optimal decomposition layer. The optimal decomposition layer is determined by the present embodiment through comprehensive analysis of the fuzzy entropy and the correlation coefficient, which can ensure that the data after singular value decomposition is sufficiently smooth, and also ensure that the approximate signal retains the data characteristics of the original signal, thereby improving the generation accuracy of the pipeline pseudo-color image.

[0086] 206. The detail pipeline deformation detection data of each acquisition channel of the optimal decomposition layer is accumulated to obtain total detail pipeline deformation detection data.

[0087] 207. The feature position corresponding to the total detail pipeline deformation detection data greater than the preset threshold value is determined, and the approximate pipeline deformation detection data of the optimal decomposition layer at the feature position in each acquisition channel is replaced by the original pipeline deformation detection data at the corresponding feature position to obtain the noise reduction pipeline deformation detection data corresponding to the original pipeline deformation detection data in each acquisition channel.

[0088] The preset threshold is set according to actual requirements. Specifically, the absolute value of the detail data of each channel is accumulated to obtain total detail pipeline deformation detection data, the position (feature position) of the data greater than the preset threshold in the total detail pipeline deformation detection data is determined, the approximate pipeline deformation detection data at the position is replaced by the original pipeline deformation detection data at the corresponding position, and the approximate pipeline deformation detection data after replacement is added to the detail pipeline deformation detection data to obtain denoised pipeline deformation detection data. The embodiment of the present application not only considers the approximate pipeline deformation detection data, but also considers the detail pipeline deformation detection data. Since the detail pipeline deformation detection data also contains relevant information of the pipeline assembly, by considering the detail pipeline deformation detection data, information loss can be avoided, the data comprehensiveness of the pipeline pseudo-color image is ensured, and therefore the construction precision of the pipeline pseudo-color image can be improved.

[0089] 208. At least one of baseline correction, data repair and data enhancement is performed on the denoised pipeline deformation detection data in each acquisition channel to obtain processed pipeline deformation detection data in each acquisition channel, and a pipeline pseudo-color image of the target pipeline is generated based on the processed pipeline deformation detection data in each acquisition channel.

[0090] For the embodiment of the present application, in order to further improve the construction precision of the pipeline pseudo-color image, baseline correction needs to be performed on the denoised pipeline deformation detection data in each acquisition channel. Based on this, step 208 specifically includes: determining the denoised pipeline deformation detection data D j (i) of each sampling point i in the denoised pipeline deformation detection data corresponding to each acquisition channel j, and determining the data median mediam(D j ) of the denoised pipeline deformation detection data corresponding to each acquisition channel j; based on the denoised pipeline deformation detection data D j (i) and the data median mediam(D j ) of each acquisition channel j, determining the baseline-corrected denoised pipeline deformation detection data G j (i) corresponding to the denoised pipeline deformation detection data of each acquisition channel j, wherein, n is the total number of acquisition channels.

[0091] Specifically, when a pipe deformation detector operates inside a pipe, due to sensor differences and the complex working environment, the detector center may not necessarily be aligned with the pipe's central axis. This leads to differences in the baseline values ​​of deformation detection data from different channels. While these baseline differences are not significant when observing the data using graphs, they appear as striped color differences in the generated pseudo-color image. This can negatively impact the assessment of the pipe's internal environment using pseudo-color images. Therefore, before further data processing, baseline correction between different channels must be performed on a single-channel basis. By subtracting the median value from each channel's data, the baseline differences between channels are eliminated, preventing striped color bands from appearing in the pseudo-color image. Figure 3 As shown in the figures (where Figure (a) is the 3D mapping image before baseline correction and Figure (b) is the 3D mapping image after baseline correction), the 3D mapping images of the data before and after baseline correction are presented. As can be seen from the figures, baseline correction of multi-channel deformation detection data can eliminate the differences in baseline values ​​between different channels while preserving the original variation trend of the channel data, making the features of the pipe components contained in the data more clearly visible.

[0092] Furthermore, to ensure the data quality required for generating the pseudo-color image, data repair can be performed on the pipeline deformation detection data. Based on this, the method includes: for each acquisition channel, determining the data acquisition point corresponding to the noise-reduced pipeline deformation detection data; dividing the data acquisition point into multiple sub-acquisition intervals; and constructing a cubic spline interpolation function with parameters to be solved for each sub-acquisition interval; determining the deformation characteristic information of the target pipeline; determining boundary constraints based on the deformation characteristic information; solving for the parameters to be solved in the cubic spline interpolation function based on the boundary constraints and the noise-reduced pipeline deformation detection data of each data acquisition point; determining the actual cubic spline interpolation function corresponding to the noise-reduced pipeline deformation detection data based on the solution results; and interpolating the noise-reduced pipeline deformation detection data based on the actual cubic spline interpolation function to obtain the repaired noise-reduced pipeline deformation detection data.

[0093] This includes deformation characteristics such as continuity and curvature constraints. Specifically, suppose there are n nodes (data acquisition points) in a given data interval [a, b], satisfying a = x0. <x1<x2…<x n = b, where x0,...,x n These represent the noise reduction pipeline deformation detection data for each acquisition point within the data interval, and the cubic spline interpolation function value for each node is f(x). i )=f i(i = 0, 1, 2,..., n), where i is the identification of the collection point. In order to realize the construction of f(x), f(x) is often expressed as the sum of a column of cubic polynomials:

[0094]

[0095] where S i (x) is the piecewise function corresponding to the i-th character collection interval, S i (x) is a cubic polynomial defined on the interval, satisfying the following three conditions:

[0096] (1) On each sub-collection interval [x i-1 , x i ](i = 1, 2,..., n), the degree of the polynomial S i (x) is not greater than three.

[0097] (2) S(x), S'(x), S''(x) are continuous and smooth on the entire interval [a, b], that is:

[0098] n is the total number of sampling points

[0099]

[0100] (3) S(x i ) = f i (i = 0, 1, 2,..., n).

[0101] If it is satisfied, S(x) is called the cubic spline interpolation function of the function f(x) with respect to n nodes.

[0102] In addition, in order to satisfy the cubic spline interpolation for the natural spline on the interval [x0, x n ], it also needs to satisfy:

[0103]

[0104] The above cubic polynomial S i (x) is usually expressed as:

[0105] S i (x i ) = a i +b i (x-x i )+c i (x-x i ) 2 +d i (x-x i ) 3

[0106] where the four parameters a i , b i , c i and d i to be solved are contained, and need to be calculated by the following steps.

[0107] (4) In each i, it is required to satisfy S i (x i ) = y i and S i (x i+1 ) = y i+1 , and the calculation thereof can obtain:

[0108] a i = y i

[0109]

[0110] where h i = x i+1 - x i represents the distance between the interval [x i , x i+1 ].

[0111] (5) In each i, it is required to satisfy the two conditions, and substitute into the first derivative , and the calculation thereof can obtain:

[0112]

[0113] Subtracting the above two equations can obtain:

[0114]

[0115] Because i = 0, 1,..., n-2, there are n-1 equations:

[0116]

[0117] (6) In each i = 0, 1,..., n-1, it is required to satisfy substitute into the second derivative , and the calculation thereof can obtain:

[0118] 2c i + 6d i h i = 2c i+1

[0119] The above formulas are combined, and the calculation thereof can obtain:

[0120]

[0121] Because i = 0, 1,..., n-2, there are n-1 equations:

[0122]

[0123] Solving the above steps one by one, the values of a i , b i , c i and d i are obtained, and then the actual cubic spline interpolation function is obtained. Finally, the actual cubic spline interpolation function is used to interpolate the noise reduction pipeline deformation detection data to obtain the repaired noise reduction pipeline deformation detection data. The effect before and after the deformation detection data interpolation is shown in Figure 4 (as shown in Fig. (a), the effect before interpolation, Fig. (b) is the effect after interpolation). Because there is a certain difference between the adjacent channels of the deformation detection data, the color is not coherent between different channels in the three-dimensional mapping grid graph, showing a ladder-shaped distribution. If a two-dimensional color image is directly generated, similar problems will occur. After cubic spline interpolation, the gap between the adjacent channel data is repaired, making the data coherent and smooth, as shown in Figure 4 , the ladder-shaped color block disappears, realizing better transition between colors, and the overall image becomes smooth and delicate.

[0124] Further, in order to further increase the construction accuracy of the pseudo-color image, the noise reduction pipeline deformation detection data can also be data enhanced. Based on this, the method comprises: for each acquisition channel, determining the data matrix corresponding to the noise reduction pipeline deformation detection data, and determining the horizontal direction convolution kernel f x and the vertical direction convolution kernel f y of the data matrix; based on the horizontal direction convolution kernel f x and the vertical direction convolution kernel f y , determine the gradient amplitude matrix g(x, y) of the data matrix, wherein, fuse the gradient amplitude matrix g(x, y) with the data matrix to obtain a fusion matrix, and perform contrast stretching on the fusion matrix, and determine the noise reduction pipeline deformation detection data after data enhancement based on the contrast stretching result.

[0125] Specifically, image enhancement is mainly realized by sharpening, which enhances the edge to strengthen the image clarity. Because the number of pipeline deformation detection data is large, in order to reduce the processing time, a first-order differential sharpening method is adopted. The first-order sharpening method mainly analyzes the edge features and texture structures contained therein of the image, and performs convolution operation on the image to highlight the edge features in the image. For example, the horizontal and vertical direction convolution kernels are as follows:

[0126]

[0127] wherein f x is a horizontal direction convolution kernel, f y is a vertical direction convolution kernel, and g(x, y) is a final gradient amplitude matrix. The convolution calculation schematic diagram is shown in Figure 5 . Then the gradient amplitude matrix is weighted and fused with the data matrix corresponding to the denoising pipeline deformation detection data, and the fused matrix is nonlinearly stretched (such as gamma correction), and finally the stretched data is determined as the denoising pipeline deformation detection data after data enhancement. The three-dimensional color drawing of the deformation detection data after the enhancement processing is shown in Figure 6 . Figure (a) is the data image before sharpening processing, and figure (b) is the data image after sharpening processing. This section of data contains multiple welds, and the raised part of the image corresponds to the pipeline weld. As shown in the figure, after the enhancement processing, the pipeline weld becomes clearer, and the difference with the non-weld position data is more obvious, so it can be known that the pipeline assembly in the image is easier to observe after the enhancement processing. Finally, based on the baseline correction, data repair, and data enhancement, the pipeline deformation detection data, the pipeline pseudo-color image of the target pipeline is generated.

[0128] According to the pipeline deformation detection data pseudo-color imaging method provided by the present application, compared with the current pipeline deformation detection method of directly generating a pseudo-color image according to pipeline contour data, the present application collects original pipeline deformation detection data of a target pipeline through multiple acquisition channels, performs layer-by-layer singular value decomposition on the original pipeline deformation detection data under each acquisition channel, determines an optimal decomposition layer in each decomposition layer based on the decomposition result, performs noise reduction on the approximate pipeline deformation detection data of the corresponding layer based on the detail pipeline deformation detection data of the optimal decomposition layer, obtains the noise reduction pipeline deformation detection data corresponding to the original pipeline deformation detection data in each acquisition channel, and then continues to perform baseline correction, data repair, data enhancement and other processing on the noise reduction pipeline deformation detection data, and finally generates a pseudo-color image of the pipeline according to the processed pipeline deformation detection data. Thus, the pipeline pseudo-color image is generated by comprehensive analysis of the pipeline deformation detection data of multiple channels, which can improve the generation accuracy of the pipeline pseudo-color image. The optimal decomposition layer is determined to avoid the problem of insufficient or excessive smoothing of the pipeline deformation detection data due to fixed decomposition layer number, thereby improving the noise reduction effect of the data. The noise reduction processing is performed by considering the detail pipeline deformation detection data of the singular value decomposition layer, which can avoid the loss of high-frequency features in the pipeline deformation detection data, thereby ensuring the comprehensiveness of the data required for generating the pseudo-color image, and further improving the generation accuracy of the pseudo-color image. The baseline correction, data repair, data enhancement and other operations are performed on the pipeline deformation detection data, which can improve the quality of the pipeline deformation detection data, thereby improving the image quality of the pseudo-color image generated based on the pipeline deformation detection data, and further improving the detection accuracy of the pipeline deformation detection based on the pseudo-color image.

[0129] Further, as a specific implementation of Figure 1 , the present application provides a pipeline deformation detection data pseudo-color imaging device, as shown in Figure 7 , the device comprises an acquisition unit 31, a decomposition unit 32, a determination unit 33, a replacement unit 34, and an image generation unit 35.

[0130] The acquisition unit 31 can be used to acquire original pipeline deformation detection data collected by multiple acquisition channels for a target pipeline in response to a pseudo-color imaging signal of the target pipeline.

[0131] The decomposition unit 32 can be used to perform layer-by-layer singular value decomposition on the original pipeline deformation detection data for each acquisition channel to obtain approximate pipeline deformation detection data and detail pipeline deformation detection data of each decomposition layer.

[0132] The determination unit 33 can be configured to determine an optimal decomposition layer meeting a decomposition requirement in each decomposition layer based on the approximate pipeline deformation detection data of each decomposition layer, and accumulate the detail pipeline deformation detection data of each acquisition channel of the optimal decomposition layer to obtain total detail pipeline deformation detection data.

[0133] The replacement unit 34 can be configured to determine a feature position corresponding to the total detail pipeline deformation detection data greater than a preset threshold, and replace the approximate pipeline deformation detection data of the optimal decomposition layer at the feature position in each acquisition channel with original pipeline deformation detection data at the corresponding feature position to obtain noise reduction pipeline deformation detection data corresponding to the original pipeline deformation detection data in each acquisition channel.

[0134] The image generation unit 35 can be configured to perform at least one of baseline correction, data repair, and data enhancement on the noise reduction pipeline deformation detection data in each acquisition channel to obtain processed pipeline deformation detection data in each acquisition channel, and generate a pipeline pseudo-color image of the target pipeline based on the processed pipeline deformation detection data in each acquisition channel.

[0135] In a specific application scenario, in order to determine the optimal decomposition layer meeting the decomposition requirement in each decomposition layer, the determination unit 33 can be configured to calculate a fuzzy entropy of the approximate pipeline deformation detection data of each decomposition layer and a correlation coefficient between the approximate pipeline deformation detection data of each decomposition layer and the original pipeline deformation detection data. Figure 8 The determination unit 33 can be configured to determine an optimal decomposition layer meeting a decomposition requirement in each decomposition layer based on the approximate pipeline deformation detection data of each decomposition layer, and accumulate the detail pipeline deformation detection data of each acquisition channel of the optimal decomposition layer to obtain total detail pipeline deformation detection data.

[0136] The calculation module 331 can be configured to calculate a fuzzy entropy of the approximate pipeline deformation detection data of each decomposition layer and a correlation coefficient between the approximate pipeline deformation detection data of each decomposition layer and the original pipeline deformation detection data.

[0137] The normalization module 332 can be configured to perform normalization processing on the fuzzy entropy and the correlation coefficient respectively, to obtain normalized fuzzy entropy and normalized correlation coefficient.

[0138] The determination module 333 can be configured to determine a decomposition effect evaluation value of each decomposition layer based on the normalized fuzzy entropy and the normalized correlation coefficient of each decomposition layer, and determine the optimal decomposition layer in each decomposition layer based on the decomposition effect evaluation value.

[0139] In specific application scenarios, in order to calculate the fuzzy entropy of the approximate pipe deformation detection data of each decomposition layer, the calculation module 331 can specifically be used to take any decomposition layer in each decomposition layer as a target decomposition layer, and construct a pipe deformation detection feature vector from the approximate pipe deformation detection data in the target decomposition layer. Based on preset embedding dimensions m and m+1, the pipe deformation detection feature vector is reconstructed into a sequence of multiple sub-pipe deformation detection feature vectors Y. i m , Y i m+1 , Among them, Y i m These are m consecutive values ​​starting from the i-th vector in the pipeline deformation detection feature vector. Y represents m consecutive values ​​starting from the j-th vector in the pipeline deformation detection feature vector. i m+1 For the (m+1) consecutive values ​​starting from the i-th vector in the pipeline deformation detection feature vector, The m+1 consecutive values ​​starting from the j-th vector in the pipeline deformation detection feature vector are used to determine the sub-pipeline deformation detection feature vector sequence Y. i m and the sub-pipe deformation detection feature vector sequence Sequence distance between And determine the feature vector sequence Y for deformation detection of the sub-pipe. i m +1 and the feature vector sequence of the sub-pipe deformation detection Sequence distance between And based on the sequence distance Determine the sequence similarity between the corresponding pairs of pipeline deformation detection feature vector sequences. in, n is the fuzzy energy coefficient, r is the width of the similarity tolerance boundary, and exp() is the similarity calculation function; based on the sequence similarity between the feature vector sequences of each pair of sub-pipe deformation detection. Determine the fuzzy entropy FuzzyEn(m,n,r,N) of the approximate pipe deformation detection data of the target decomposition layer, where FuzzyEn(m,n,r,N)=lnφ m (n,r)-lnφ m+1 (n,r), N is the total number of vectors in the pipeline deformation detection feature vector.

[0140] In a specific application scenario, in order to calculate the correlation coefficient between the approximate pipeline deformation detection data of each decomposition layer and the original pipeline deformation detection data, the calculation module 331 can be specifically used to determine the covariance Cov(X f ,Y) between the approximate pipeline deformation detection data X f of each decomposition layer f and the original pipeline deformation detection data Y, and determine the approximate data variance Var(X f ) of the approximate pipeline deformation detection data X f of each decomposition layer f and the original data variance Var(Y) of the original pipeline deformation detection data; based on the covariance Cov(X f ,Y), the approximate data variance Var(X f ), and the original data variance Var(Y) corresponding to each decomposition layer f, determine the correlation coefficient r(X f ,Y) between the approximate pipeline deformation detection data X f of each decomposition layer f and the original pipeline deformation detection data Y, wherein,

[0141] In a specific application scenario, in order to baseline correct, data repair, and data enhance the denoising pipeline deformation detection data in each acquisition channel, the image generation unit 35 includes a data correction module 351, a data repair module 352, and a data enhancement module 353.

[0142] The data correction module 351 can be used to determine the denoising pipeline deformation detection data D j (i) of each sampling point i in the denoising pipeline deformation detection data corresponding to each acquisition channel j, and determine the data median mediam(D j ) of the denoising pipeline deformation detection data corresponding to each acquisition channel j; based on the denoising pipeline deformation detection data D j (i) and the data median mediam(D j ) corresponding to each acquisition channel j, determine the baseline corrected denoising pipeline deformation detection data G j (i) corresponding to the denoising pipeline deformation detection data of each acquisition channel j, wherein, n is the total number of acquisition channels.

[0143] The data repair module 352 can be configured to determine, for each acquisition channel, a data acquisition point corresponding to the noise reduction pipeline deformation detection data, divide the data acquisition point into a plurality of sub-acquisition intervals, and construct a cubic spline interpolation function with to-be-solved parameters corresponding to each sub-acquisition interval respectively; determine deformation characteristic information of the target pipeline, and determine a boundary constraint condition based on the deformation characteristic information; based on the boundary constraint condition, the noise reduction pipeline deformation detection data of each data acquisition point, solve the to-be-solved parameters in the cubic spline interpolation function, and determine an actual cubic spline interpolation function corresponding to the noise reduction pipeline deformation detection data based on a solving result; perform interpolation processing on the noise reduction pipeline deformation detection data based on the actual cubic spline interpolation function, and obtain repaired noise reduction pipeline deformation detection data.

[0144] The data enhancement module 353 can be configured to determine, for each acquisition channel, a data matrix corresponding to the noise reduction pipeline deformation detection data, and determine a horizontal direction convolution kernel f x and a vertical direction convolution kernel f y based on the noise reduction pipeline deformation detection data; determine a gradient amplitude matrix g(x, y) of the data matrix based on the horizontal direction convolution kernel f x and the vertical direction convolution kernel f y . Fuse the gradient amplitude matrix g(x, y) and the data matrix to obtain a fusion matrix, perform contrast stretching on the fusion matrix, and determine the noise reduction pipeline deformation detection data after data enhancement based on a contrast stretching result.

[0145] It should be noted that other corresponding descriptions of each functional module involved in the pipeline deformation detection data pseudo-color imaging device provided by the embodiments of the present application can be referred to the corresponding descriptions of the method shown in Figure 1 , which will not be described here.

[0146] Based on the above as Figure 1The method shown in the invention, correspondingly, also provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs the following steps: In response to a pseudo-color imaging signal of a target pipe, acquiring raw pipe deformation detection data acquired by multiple acquisition channels for the target pipe; for each acquisition channel, performing layer-by-layer singular value decomposition on the raw pipe deformation detection data to obtain approximate pipe deformation detection data and detailed pipe deformation detection data for each decomposition layer; based on the approximate pipe deformation detection data of each decomposition layer, determining the optimal decomposition layer that meets the decomposition requirements in each decomposition layer, and accumulating the detailed pipe deformation detection data of each acquisition channel of the optimal decomposition layer to obtain... Total detailed pipe deformation detection data; determine the feature positions corresponding to the total detailed pipe deformation detection data that exceed a preset threshold, and replace the approximate pipe deformation detection data of the optimal decomposition layer at the feature positions in each acquisition channel with the original pipe deformation detection data at the corresponding feature positions to obtain the noise-reduced pipe deformation detection data corresponding to the original pipe deformation detection data in each acquisition channel; perform at least one operation of baseline correction, data repair, and data enhancement on the noise-reduced pipe deformation detection data in each acquisition channel to obtain the processed pipe deformation detection data in each acquisition channel, and generate a pseudo-color image of the target pipe based on the processed pipe deformation detection data in each acquisition channel.

[0147] Based on the above, Figure 1 The method shown and as Figure 7 The embodiment of the device shown in the invention also provides a physical structure diagram of a computer device, such as... Figure 9As shown, the computer device comprises a processor 41, a memory 42, and a computer program stored on the memory 42 and executable on the processor, wherein the memory 42 and the processor 41 are both arranged on a bus 43, and the processor 41 implements the following steps when executing the program: in response to a pseudo-color imaging signal of a target pipeline, acquiring original pipeline deformation detection data collected by multiple acquisition channels for the target pipeline; for each acquisition channel, performing layer-by-layer singular value decomposition on the original pipeline deformation detection data to obtain approximate pipeline deformation detection data and detail pipeline deformation detection data of each decomposition layer; based on the approximate pipeline deformation detection data of each decomposition layer, determining an optimal decomposition layer that meets decomposition requirements in each decomposition layer, and accumulating the detail pipeline deformation detection data of each acquisition channel of the optimal decomposition layer to obtain total detail pipeline deformation detection data; determining a feature position corresponding to the total detail pipeline deformation detection data greater than a preset threshold, and replacing the approximate pipeline deformation detection data of the optimal decomposition layer at the feature position in each acquisition channel with original pipeline deformation detection data at the corresponding feature position to obtain noise reduction pipeline deformation detection data corresponding to the original pipeline deformation detection data in each acquisition channel; performing at least one operation of baseline correction, data repair, and data enhancement on the noise reduction pipeline deformation detection data in each acquisition channel to obtain processed pipeline deformation detection data in each acquisition channel, and generating a pipeline pseudo-color image of the target pipeline based on the processed pipeline deformation detection data in each acquisition channel.

[0148] Through the technical solution of this invention, the original pipeline deformation detection data of the target pipeline is acquired using multiple acquisition channels, and singular value decomposition is performed on the original pipeline deformation detection data under each acquisition channel. Based on the decomposition results, the optimal decomposition layer is determined in each decomposition layer. Based on the detailed pipeline deformation detection data of the optimal decomposition layer, the approximate pipeline deformation detection data of the corresponding layer is denoised to obtain the denoised pipeline deformation detection data corresponding to the original pipeline deformation detection data in each acquisition channel. Then, the denoised pipeline deformation detection data is further processed by baseline correction, data repair, data enhancement, etc. Finally, a pseudo-color image of the pipeline is generated based on the processed pipeline deformation detection data. Therefore, by comprehensively analyzing pipeline deformation detection data from multiple channels to generate pseudo-color pipeline images, the generation accuracy of pseudo-color pipeline images can be improved. Determining the optimal decomposition layer avoids the problems of insufficient noise reduction or excessive smoothing of pipeline deformation detection data due to a fixed number of decomposition layers, thus improving the noise reduction effect. Considering the detailed pipeline deformation detection data from the singular value decomposition layer for noise reduction avoids the loss of high-frequency features in the pipeline deformation detection data, ensuring the comprehensiveness of the data required for generating pseudo-color images, thereby improving the generation accuracy of pseudo-color images. By performing baseline correction, data repair, and data enhancement on the pipeline deformation detection data, the quality of the pipeline deformation detection data can be improved, thereby improving the image quality of the pseudo-color images generated based on the pipeline deformation detection data, and ultimately improving the detection accuracy of pipeline deformation detection based on pseudo-color images.

[0149] It is obvious to those skilled in the art that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby storing them in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those presented herein, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular combination of hardware and software.

[0150] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method of pseudo-color imaging of pipeline deformation detection data, comprising: The method comprises the following steps: In response to the pseudo-color imaging signal of the target pipeline, raw pipeline deformation detection data collected by multiple acquisition channels for the target pipeline are acquired; For each acquisition channel, the raw pipeline deformation detection data are subjected to layer-by-layer singular value decomposition to obtain approximate pipeline deformation detection data and detail pipeline deformation detection data of each decomposition layer; Based on the approximate pipeline deformation detection data of each decomposition layer, an optimal decomposition layer meeting the decomposition requirement is determined in each decomposition layer, and the detail pipeline deformation detection data of each acquisition channel of the optimal decomposition layer are accumulated to obtain total detail pipeline deformation detection data; A feature position corresponding to the total detail pipeline deformation detection data greater than a preset threshold is determined, and the approximate pipeline deformation detection data of the optimal decomposition layer at the feature position in each acquisition channel is replaced by the raw pipeline deformation detection data at the corresponding feature position to obtain noise reduction pipeline deformation detection data corresponding to the raw pipeline deformation detection data in each acquisition channel; At least one of baseline correction, data repair, and data enhancement is performed on the noise reduction pipeline deformation detection data in each acquisition channel to obtain processed pipeline deformation detection data in each acquisition channel, and a pipeline pseudo-color image of the target pipeline is generated based on the processed pipeline deformation detection data in each acquisition channel.

2. The method of claim 1, wherein, Based on the approximate pipeline deformation detection data of each decomposition layer, an optimal decomposition layer meeting the decomposition requirement is determined in each decomposition layer, comprising: The fuzzy entropy of the approximate pipeline deformation detection data of each decomposition layer is calculated respectively, and the correlation coefficient between the approximate pipeline deformation detection data of each decomposition layer and the raw pipeline deformation detection data is calculated; The fuzzy entropy and the correlation coefficient are normalized respectively to obtain normalized fuzzy entropy and normalized correlation coefficient; Based on the normalized fuzzy entropy and the normalized correlation coefficient of each decomposition layer, a decomposition effect evaluation value of each decomposition layer is determined, and the optimal decomposition layer is determined in each decomposition layer based on the decomposition effect evaluation value.

3. The method of claim 2, wherein, The fuzzy entropy of the approximate pipeline deformation detection data of each decomposition layer is calculated respectively, comprising: any one of the decomposition layers as a target decomposition layer, constructing a pipeline deformation detection feature vector from approximate pipeline deformation detection data in the target decomposition layer, reconstructing the pipeline deformation detection feature vector into a plurality of sub-pipeline deformation detection feature vector sequences based on a preset embedding dimension and , , , , wherein, is a continuous value starting from the th vector in the pipeline deformation detection feature vector, is a continuous value starting from the th vector in the pipeline deformation detection feature vector, is a continuous value starting from the th vector in the pipeline deformation detection feature vector, is a continuous value starting from the th vector in the pipeline deformation detection feature vector, is a continuous value starting from the th vector in the pipeline deformation detection feature vector, is a continuous value starting from the th vector in the pipeline deformation detection feature vector; determining a sequence distance between the sub-pipeline deformation detection feature vector sequence and the sub-pipeline deformation detection feature vector sequence determining a sequence distance between the sub-pipeline deformation detection feature vector sequence and the sub-pipeline deformation detection feature vector sequence determining a sequence distance between the sub-pipeline deformation detection feature vector sequence and the sub-pipeline deformation detection feature vector sequence determining a sequence similarity between the respective pairs of sub-pipeline deformation detection feature vector sequences based on the sequence distances , , , wherein, , , a fuzzy energy coefficient, a width of a similarity tolerance boundary, a similarity calculation function; determining a sequence similarity between each pair of sub-pipeline deformation detection feature vector sequences , , determining a fuzzy entropy of the approximate pipeline deformation detection data of the target decomposition layer wherein, , , , is the total number of vectors in the pipeline deformation detection feature vectors.

4. The method of claim 2, wherein, The correlation coefficient between the approximate pipeline deformation detection data of each decomposition layer and the raw pipeline deformation detection data is calculated, comprising: Determine each decomposition layer separately The approximate pipeline deformation detection data and the original pipeline deformation detection data covariance between and determine each decomposition layer The approximate pipeline deformation detection data Approximate data variance The original data variance of the original pipeline deformation detection data ; Based on each decomposition layer The corresponding covariance The approximate data variance The original data variance Determine each decomposition layer Approximate pipe deformation detection data and the original pipeline deformation detection data Correlation coefficient between ,in, .

5. The method of claim 1, wherein, The method for performing baseline correction on the noise reduction pipeline deformation detection data in each acquisition channel comprises: determining each acquisition channel corresponding to each sampling point in the noise reduction pipeline deformation detection data noise reduction pipeline deformation detection data and determining each acquisition channel corresponding to the median value of the noise reduction pipeline deformation detection data ; based on each acquisition channel corresponding noise reduction pipeline deformation detection data and data median determining, for each acquisition channel the baseline-corrected noise reduction pipeline deformation detection data corresponding to the noise reduction pipeline deformation detection data wherein, , is the total number of acquisition channels.

6. The method of claim 1, wherein, The method for performing data repair on the noise reduction pipeline deformation detection data in each acquisition channel comprises: For each acquisition channel, a data acquisition point corresponding to the noise reduction pipeline deformation detection data is determined, the data acquisition point is divided into a plurality of sub-acquisition intervals, and a cubic spline interpolation function with a to-be-solved parameter corresponding to each sub-acquisition interval is constructed respectively; determining deformation characteristic information of the target pipeline, determining a boundary constraint condition based on the deformation characteristic information, solving the to-be-solved parameters in the cubic spline interpolation function based on the boundary constraint condition and the denoised pipeline deformation detection data of each data collection point, and determining an actual cubic spline interpolation function corresponding to the denoised pipeline deformation detection data based on a solving result; performing interpolation processing on the denoised pipeline deformation detection data based on the actual cubic spline interpolation function to obtain repaired denoised pipeline deformation detection data.

7. The method of claim 1, wherein, The method for performing data enhancement on the denoised pipeline deformation detection data in each collection channel comprises the following steps: For each acquisition channel, determine a data matrix corresponding to the noise reduction pipeline deformation detection data, and determine a horizontal direction convolution kernel of the data matrix and a vertical direction convolution kernel ; based on the horizontal direction convolution kernel and the vertical direction convolution kernel , determine a gradient amplitude matrix of the data matrix wherein, ; The gradient magnitude matrix The data matrix is ​​fused to obtain a fusion matrix, and the fusion matrix is ​​contrast stretched. Based on the contrast stretching result, the noise-reduced pipeline deformation detection data after data enhancement is determined.

8. A pseudo-color imaging apparatus for pipeline deformation detection data, characterized by, comprises: an acquisition unit configured to acquire original pipeline deformation detection data collected by a plurality of collection channels for a target pipeline in response to a pseudo-color imaging signal of the target pipeline; a decomposition unit configured to perform layer-by-layer singular value decomposition on the original pipeline deformation detection data for each collection channel to obtain approximate pipeline deformation detection data and detail pipeline deformation detection data of each decomposition layer; a determination unit configured to determine an optimal decomposition layer that meets a decomposition requirement in each decomposition layer based on the approximate pipeline deformation detection data of each decomposition layer, and accumulate the detail pipeline deformation detection data of each collection channel of the optimal decomposition layer to obtain total detail pipeline deformation detection data; a replacement unit configured to determine a feature position corresponding to total detail pipeline deformation detection data greater than a preset threshold, and replace the approximate pipeline deformation detection data of the optimal decomposition layer at the feature position in each collection channel with original pipeline deformation detection data at the corresponding feature position to obtain denoised pipeline deformation detection data corresponding to the original pipeline deformation detection data in each collection channel; an image generation unit configured to perform at least one of baseline correction, data repair and data enhancement on the denoised pipeline deformation detection data in each collection channel to obtain processed pipeline deformation detection data in each collection channel, and generate a pipeline pseudo-color image of the target pipeline based on the processed pipeline deformation detection data in each collection channel.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program, when executed by a processor, implements the steps of the method of any one of claims 1 to 7.

10. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The computer program, when executed by a processor, implements the steps of the method of any one of claims 1 to 7.

Images