A method, apparatus and device for identifying a petroleum pipeline

Abstract

The embodiment of the specification relates to the technical field of oil pipeline identification, and relates to an oil pipeline identification method, device and equipment, which comprises the following steps: acquiring multiple radar images collected by a drone; dividing each pixel point according to the relative difference of the gray value of each pixel point of each radar image to obtain multiple missing pixel points and multiple redundant pixel points; determining the missing redundancy ratio of each radar image according to the multiple missing pixel points and the multiple redundant pixel points; calculating the spatial discontinuity and length of the oil pipeline in each radar image; calculating the interpolation order of the multiple radar images according to the missing redundancy ratio of the multiple radar images and the spatial discontinuity and length of the oil pipeline in the multiple radar images; performing spline interpolation on the multiple missing pixel points and the multiple redundant pixel points of each radar image according to the interpolation order of each radar image; and identifying the oil pipeline according to the multiple radar images after spline interpolation.

Description

TECHNICAL FIELD

[0001] Embodiments of the present specification relate to the technical field of oil pipeline identification, and in particular to an oil pipeline identification method, device and equipment. BACKGROUND

[0002] Oil pipelines need to be regularly inspected and maintained to ensure their normal operation. In the event of an oil leak emergency, quickly and accurately identifying the location and direction information of the oil pipeline is crucial for effective emergency rescue. Since drones are not limited by traffic and terrain factors, for long-distance oil pipelines distributed in a wide geographical area, the inspection time and route can be flexibly arranged according to actual needs. Therefore, using a drone to inspect an oil pipeline can improve the speed and efficiency of the inspection.

[0003] Oil pipelines are buried in the ground, and geological activities can cause the position of the oil pipeline to shift. When using a drone to inspect such an oil pipeline, it first needs to be identified. Therefore, a ground penetrating radar can be installed on the drone to identify the position and direction of the oil pipeline in the ground through images collected by the ground penetrating radar. However, since the flight control system of the drone and the data acquisition system of the ground penetrating radar are two independent subsystems, when the flight speed of the drone and the data acquisition frequency of the radar are not synchronized, the collected images can be missing or redundant, affecting the identification of the position and direction of the underground oil pipeline. Using a spline interpolation algorithm to interpolate and supplement the missing area can improve the identification probability of the oil pipeline, but the existing method uses a fixed order for interpolation, which has poor adaptability, and thus leads to low accuracy and unsatisfactory efficiency of oil pipeline identification. Specifically, too low an order will result in non-smoothness after interpolation, which cannot improve the accuracy of oil pipeline identification; and too high an order will greatly increase the computational load, which is not conducive to real-time oil pipeline identification.

[0004] Therefore, how to overcome the problem of poor adaptability in existing methods for interpolating oil pipeline radar images, which in turn leads to low accuracy and unsatisfactory efficiency of oil pipeline identification, and improve the flexibility of interpolating oil pipeline radar images, and thus improve the accuracy and efficiency of oil pipeline identification, is a key problem that needs to be solved urgently. SUMMARY

[0005] The purpose of the embodiments of the present specification is to provide an oil pipeline identification method, device and equipment to overcome the problem of poor adaptability in existing oil pipeline identification methods for interpolating oil pipeline radar images, which in turn leads to low accuracy and unsatisfactory efficiency of oil pipeline identification.

[0006] To solve the above technical problems, the specific technical solutions of the embodiments of the present specification are as follows:

[0007] In one aspect, the embodiments of the present specification provide a method for identifying a petroleum pipeline, comprising:

[0008] acquiring a plurality of radar images collected by a UAV;

[0009] dividing each pixel point according to the relative difference of the gray value of each pixel point in each radar image to obtain a plurality of missing pixel points and a plurality of redundant pixel points;

[0010] determining a missing-redundant ratio of each radar image according to the plurality of missing pixel points and the plurality of redundant pixel points;

[0011] calculating the spatial discontinuity and length of the petroleum pipeline in each radar image according to the relative difference of the gray value of each pixel point in the pipeline region in each radar image;

[0012] calculating the interpolation order of the plurality of radar images according to the missing-redundant ratio of the plurality of radar images and the spatial discontinuity and length of the petroleum pipeline in the plurality of radar images;

[0013] performing spline interpolation on the plurality of missing pixel points and the plurality of redundant pixel points of each radar image according to the interpolation order of each radar image;

[0014] identifying the petroleum pipeline according to the plurality of radar images after spline interpolation.

[0015] In another aspect, the embodiments of the present specification provide an identification device for a petroleum pipeline, comprising:

[0016] an acquisition module configured to acquire a plurality of radar images collected by a UAV;

[0017] a division module configured to divide each pixel point according to the relative difference of the gray value of each pixel point in each radar image to obtain a plurality of missing pixel points and a plurality of redundant pixel points;

[0018] a determination module configured to determine a missing-redundant ratio of each radar image according to the plurality of missing pixel points and the plurality of redundant pixel points;

[0019] a first calculation module configured to calculate the spatial discontinuity and length of the petroleum pipeline in each radar image according to the relative difference of the gray value of each pixel point in the pipeline region in each radar image;

[0020] a second calculation module configured to calculate the interpolation order of the plurality of radar images according to the missing-redundant ratio of the plurality of radar images and the spatial discontinuity and length of the petroleum pipeline in the plurality of radar images;

[0021] an interpolation module configured to perform spline interpolation on the plurality of missing pixel points and the plurality of redundant pixel points of each radar image according to the interpolation order of each radar image;

[0022] The identification module is configured to identify the oil pipeline according to the plurality of radar images after the spline interpolation.

[0023] In another aspect, a computer device is provided, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor executes the above-mentioned method for identifying the oil pipeline.

[0024] In another aspect, the embodiments of the present specification also provide a computer program product, which executes the instructions of any one of the above-mentioned methods when executed by the processor of the computer device.

[0025] As can be seen from the technical solutions provided by the above embodiments of the present specification, the embodiments of the present specification can obtain a plurality of radar images collected by a UAV; divide each pixel point according to the relative difference of the gray value of each pixel point in each radar image to obtain a plurality of missing pixel points and a plurality of redundant pixel points; determine the missing-redundant ratio of each radar image according to the plurality of missing pixel points and the plurality of redundant pixel points; calculate the spatial discontinuity and length of the oil pipeline in each radar image according to the relative difference of the gray value of each pixel point in the pipeline region of each radar image; calculate the interpolation order of the plurality of radar images according to the missing-redundant ratio of the plurality of radar images and the spatial discontinuity and length of the oil pipeline in the plurality of radar images; perform spline interpolation on the plurality of missing pixel points and the plurality of redundant pixel points of each radar image according to the interpolation order of each radar image; and identify the oil pipeline according to the plurality of radar images after the spline interpolation. Compared with the prior art, the embodiments of the present specification can determine the missing-redundant ratio of each radar image, the spatial discontinuity and length of the oil pipeline in each radar image by the difference of the pixel gray value in each radar image, directly associate the influence of the synchronization of the flight speed of the UAV and the radar collection frequency on the missing and redundancy of the pixel points in the radar image, the continuity of the oil pipeline, and the similar length of the oil pipeline. In addition, through joint analysis of multiple frames, the influence of noise or occlusion in a single image on pipeline identification can be reduced, the interpolation order can be adaptively selected according to the severity of the data asynchronization, overfitting or underfitting can be avoided, the smoothness and calculation efficiency of the interpolation result can be improved, the stability of the interpolation can be enhanced, and thus the accuracy and efficiency of the oil pipeline identification can be greatly improved. BRIEF DESCRIPTION OF DRAWINGS

[0026] In order to more clearly illustrate the technical solutions in the embodiments of the present specification or the prior art, the drawings needed in the embodiments or the prior art description will be briefly introduced as follows.

[0027] Figure 1 is a flowchart of an oil pipeline identification method provided by the embodiments of the present specification;

[0028] Figure 2is a schematic diagram of oil pipeline image acquisition based on a UAV radar provided by an embodiment of the present specification;

[0029] Figure 3 is a flowchart of a method for dividing missing pixel points and redundant pixel points of a radar image provided by an embodiment of the present specification;

[0030] Figure 4 is a flowchart of a method for determining the spatial discontinuity and length of an oil pipeline in a radar image provided by an embodiment of the present specification;

[0031] Figure 5 is a flowchart of a method for determining the interpolation order of a radar image provided by an embodiment of the present specification;

[0032] Figure 6 is a flowchart of a method for identifying the position and orientation of an oil pipeline provided by an embodiment of the present specification;

[0033] Figure 7 is a flowchart of a method for determining the spatial discontinuity of an oil pipeline in a radar image provided by an embodiment of the present specification;

[0034] Figure 8 is a flowchart of a method for determining the length of an oil pipeline in a radar image provided by an embodiment of the present specification;

[0035] Figure 9 is a schematic diagram of a radar image B acquired at a certain moment provided by an embodiment of the present specification;

[0036] Figure 10 is a schematic diagram of a radar image C obtained after spline interpolation of the radar image B provided by an embodiment of the present specification;

[0037] Figure 11 is a schematic diagram of the structural composition of an oil pipeline identification device provided by an embodiment of the present specification;

[0038] Figure 12 is a schematic diagram of the structural composition of a computer device provided by an embodiment of the present specification. DETAILED DESCRIPTION

[0039] The technical solutions in the embodiments of the present specification will be described clearly and completely below with reference to the drawings in the embodiments of the present specification. Obviously, the described embodiments are only part of the embodiments of the present specification, rather than all the embodiments. Based on the embodiments in the present specification, all other embodiments obtained by those of ordinary skill in the art without creative labor should fall within the scope of protection of the present specification.

[0040] It should be noted that the terms "first," "second," and the like in this specification, the claims, and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a particular order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments of this specification described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having," and any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, apparatus, product, or device comprising a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to such processes, methods, products, or devices.

[0041] Figure 1 This is a flow chart of a method for identifying an oil pipeline provided in an embodiment of this specification. When implemented, the method includes the following steps:

[0042] S101: Acquire multiple frames of radar images collected by the UAV.

[0043] In some embodiments, multiple frames of radar images collected by a drone may be acquired.

[0044] Oil pipelines are laid over long distances, covering vast areas, and operating in complex terrain. This complex environment undoubtedly places extremely high demands on pipeline measurement, requiring precise understanding of both the pipeline's direction and its location. Traditional measurement methods struggle to meet these requirements in terms of efficiency and accuracy. However, the use of drones equipped with ground-penetrating radar (GPR) equipment has opened up a new avenue for oil pipeline measurement, enabling efficient and accurate identification of pipelines, thereby clearly defining their direction and location.

[0045] To ensure the accuracy and stability of detection data, the ground-penetrating radar device can be installed at the center of the bottom of the drone. This installation method helps the device to evenly receive and transmit signals. After the drone can smoothly take off from the pre-set take-off point, it can begin cruising at a preset stable flight speed according to the pre-planned flight route. At the same time, the signal transmitter of the ground-penetrating radar continuously emits high-frequency electromagnetic pulses to the ground at a pre-set transmission frequency. Figure 2 As shown in the figure, as these pulse signals propagate through the soil, if they encounter an interface between different media, such as an oil pipeline, some of the pulse signals will be reflected back, like an echo, due to the differences in the physical properties of the media. These reflected signals are accurately captured by the signal receiver of the ground-penetrating radar equipment. To fully record the detection data, the collected data is integrated into a detection radar image frame at scientifically calculated intervals.

[0046] The connection between the ground penetrating radar device and the unmanned aerial vehicle is firm, so that loosening or displacement during flight can be avoided. During flight, the unmanned aerial vehicle always maintains a predetermined flight speed, and every time a preset length is flown, the system triggers a data acquisition program to generate a new detection radar image. Through such an orderly acquisition method, the oil pipeline can be continuously and comprehensively detected.

[0047] Each frame of continuously acquired detection radar image can be filtered, denoised and image enhanced. In the filtering and denoising link, a Gaussian filtering algorithm can be used, which can effectively suppress the Gaussian noise in the image, and through weighted average of image pixels, the image is smoothed while the edge information of the image is preserved to the greatest extent. Secondly, a histogram equalization algorithm can be used for image enhancement, which adjusts the gray scale distribution of the image to enhance the contrast of the image, so that the details in the image are more clear and visible. After filtering, denoising and image enhancement, each frame of detection radar image can also be converted into a gray image. The gray image can simplify the image data structure, highlight the gray difference between the pipeline and the surrounding medium, and facilitate subsequent image analysis and recognition. Through a series of rigorous data acquisition and processing processes, each frame of clear and accurate radar image is obtained, providing a solid data foundation for the analysis and management of oil pipelines.

[0048] S102: According to the relative difference of the gray value of each pixel point of each frame of radar image, each pixel point is divided to obtain a plurality of missing pixel points and a plurality of redundant pixel points.

[0049] From the perspective of digital image processing, the radar image is essentially a two-dimensional gray matrix, and the gray value of each pixel point corresponds to the intensity information of the reflection signal at a specific position, with a value range of 0 to 255. When the reflection signal intensity is strong, it means that more electromagnetic pulse energy is reflected back to the receiving antenna of the ground penetrating radar, and after analog-to-digital conversion and quantization by the signal processing unit, a higher gray value is assigned to the corresponding pixel point, which presents a bright visual effect in the visualized radar image; on the contrary, if the reflection signal intensity is weak, only a small amount of electromagnetic pulse energy returns, and the gray value of the pixel point will be lower, which appears as a dark tone in the image.

[0050] At the interface between the oil pipeline and the surrounding medium, due to the significant difference in dielectric constant between the two, electromagnetic pulses will be strongly reflected at this interface, and a large amount of reflected signals will be received and converted into high gray value pixel points. These high gray value pixel points are connected to each other, outlining bright lines or areas on the radar image, thus clearly presenting the contour and direction of the oil pipeline.

[0051] However, when the unmanned aerial vehicle carries the ground penetrating radar for detection, the coordination problem between the flight speed of the unmanned aerial vehicle and the data acquisition frequency of the ground penetrating radar is faced. The flight speed of the unmanned aerial vehicle can be in the range of 10-30 kilometers per hour, while the data acquisition frequency of the ground penetrating radar can be as high as hundreds or even thousands of times per second. The lack of precise time synchronization between the two can cause a series of data acquisition problems. For example, when the unmanned aerial vehicle flies at a faster speed and the radar data acquisition frequency is relatively low, the unmanned aerial vehicle can have flown a long distance in the time interval between the adjacent two data acquisitions, resulting in a large deviation of the ground area corresponding to the adjacent pixel points, and the radar image collected will appear to be misaligned. Conversely, if the data acquisition frequency is too high and the flight speed of the unmanned aerial vehicle is too slow, the same ground area can be collected multiple times, causing data redundancy. In addition, time synchronization errors can cause the features of the oil pipeline in the radar image to be abnormal, that is, the continuity of the oil pipeline can be broken, which appears as a jump or break in the line on the radar image.

[0052] For the image data missing or redundancy caused by the asynchronization of the flight speed of the unmanned aerial vehicle and the data acquisition frequency of the radar, a spline interpolation algorithm can be used to improve it. If there is a data missing area in the detected radar image, a spline interpolation algorithm can be used to fill in the missing data. The spline interpolation algorithm is a mathematical method for constructing a smooth curve or surface between given data points. In the ground penetrating radar image, the order of spline interpolation determines the smoothing effect. High-order spline interpolation can make the curve smoother, but the calculation time is greatly prolonged, and the risk of overfitting is increased, which may identify noise as pipeline features or distort the shape of the pipeline, thereby affecting the correct judgment of the pipeline position, shape and depth information. Therefore, too high an order of the spline interpolation algorithm can cause large fluctuations in the interpolation results, which may appear as unnatural jumping or blurring in the interpolated detection radar image. Therefore, a suitable order can be selected to smooth the radar image.

[0053] In some embodiments, according to the relative difference between the gray value of each pixel point of each frame of radar image and the gray value of each edge pixel point in a preset spatial window centered on each pixel point, a gray difference value of each pixel point of each frame of radar image is determined; a histogram of gray difference values of all pixel points of each frame of radar image is obtained; a first gray difference threshold is determined according to a valley between two wave crests with the smallest gray difference values in the histogram; a second gray difference threshold is determined according to a valley between two wave crests with the largest gray difference values in the histogram; pixel points with a gray difference value less than the first gray difference threshold are divided into missing pixel points; and pixel points with a gray difference value greater than the second gray difference threshold are divided into redundant pixel points.

[0054] The data acquisition frequency asynchronization leads to the unevenness of signal intensity data in the spatial distribution in the detection radar image, which can be manifested as the absence and redundancy of pixel points in some areas. The unevenness degree can be represented by the gray distribution of the pixel points in each frame of radar image, so that the pixel points in the redundant and missing areas of each frame of radar image can be accurately divided, and then the pixel points in the redundant and missing areas of each frame of radar image can be accurately interpolated to obtain a clear and continuous pipeline image.

[0055] Figure 3 is a flowchart of a method for dividing missing and redundant pixel points of a radar image provided by an embodiment of the present specification. In specific implementation, the following steps are included:

[0056] S1021: determining the gray difference value of each pixel point of each frame of radar image according to the relative difference of the gray value of each pixel point and the gray value of each edge pixel point in a preset spatial window centered on each pixel point.

[0057] A preset spatial window is demarcated with each pixel point as the center. The selection of the window size can comprehensively consider the resolution of the radar image, the detail features of the detected target and other factors. For example, for a radar image with high resolution and rich pipeline details, a 3x3 or 5x5 window can be selected; and for an image with relatively low resolution and macroscopic target, a 7x7 or even larger window can be selected. Within this preset spatial window, the relative difference between the gray value of the center pixel point and the gray value of each edge pixel point can be calculated. The calculation method of the relative difference can adopt the absolute value of the difference, that is, the difference between the gray value of the center pixel point and the gray value of the edge pixel point is calculated to measure the difference degree between them. Specifically, all the pixel points (except the center pixel point) in the window can be traversed, the absolute value of the difference between the gray value of the center pixel point and each traversed pixel point is calculated, and then the multiple absolute values are processed, such as taking the weighted average, maximum value, etc., to determine the final gray difference value of the center pixel point. This calculation method can directly reflect the change of the center pixel point and the surrounding pixel points in the gray value. If the center pixel point is in a uniform area, the gray difference value with the edge pixel point is small; on the contrary, if it is in an edge or a feature change obvious area, the gray difference value will be larger.

[0058] S1022: obtaining a histogram of the gray difference values of all pixel points of each frame of radar image.

[0059] The histogram takes the gray difference value as the abscissa and the number of pixels with the same gray difference value as the ordinate, and can intuitively show the distribution of the gray difference value in the whole radar image. In the radar image obtained by oil pipeline detection, the internal composition is complex and diverse, mainly covering background area, oil pipeline area and noise and other different components. The gray distribution characteristics of these different areas in the image are intuitively presented through the distribution histogram.

[0060] The background area occupies a large area in the radar image and has a relatively uniform characteristic. This is because in the underground environment where the oil pipeline is laid, the surrounding soil, rock and other media are relatively stable within a certain range, so that the intensity of the electromagnetic pulse reflection signal changes little, and on the image, it is a large number of pixel points with similar gray values. This uniformity is reflected in the distribution histogram, forming a relatively obvious and relatively flat peak. The gray difference value interval corresponding to the peak concentrates a large number of gray difference data from the background area pixels.

[0061] The oil pipeline area has a unique feature in the radar image. Due to the significant difference in dielectric constant between the oil pipeline and the surrounding medium, the electromagnetic pulse is strongly reflected on the surface of the pipeline, making the gray value of the pipeline area pixel significantly higher than that of the background area, and the gray change is violent. In the distribution histogram, this characteristic can be represented as another peak that is distinguished from the background area peak. The gray difference value interval corresponding to the peak can represent the gray difference concentration range of the pipeline area pixels.

[0062] The peak and valley values in the distribution histogram are closely related to the image features. The peak value represents a region with relatively high gray difference, which can correspond to the part of the image with obvious feature change, such as the oil pipeline area; while the valley value can represent a region with relatively low gray difference, which is located in the transition zone between different feature regions, or a part of the image with relatively flat gray change, such as the internal part of the background area and the transition part between the background area and the pipeline area. Through the analysis of the peak and valley values, not only can the different features in the image such as oil pipeline, background area and noise be identified, but also the distribution range, feature strength and boundary relationship between each region can be further understood, thereby providing an important basis for subsequent image filtering, feature extraction and target recognition and other processing.

[0063] S1023: Determine a first gray difference threshold according to the valley between the two peaks with the smallest gray difference value in the histogram.

[0064] The first gray difference threshold value can be determined according to a wave trough between two wave crests with minimum gray difference values in the histogram. The data other than zero in the histogram can be fitted to obtain a distribution curve of the gray difference values, each wave crest in the histogram corresponds to a maximum value of the distribution curve, and each wave trough can correspond to a minimum value of the distribution curve, and the wave trough between the two wave crests can correspond to the minimum minimum value between the two maximum values. The two wave crests can correspond to the background region and the relatively smooth pipeline region respectively, and the wave trough between them reflects the transition point from the background to the pipeline feature. By selecting the gray difference value at the wave trough as the first threshold value, the pixel points in the uniform or similar region with small gray difference can be effectively screened out.

[0065] S1024: determining a second gray difference threshold value according to a wave trough between two wave crests with maximum gray difference values in the histogram.

[0066] The second gray difference threshold value can be determined according to a wave trough between two wave crests with maximum gray difference values in the histogram. The two wave crests can represent regions with obvious pipeline feature changes and other regions, and the gray difference value at the wave trough can be used as a limit to distinguish between significant feature changes and normal changes.

[0067] S1025: dividing the pixel points with gray difference values less than the first gray difference threshold value into missing pixel points.

[0068] Based on the two determined threshold values, the pixel points can be divided. The pixel points with gray difference values less than the first gray difference threshold value are likely to be in the missing region in the image, because the gray difference between them and the surrounding pixel points is small, that is, the effective information in these regions can be missing due to various reasons in the detection process (such as signal shielding, acquisition error, etc.), so they are divided into missing pixel points.

[0069] S1026: dividing the pixel points with gray difference values greater than the second gray difference threshold value into redundant pixel points.

[0070] On the contrary, the pixel points with gray difference values greater than the second gray difference threshold value have too large gray difference with the surrounding pixel points, which exceeds the normal feature change range, and such pixel points are likely to be generated due to data acquisition redundancy or noise interference, so they are divided into redundant pixel points.

[0071] S103: determining a missing redundancy ratio of each frame of radar image according to the plurality of missing pixel points and the plurality of redundant pixel points.

[0072] In some embodiments, a missing redundancy ratio of each frame of radar image is determined according to the plurality of missing pixel points and the plurality of redundant pixel points.

[0073] The ratio of the number of missing pixels to the number of redundant pixels in all regions in each frame of the detection radar image can be taken as the missing-redundancy ratio of each frame of the radar image.

[0074] The smaller the missing-redundancy ratio is, the more serious the redundancy of data collection in the corresponding detection radar image is, which can be caused by too high collection frequency or repeated collection in the collection process. On the contrary, the more serious the missing of data collection in the corresponding detection radar image is, which can be caused by the problem of synchronous collection.

[0075] When the missing-redundancy ratio is at a low level, the high-density redundancy of data will cause the pixel distribution to present an irregular discrete state. At this time, if a high-order spline interpolation algorithm is used, the redundant and discrete data points will be over-fitted, causing the texture of the interpolated image to be distorted, the edges to be blurred, and other problems, which will instead hide the true shape of the oil pipeline. Therefore, by reducing the order of the spline interpolation algorithm, the interference caused by redundant data can be effectively smoothed, while the key features of the pipeline are retained and the overall clarity of the image is improved.

[0076] When the missing-redundancy ratio is high, i.e., the data missing problem is prominent, there are a large number of information "blind areas" in the radar image. When detecting oil pipelines in complex geological structures, due to the strong attenuation of electromagnetic signals by rock layers, only scattered pixel points are present in the radar image for some regions of the pipeline. If a low-order spline interpolation algorithm is used, it is difficult to construct a reasonable pixel transition relationship in a large area of missing regions, and the reconstructed image will have problems such as pipeline shape rupture and detail loss. At this time, the order of the spline interpolation algorithm can be increased. High-order spline interpolation can infer the pixel distribution rule of the missing region based on the known pixel points through complex curve fitting calculation, especially when dealing with complex curved shapes, branch structures, and other features of oil pipelines. The spatial topological relationship of the pipeline can be more accurately restored, and the true restoration degree of the image to the actual condition of the underground pipeline can be ensured.

[0077] S104: According to the relative difference of the gray value of each pixel point in the pipeline region in each frame of the radar image, the spatial discontinuity and length of the oil pipeline in each frame of the radar image are calculated.

[0078] In some embodiments, a grayscale difference value of each pixel in each radar image frame is determined based on the relative difference between the grayscale value of each pixel in each radar image frame and the grayscale values ​​of each edge pixel in a preset spatial window centered on each pixel; each radar image frame is segmented into multiple image regions based on the grayscale difference value of each pixel in each radar image frame; the image region containing the largest number of edge contours among the multiple image regions is used as the pipeline region of each radar image frame; a texture feature extraction algorithm is used to extract the texture feature value of each pixel in the pipeline region of each radar image frame; the spatial discontinuity of the oil pipeline in each radar image frame is calculated based on the missing redundancy ratio of each radar image frame and the texture feature value of each pixel in the pipeline region of each radar image frame; and based on a traversal search algorithm, each edge pixel in the pipeline region of each radar image frame is searched to obtain the length of the oil pipeline in each radar image frame.

[0079] Oil pipelines exhibit a certain degree of spatial continuity in radar detection images. If data acquisition synchronization is compromised, the edges of the pipelines in the radar images may appear jagged or broken, resulting in discontinuities. Therefore, by calculating pixel grayscale differences, the contrast between the pipeline and the background in the image is effectively enhanced. Image segmentation and pipeline region recognition are used to successfully extract the pipeline region from the complex image. Furthermore, texture feature extraction and spatial discontinuity calculation are combined to assess the pipeline's integrity in the image. Finally, a traversal search algorithm is used to accurately measure the pipeline's length. These processing results provide a reliable basis for subsequent pipeline identification decisions.

[0080] Figure 4 This is a flow chart of a method for determining the spatial discontinuity and length of an oil pipeline in a radar image provided by an embodiment of this specification. When implemented, the method includes the following steps:

[0081] S1041: Determine a grayscale difference value for each pixel in each radar image frame based on a relative difference between the grayscale value of each pixel in each radar image frame and the grayscale values ​​of each edge pixel in a preset spatial window centered on each pixel.

[0082] For each pixel in each frame of radar image, a preset spatial window can be drawn with it as the center. After the window is determined, the relative difference of the gray scale values of the center pixel and the edge pixels in the window can be calculated. For example, the absolute value of the difference of the gray scale values can be calculated. By traversing all the edge pixels in the window and performing the above calculation, a plurality of gray scale difference values can be obtained, and these values can be processed, such as taking the average value, the maximum value, etc., to determine the final gray scale difference value of the center pixel. For example, in a 5x5 window, the gray scale value of the center pixel is 120, and the gray scale values of its eight edge pixels are 115, 122, 118, 125, 110, 128, 116, and 123. By calculating the absolute value of each difference and taking the average value, the gray scale difference value of the pixel is about 3.75.

[0083] S1042: Based on the gray scale difference values of each pixel in each frame of radar image, each frame of radar image is segmented into a plurality of image regions.

[0084] Based on the calculated gray scale difference values of each pixel in each frame of radar image, the image can be further segmented into a plurality of image regions. The image segmentation method can include threshold segmentation method, region growing method, clustering algorithm, etc. The threshold segmentation method can set one or more thresholds according to the gray scale difference value, and divide the pixels with similar gray scale difference values into the same region; the region growing method can start from a seed point and gradually merge adjacent pixels to form a region according to the similarity criterion of the gray scale difference values of the pixels; the clustering algorithm, such as K-means algorithm, can divide the image into a plurality of category regions by clustering analysis of the gray scale difference values of the pixels. For example, using the threshold segmentation method, if the pixels with a gray scale difference value less than 5 are divided into a background region, the pixels with a gray scale difference value greater than 15 are divided into a pipeline suspected region, and the pixels with a gray scale difference value between 5 and 15 are divided into a transition region, the radar image can be preliminarily segmented into different regions.

[0085] S1043: The image region containing the most edge contours in the plurality of image regions is taken as the pipeline region of each frame of radar image.

[0086] In the segmented image regions, an image region with the most edge contours can be identified as the pipeline region of each frame of radar image. The oil pipeline has obvious edge features in the radar image, and the difference in dielectric constant between the pipeline and the surrounding medium leads to different electromagnetic pulse reflection characteristics, which is reflected in the image that the pipeline region has rich edge contours. In order to accurately count the number of edge contours of each region, edge detection algorithms such as Canny edge detection algorithm, Sobel operator, etc. can be used. Taking the Canny edge detection algorithm as an example, the edge information in the image can be accurately extracted through the steps of Gaussian filtering to smooth the image, calculating the gradient amplitude and direction, non-maximum suppression and hysteresis threshold processing. When counting the number of edge contours of each region, the connected component analysis can be performed on the image after edge detection, and the number of independent edge connected components in each region is calculated. The region with the most number is the pipeline region.

[0087] S1044: According to the gray difference values of each pixel point in the pipeline region of each frame of radar image, the texture feature value of each pixel point is extracted using a texture feature extraction algorithm.

[0088] After determining the pipeline region, the texture feature value of each pixel point can be extracted using a texture feature extraction algorithm according to the gray difference value of each pixel point. Texture features can reflect the spatial distribution pattern of pixel gray values in the image, and are of great significance for identifying oil pipeline features. Commonly used texture feature extraction algorithms include gray level co-occurrence matrix (GLCM) and local binary pattern (LBP). Taking the gray level co-occurrence matrix as an example, the co-occurrence matrix can be constructed by calculating the frequency of pixel pairs with specific gray values in a specific direction and distance. Based on the co-occurrence matrix, texture feature parameters such as energy, entropy, contrast, and correlation can be further calculated, which describe the texture characteristics of the image from different angles. For example, the contrast parameter reflects the clarity of the image texture, and the higher the contrast, the greater the difference between the pixel gray values in the image, and the clearer the texture; the energy parameter represents the uniformity of the image texture, and the greater the energy value, the more concentrated the distribution of pixel gray values in the image, and the more regular the texture.

[0089] S1045: According to the missing redundancy ratio of each frame of radar image and the texture feature value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated.

[0090] The missing redundancy ratio can reflect the integrity and effectiveness of image data acquisition, and the texture feature value embodies the texture characteristics of the pipeline area, and the combination of the two can comprehensively evaluate the continuity of the oil pipeline in space. In specific calculation, a mathematical model containing the missing redundancy ratio and the texture feature parameter can be constructed. For example, the missing redundancy ratio is taken as a weight coefficient, and the contrast, correlation and other parameters in the texture feature are weighted and summed to obtain a comprehensive index, which can be normalized and used as the spatial discontinuity degree. If the missing redundancy ratio of a certain frame of radar image is large, it means that the data is missing seriously, which may lead to the breakage of the pipeline in the image; at the same time, if the contrast of the texture feature is low, it means that the texture of the pipeline area is fuzzy, which can also reflect the discontinuity of the pipeline. Through the calculation of the spatial discontinuity degree, the spatial integrity of the oil pipeline in the radar image can be quantitatively evaluated, which provides an important reference for the subsequent pipeline identification.

[0091] S1046: Based on the traversal search algorithm, each edge pixel point in the pipeline area of each frame of radar image is searched to obtain the length of the oil pipeline in each frame of radar image.

[0092] Based on the traversal search algorithm, each edge pixel point in the pipeline area of each frame of radar image is searched to obtain the length of the oil pipeline in each frame of radar image. The traversal search algorithm can use breadth-first search (BFS) or depth-first search (DFS) strategy. Taking breadth-first search as an example, starting from a certain edge pixel point of the pipeline area, in the order from near to far, the adjacent edge pixel points of the point are visited in turn, and the visited points are marked to avoid repeated access. In the access process, the number of pixel points passed is recorded, and when all the reachable edge pixel points are traversed, the length of the oil pipeline in the image can be calculated according to the number of pixel points and the scale of the image. In complex radar images, the pipeline may have branches, bends and other situations, at this time the traversal search algorithm can accurately track and calculate the total length of the pipeline along the edge path, providing accurate data support for the surveying and management of the oil pipeline.

[0093] S105: According to the missing redundancy ratio of the multiple frames of radar images and the spatial discontinuity degree and length of the oil pipeline in the multiple frames of radar images, the interpolation order of the multiple frames of radar images is calculated.

[0094] In some embodiments, a feature vector of the oil pipeline is constructed for each radar image frame based on the spatial discontinuity and length of the oil pipeline in each radar image frame; a radar image set centered on each radar image frame is determined based on a preset time window; the integrity of the oil pipeline in each radar image frame is determined based on the similarity between the feature vector of the oil pipeline in each radar image frame and the feature vectors of the oil pipelines in each radar image in the radar image set; and an interpolation order for each radar image frame is determined based on the missing-to-redundancy ratio of each radar image frame and the spatial discontinuity and integrity of the oil pipeline in each radar image frame.

[0095] By constructing pipeline feature vectors, massive amounts of image data can be rapidly structured. By building image collections based on preset time windows and calculating their completeness, images with pipeline status changes can be automatically identified. Furthermore, by determining the interpolation order based on the missing-to-redundancy ratio, spatial discontinuity, and completeness, problematic images can be repaired, significantly improving image usability and the accuracy of pipeline status assessments.

[0096] Figure 5 : is a flowchart of a method for determining a radar image interpolation order provided in an embodiment of this specification. When implemented, the method includes the following steps:

[0097] S1051: Constructing a feature vector of the oil pipeline in each radar image frame according to the spatial discontinuity and length of the oil pipeline in each radar image frame.

[0098] The spatial discontinuity and length of the oil pipeline in each radar image frame can be combined into an ordered vector, known as a pipeline feature vector, for example, represented as [spatial discontinuity, length]. This vector representation effectively integrates the multidimensional characteristics of the pipeline, providing a standardized data structure for subsequent comparative analysis and condition assessment.

[0099] S1052: Determine a radar image set centered around each radar image frame according to a preset time window.

[0100] After constructing the feature vector, we can determine the radar image set centered around each radar image frame based on a preset time window. The size of the time window can be preset based on the drone's flight speed and the frequency of radar image acquisition.

[0101] S1053: Determine the integrity of the oil pipeline in each radar image frame according to the similarity between the oil pipeline feature vector in each radar image frame and the oil pipeline feature vector in each radar image in the radar image set.

[0102] Methods for calculating feature vector similarity include Euclidean distance and cosine similarity. Euclidean distance measures the spatial distance between vectors by taking the square root of the sum of the squared differences between corresponding elements of two vectors. A closer distance indicates a higher similarity. Cosine similarity measures the directional similarity of two vectors by taking the cosine of the angle between them. A value closer to 1 indicates a higher similarity.

[0103] Taking cosine similarity as an example, the following formula can be used to calculate the similarity between the oil pipeline feature vector of each radar image frame and the oil pipeline feature vector of each radar image in the radar image set:

[0104] ;

[0105] Where, Indicates the The integrity of oil pipelines in frame radar images, Indicates the The oil pipeline feature vector corresponding to the frame radar image.

[0106] The higher the completeness, the higher the similarity of the oil pipelines between consecutive radar images, that is, the more complete the extracted features are, and the higher the synchronization of data acquisition; on the contrary, the worse the synchronization is when collecting data for the same oil pipeline.

[0107] The greater the degree of integrity, the lower the order of spline interpolation, because the oil pipeline features extracted from the image are more complete. This means that data acquisition synchronization has a minimal impact on the features, the image data quality is relatively high, and the data can better reflect the true shape, texture, and other characteristics of the oil pipeline. In this case, complex methods such as high-order spline interpolation cannot be used to repair or reconstruct the data. On the contrary, the order of spline interpolation should be higher.

[0108] S1054: Determine an interpolation order for each radar image frame according to a missing-to-redundancy ratio of each radar image frame and a spatial discontinuity and integrity of the oil pipeline in each radar image frame.

[0109] After determining the completeness of the oil pipeline in each frame of radar image, the interpolation order of the image is finally determined in combination with the missing redundancy ratio and the spatial discontinuity of each frame of image. The missing redundancy ratio reflects the severity of the missing or redundancy problem existing in the data acquisition process. The smaller the ratio, the more serious the data redundancy; the larger, the more prominent the data missing problem. The spatial discontinuity reflects the morphological integrity of the pipeline in the image. The higher the value, the more obvious the discontinuity of the pipeline, such as fracture, misplacement, etc. The selection of the interpolation order directly affects the repair effect of the radar image and the reconstruction quality of the pipeline features. When the missing redundancy ratio is large, and the spatial discontinuity is high, and the completeness is low, it means that there is a large amount of data missing in the image, the pipeline form is seriously damaged, and the difference with the historical state is large. At this time, a higher interpolation order can be selected, such as cubic spline interpolation or higher order interpolation algorithm. High-order interpolation can infer the pixel information of the missing area as much as possible through complex curve fitting and surface construction according to the limited known data points, and restore the complete form of the pipeline.

[0110] On the contrary, if the missing redundancy ratio is small, the spatial discontinuity is low, and the completeness is high, it means that the image data quality is good, and the pipeline form is relatively complete. At this time, a lower interpolation order can meet the image repair requirements, such as linear spline interpolation or quadratic spline interpolation. The low-order interpolation algorithm has low computational complexity and can quickly complete the repair work on the premise of ensuring the basic quality of the image, avoiding overfitting problems that may be caused by high-order interpolation. For example, in daily pipeline inspection, most radar image data is collected normally, and the pipeline form is stable. Linear spline interpolation can effectively handle a small amount of local data anomalies, ensuring the clarity of the image and improving data processing efficiency.

[0111] S106: Perform spline interpolation on the plurality of missing pixel points and the plurality of redundant pixel points of each frame of radar image according to the interpolation order of each frame of radar image.

[0112] When performing spline interpolation on the missing pixel points, the set of known pixel points participating in the interpolation calculation can be determined. A specific range of neighborhood can be defined around the missing pixel point, and the known pixel points in the neighborhood can be selected as the basis for interpolation. For example, for a missing pixel point, the known pixel points within a range of 3x3 or 5x5 around it can be selected, a spline function can be constructed according to the selected interpolation order, and the estimated value of the missing pixel point can be calculated by the function to complete the filling of the missing area. In the filling process, the texture correlation and gray level change trend between pixel points can also be considered to make the newly generated pixel value naturally blend with the surrounding pixels, avoiding obvious splicing marks.

[0113] For redundant pixels, spline interpolation mainly plays a role in smoothing and noise reduction. By including redundant pixels in the interpolation calculation range, the gray values of these pixels can be adjusted to be consistent with the gray change trend of the surrounding pixels by using the smoothing characteristics of the spline function. For example, when there is a local gray abnormal area in the radar image caused by repeated collection, through spline interpolation processing, the gray fluctuation of this area can be effectively reduced, making the image overall smoother and more natural, and improving the visual effect of the image and the accuracy of subsequent analysis.

[0114] After processing, the pipeline features are clear and continuous, providing high-quality data support for subsequent pipeline identification.

[0115] S107: Identify the position and direction of the oil pipeline according to the multiple radar images after spline interpolation.

[0116] In some embodiments, the multiple radar images after spline interpolation are spliced; and the position and direction of the oil pipeline are identified according to the spliced radar images.

[0117] Figure 6 is a flowchart of a method for identifying the position and direction of an oil pipeline provided by an embodiment of the present specification, and in specific implementation, the following steps are included:

[0118] S1071: Splice the multiple radar images after spline interpolation.

[0119] Due to the changes in attitude, fluctuations in height, and other situations during the flight of the unmanned aerial vehicle, the radar images collected at different times differ in spatial position, rotation angle, and scale. Registration algorithms can be used to accurately splice the images, including feature point-based registration and region-based registration. Feature point-based registration methods, such as SIFT (Scale-Invariant Feature Transform) and SURF (Speeded-Up Robust Features) algorithms, can extract significant feature points such as corners and edges from images, calculate the descriptors of the feature points, and then match the feature points based on the similarity of the descriptors to determine the transformation relationship between the images. Region-based registration methods can find the best matching position by calculating the similarity (such as mean square error and cross-correlation coefficient) between image blocks.

[0120] After image registration, geometric correction can be performed on the images. Due to the influence of radar detection principles and the flight attitude of the unmanned aerial vehicle, the collected radar images may have perspective distortion, nonlinear distortion, and other problems. By establishing geometric transformation models such as affine transformation and perspective transformation, the images can be corrected. Affine transformation can handle translation, rotation, and scaling of images, while perspective transformation can correct more complex perspective distortion. For example, for image distortion caused by tilting the unmanned aerial vehicle, perspective transformation can be used to correct the image to an orthographic view, making the position and shape of the pipeline in the image more consistent with the actual situation.

[0121] S1072: Identify the position and direction of the oil pipeline according to the spliced radar image.

[0122] The image registration and geometric correction radar image can be spliced, and the position and direction of the oil pipeline can be identified based on the spliced radar image. Median filtering, Gaussian filtering and other methods can be used to remove noise in the image, and histogram equalization, contrast limited adaptive histogram equalization and other techniques can be used to enhance the contrast of the image, so that the difference between the oil pipeline and the background is more obvious.

[0123] The edge information of the pipeline can be extracted using edge detection algorithms. Classical edge detection algorithms such as Canny and Sobel operators can effectively detect areas with sharp gray level changes in the image, i.e. the edges of the pipeline. Taking Canny edge detection as an example, it extracts continuous and accurate pipeline edges through steps such as Gaussian filtering, gradient amplitude and direction calculation, non-maximum suppression and hysteresis threshold processing.

[0124] After obtaining the pipeline edge, a line detection algorithm can be used to determine the position and direction of the pipeline. For example, Hough transform can be used to convert the straight line detection problem in the image space to the peak detection problem in the parameter space. By voting in the parameter space, the peak point representing the straight line is found, and then converted back to the image space to determine the straight line segment of the pipeline. For curved oil pipelines, a segmented detection method can be used to divide the pipeline into multiple small segments, and straight line detection is performed on each segment, and then curve fitting (such as least squares fitting quadratic curve, spline curve fitting, etc.) is used to restore the overall direction of the pipeline.

[0125] In addition, considering that the underground medium is a complex material system mainly composed of soil, rock and oil pipeline itself. The composition of soil is complex and variable, and according to its particle composition and water content, the dielectric constant can fluctuate between 2-30. The dielectric constant of rock is closely related to its mineral composition, porosity and water saturation, and the dielectric constant of common rock is in the range of 4-20. The oil pipeline is generally made of steel or plastic, and the dielectric constant of steel is close to 1, and the dielectric constant of plastic material is usually between 2-5. The dielectric constant difference between the oil pipeline and the surrounding soil is particularly significant. Taking the common steel oil pipeline as an example, the difference in dielectric constant between the steel oil pipeline and the surrounding moist soil (dielectric constant can reach more than 20) can reach dozens of times. This huge difference in dielectric constant makes the electromagnetic pulse signal produce strong reflection signal when it encounters the wall of the oil pipeline.

[0126] According to the propagation speed of electromagnetic pulses in the underground medium and the time delay of the reflected signal, the depth of the oil pipeline is calculated by the time-depth conversion algorithm. The propagation speed of electromagnetic waves in the underground medium is known (between 0.06-0.12 m / ns, the specific value depends on the dielectric constant of the medium), combined with the time delay of the reflected signal from transmission to reception, through a simple formula: depth = propagation speed x time delay / 2 (divided by 2 because the signal needs to propagate back and forth), the depth of the oil pipeline in the underground can be accurately calculated.

[0127] In order to determine the position coordinates of the oil pipeline in the geographical space, the unmanned aerial vehicle detection system can also carry a high-precision GPS positioning system. When the ground penetrating radar receives the reflected signal, the GPS positioning system synchronously records the longitude, latitude and altitude information of the unmanned aerial vehicle. By combining the depth information of the oil pipeline with the position information of the unmanned aerial vehicle, the three-dimensional spatial position of the oil pipeline can finally be accurately marked in the geographic information system (GIS).

[0128] In some embodiments, according to a preset spatial window, a set of pixel points centered on each pixel point in the pipeline region is determined; according to the texture feature value of each pixel point in the pipeline region and the sum of the texture feature values of each pixel point in the corresponding pixel point set, the texture difference value of each pixel point in the pipeline region is obtained; according to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region in each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: ; In the formula, the spatial discontinuity of the oil pipeline in the i-th frame of radar image is represented as Di, the missing redundancy ratio of the i-th frame of radar image is represented as Ri, the number of pixel points in the pipeline region in the i-th frame of radar image is represented as Ni, the sum of the texture difference values of all pixel points in the pipeline region in the i-th frame of radar image is represented as Si.

[0129] Figure 7 is a flowchart of a method for determining the spatial discontinuity of an oil pipeline in a radar image provided by an embodiment of the present specification. In specific implementation, the following steps are included:

[0130] S701: According to a preset spatial window, a set of pixel points centered on each pixel point in the pipeline region is determined.

[0131] ​​​​The size of the spatial window can be preset based on the radar image resolution, the thickness of the oil pipeline, and the required detection accuracy. For example, for high-resolution radar images with rich pipeline details, a smaller 3×3 or 5×5 window can be used to precisely focus on local texture features. For lower-resolution images with relatively macroscopic pipeline targets, a larger 7×7 or even 9×9 window can be used to better capture overall texture information.

[0132] S702: Obtain a texture difference value for each pixel in the pipeline area according to the sum of the differences between the texture feature value of each pixel in the pipeline area and the texture feature values ​​of each pixel in the corresponding pixel set.

[0133] After determining the preset spatial window, a corresponding window is delineated with each pixel point in the pipeline area as the center, and all the pixels contained in the window can be included in the set. After constructing the pixel point set, the texture difference value of each pixel point in the pipeline area can be calculated. This calculation process can be based on the comparative analysis of the texture feature value of each pixel point itself and the texture feature values ​​of other pixels in the corresponding pixel point set. Specifically, the texture feature value of each pixel point can be extracted, and texture feature methods such as gray level co-occurrence matrix (GLCM) and local binary pattern (LBP) can be used. Taking the gray level co-occurrence matrix as an example, the co-occurrence matrix is ​​constructed by calculating the frequency of occurrence of pixel pairs with specific gray values ​​in a specific direction and distance in the image, and based on this, texture feature parameters such as energy, entropy, contrast, and correlation are further calculated.

[0134] After obtaining the texture feature value, the difference between each pixel and the texture feature value of other pixels in the corresponding set can be calculated. For example, the sum of the absolute values ​​of the differences of each texture feature parameter can be calculated. Taking the contrast parameter as an example, if the contrast value of the center pixel is , the contrast value of a pixel in the set is , then the difference between the two is By performing similar difference calculations on all pixels in the set and accumulating them, the texture difference value of the central pixel can be obtained. For example, for a certain pixel, the contrast, energy, entropy, and correlation parameter values ​​calculated by its gray level co-occurrence matrix are 、 、 、 , corresponding to the pixel point set The parameter values ​​of the pixels are 、 、 、 , then the texture difference value of the pixel The formula Calculate. Traverse all pixels in the pipeline area and repeat the above calculation process to obtain the texture difference value of each pixel.

[0135] S703: Calculate the spatial discontinuity of the oil pipeline in each radar image frame according to the missing redundancy ratio of each radar image frame and the texture difference value of each pixel point in the pipeline area in each radar image frame.

[0136] After obtaining the texture difference values ​​for each pixel in the pipeline area, the spatial discontinuity of the oil pipeline can be determined by combining the missing / redundant ratio of each radar image frame. The missing / redundant ratio is an important indicator of image data acquisition quality, reflecting the severity of missing or redundant issues during the data acquisition process. The spatial discontinuity is calculated by comparing the number of missing pixels to the number of redundant pixels in all regions of each radar image frame. A smaller ratio indicates a more severe redundancy in the data acquisition; a higher ratio indicates a more prominent data missing problem.

[0137] For example, the following formula can be used to calculate the spatial discontinuity of the oil pipeline in each frame of radar image:

[0138] ;

[0139] Where, Indicates the Spatial discontinuity of oil pipelines in frame radar images, Indicates the The missing redundancy ratio of the frame radar image, Indicates the The number of pixels in the pipeline area of ​​the frame radar image, Indicates the The sum of the texture difference values ​​of all pixels in the pipeline area in the frame radar image.

[0140] The larger the mean of the local texture difference values ​​of all pixels in the pipeline area, the more dramatic the texture changes in the pipeline area, which may contain more edges or irregular structures and a more complex texture. The larger the spatial discontinuity, the more spatially discontinuous areas there are in the detection radar image. This may be due to the data anomalies of some pixels caused by asynchronous data collection, which in turn destroys the overall spatial morphology of the oil pipeline and affects the accurate judgment of features such as the pipeline direction and curvature. The smaller the spatial discontinuity, the more continuous the edges of the oil pipeline in the detection radar image.

[0141] The greater the spatial discontinuity, the higher the order of the spline interpolation algorithm can be increased, because when the shape of the oil pipeline in the detected radar image is discontinuous, the edges may appear jagged or some areas may be completely missing, making the spatial shape more complex. High-order spline interpolation can provide more degrees of freedom to fit such complex shapes. Low-order spline interpolation can only generate straight line segments to fill or connect data points, and it is difficult to accurately restore the shape of the oil pipeline with complex bending, breaking, and other conditions. Secondly, high-order spline interpolation can better fit these complex shape changes through more coefficients and control points, more accurately describe the bending degree and direction change of the curve, and effectively repair the damaged oil pipeline shape due to poor spatial continuity.

[0142] In some embodiments, based on the traversal search algorithm, each edge pixel point in the pipeline region of each frame of radar image is searched to obtain a plurality of edge chains composed of edge pixel points; the length of each edge chain is determined according to the number of edge pixel points contained in each edge chain; and the maximum length of the plurality of edge chains is taken as the length of the oil pipeline in each frame of radar image.

[0143] Based on the edge pixel points of the pipeline region, the system's search strategy and length calculation method can obtain quantitative data that accurately reflect the actual length of the pipeline.

[0144] Figure 8 is a flowchart of a method for determining the length of an oil pipeline in a radar image provided by an embodiment of the present specification. In specific implementation, the following steps are included:

[0145] S801: Based on the traversal search algorithm, each edge pixel point in the pipeline region of each frame of radar image is searched to obtain a plurality of edge chains composed of edge pixel points.

[0146] The traversal search algorithm can include breadth-first search (BFS) and depth-first search (DFS). The breadth-first search algorithm follows the search logic of “from near to far”, starting from the starting point, preferentially accessing the adjacent nodes closest to the starting point, and expanding the search range layer by layer until all reachable nodes are traversed. The depth-first search algorithm uses the strategy of “walking down one path to the end”, starting from the starting point, exploring as deeply as possible along one path, and when it cannot continue, backtracking to the last node and selecting other paths to continue searching. In the search of edge pixel points in the pipeline region of the radar image, both algorithms have their advantages and disadvantages. The breadth-first search algorithm can quickly and comprehensively traverse the edge pixel points when dealing with relatively simple pipeline structures without complex branches, ensuring that no part is missed. The depth-first search algorithm can efficiently explore along a specific path for pipelines with complex branches, intersections, and other structures, avoiding getting stuck in ineffective searches in complex structures.

[0147] The edge pixels of the pipeline region in each radar image can be identified based on an edge monitoring algorithm, which will not be described herein. Based on the identified edge pixels, a traversal search algorithm can be used for searching, so as to obtain a plurality of edge chains composed of edge pixels. In the searching process, a suitable starting point can be set for the algorithm. The selection of the starting point can be specifically based on the morphological features of the pipeline region, and any pixel point on the edge of the pipeline region can be selected as the starting point. Taking the breadth-first search algorithm as an example, the selected starting point is marked as an accessed state, and its adjacent unvisited edge pixels are added to a search queue. Each time a pixel point is taken out from the queue, the point is accessed and its adjacent unvisited edge pixels are added to the queue, and the process is repeated until the queue is empty. In this process, in order to distinguish different edge chains, each independent search path can be marked and recorded. When a marked edge pixel is searched and the pixel belongs to an existing edge chain, it indicates that the current search path intersects with the existing path, and the current search is stopped and a new search path is started, so as to obtain a plurality of independent edge chains.

[0148] S802: Determine the length of each edge chain according to the number of edge pixels contained in each edge chain.

[0149] After obtaining the edge chains, the length of each edge chain can be determined according to the number of edge pixels contained in each edge chain. When calculating the length, the scale information of the image can be considered. Since the radar image is a two-dimensional projection of the actual underground scene, there is a certain scale relationship between the pixel points in the image and the actual geographical space. For example, if the image scale is 1:100, that is, 1 pixel point in the image represents 0.01 meters in actual distance, and a certain edge chain contains 500 edge pixels, then the actual length of the edge chain is 500 x 0.01 = 5 meters. In the actual calculation process, the connection mode between the pixel points can also be considered. For adjacent pixel points, if they are horizontally or vertically connected, their distance is 1 pixel unit; if they are connected by a diagonal line, their distance is pixel units according to the Pythagorean theorem. By accurately calculating the actual distance between the pixel points in each edge chain and accumulating it, the accurate length of each edge chain can be obtained.

[0150] S803: Take the maximum length of the plurality of edge chains as the length of the oil pipeline in each radar image.

[0151] After obtaining the lengths of the plurality of edge chains, the maximum length thereof can be taken as the length of the oil pipeline in each radar image. Selecting the maximum length can highlight the main morphological features of the pipeline and avoid length calculation deviation caused by branches, local protrusions and the like.

[0152] In some embodiments, according to the missing redundancy ratio of each frame of radar image and the spatial discontinuity and completeness of the oil pipeline in each frame of radar image, the interpolation order corresponding to each frame of radar image is determined using the following formula:

[0153] ;

[0154] In the formula, k represents the interpolation order corresponding to the i-th frame of radar image. In the formula, k represents the interpolation order corresponding to the i-th frame of radar image.

[0155] In the formula, k represents the interpolation order corresponding to the i-th frame of radar image. In the formula, r represents the missing redundancy ratio of the i-th frame of radar image. The size of the missing redundancy ratio intuitively reflects the severity of the missing or redundancy problem existing in the data acquisition process. If the missing redundancy ratio is small, for example, less than 1, it means that the data redundancy is prominent, which may be due to the mismatch between the flight speed of the unmanned aerial vehicle and the radar acquisition frequency, resulting in repeated acquisition of the same area, or the existence of overlapping in the acquisition path planning, so that a large amount of repeated data appears in the image.

[0156] In the formula, k represents the interpolation order corresponding to the i-th frame of radar image. In the formula, s represents the spatial discontinuity of the oil pipeline in the i-th frame of radar image. The spatial discontinuity can reflect the texture difference of the pipeline area and the influence of the data acquisition quality on the pipeline form. The higher the spatial discontinuity value, the more serious the discontinuity of the pipeline in the image, such as fracture, misplacement, etc. Such discontinuity may be caused by physical damage to the actual pipeline, such as pipeline fracture caused by earthquakes, geological subsidence, or may be caused by visual discontinuity due to missing image data or improper processing.

[0157] In the formula, k represents the interpolation order corresponding to the i-th frame of radar image. In the formula, c represents the completeness of the oil pipeline in the i-th frame of radar image. The higher the completeness, the lower the order of spline interpolation, because the extracted oil pipeline features in the image have better integrity. This means that the influence of data acquisition synchronization on the features is smaller, and the data quality in the image is relatively high, and the data can better reflect the real shape, texture and other features of the oil pipeline.

[0158] In the formula, a represents a preset scaling coefficient. In the formula, f represents a normalization function. In the formula, f represents a floor function.

[0159] The following provides a specific embodiment of the present specification:

[0160] ​The ground penetrating radar is installed at the center of the bottom of the UAV, and the connection between the radar device and the UAV is ensured to be firm to avoid loosening or displacement during flight. The technical parameters of the ground penetrating radar are: dynamic range 192 dB (32 bit), maximum depth range 120 m (RDP=4), superposition number 32000 times / s, pulse repetition frequency 156 kHz, antenna 120 MHz, frequency range 20-140 MHz, longitudinal / lateral resolution 31 / 88 cm, and working temperature -20-+50℃. The UAV starts flight from the preset take-off point A according to the predetermined flight route and flight speed of 2 m / s, and the ground penetrating radar emits high-frequency electromagnetic pulses to the underground at the set pulse repetition frequency. During the propagation of the pulses in the underground, part of the pulses is reflected back when encountering the interface of different media such as oil pipelines, and is received by the receiving antenna of the radar. The collected data is formed into a radar image every certain distance, and a radar image is formed every 10 meters of flight of the UAV. The radar image B obtained at a certain moment is shown in FIG. 4. Figure 9 According to the missing redundancy ratio of the front and rear multiple radar images and the spatial discontinuity and length of the oil pipeline in the front and rear multiple radar images, the corresponding interpolation order is calculated and spline interpolation is performed. According to the spline interpolated radar image C, the oil pipeline is identified, as shown in FIG. 5. Figure 10

[0161] ​The method for identifying the oil pipeline provided in the embodiments of the present specification can obtain multiple radar images collected by a UAV; each pixel point is divided according to the relative difference of the gray value of each pixel point in each radar image, and multiple missing pixel points and multiple redundant pixel points are obtained; the missing-redundant ratio of each radar image is determined according to the multiple missing pixel points and the multiple redundant pixel points; the spatial discontinuity and length of the oil pipeline in each radar image are calculated according to the relative difference of the gray value of each pixel point in the pipeline region in each radar image; the interpolation order of the multiple radar images is calculated according to the missing-redundant ratio of the multiple radar images and the spatial discontinuity and length of the oil pipeline in the multiple radar images; the multiple missing pixel points and the multiple redundant pixel points in each radar image are subjected to spline interpolation according to the interpolation order of each radar image; and the oil pipeline is identified according to the multiple radar images subjected to spline interpolation. Compared with the prior art, the missing-redundant ratio of each radar image, the spatial discontinuity and length of the oil pipeline in each radar image can be determined according to the difference in the gray value of each pixel point in each radar image, and the influence of the synchronization of the flight speed of the UAV and the radar collection frequency on the missing and redundancy of the pixel points in the radar image, the continuity of the oil pipeline and the similar length of the oil pipeline is directly related. In addition, through joint analysis of multiple frames, the influence of noise or shielding in a single frame image on pipeline identification is reduced, the interpolation order is adaptively selected according to the severity of the data asynchronization, overfitting or underfitting is avoided, the smoothness and calculation efficiency of the interpolation result are improved, the stability of the interpolation is enhanced, and thus the accuracy and efficiency of the oil pipeline identification are greatly improved.

[0162] Based on the above-mentioned method for identifying the oil pipeline, the present specification further proposes an embodiment of a device for identifying the oil pipeline. As shown in the figure, the device for identifying the oil pipeline 1100 can specifically include the following modules: Figure 11

[0163] The acquisition module 1101 can be used to acquire multiple radar images collected by a UAV.

[0164] The division module 1102 can be used to divide each pixel point according to the relative difference of the gray value of each pixel point in each radar image, and obtain multiple missing pixel points and multiple redundant pixel points.

[0165] The determination module 1103 can be used to determine the missing-redundant ratio of each radar image according to the multiple missing pixel points and the multiple redundant pixel points.

[0166] The first calculation module 1104 can be used to calculate the spatial discontinuity and length of the oil pipeline in each radar image according to the relative difference of the gray value of each pixel point in the pipeline region in each radar image.

[0167] ​The second calculation module 1105 can be configured to calculate an interpolation order of the plurality of radar images according to the missing redundancy ratio of the plurality of radar images and the spatial discontinuity and length of the oil pipeline in the plurality of radar images.

[0168] The interpolation module 1106 can be configured to perform spline interpolation on the plurality of missing pixel points and the plurality of redundant pixel points of each radar image according to the interpolation order of each radar image.

[0169] The identification module 1107 can be configured to identify the oil pipeline according to the plurality of radar images after the spline interpolation.

[0170] In some embodiments, the division module 1102 can be specifically configured to:

[0171] determine a gray difference value of each pixel point of each radar image according to a relative difference between a gray value of each pixel point and a gray value of each edge pixel point in a preset spatial window centered on each pixel point;

[0172] obtain a histogram of the gray difference values of all pixel points of each radar image;

[0173] determine a first gray difference threshold according to a valley between two wave crests with the smallest gray difference values in the histogram;

[0174] determine a second gray difference threshold according to a valley between two wave crests with the largest gray difference values in the histogram;

[0175] divide the pixel points with the gray difference values less than the first gray difference threshold into missing pixel points;

[0176] divide the pixel points with the gray difference values greater than the second gray difference threshold into redundant pixel points.

[0177] In some embodiments, the first calculation module 1104 can be specifically configured to:

[0178] determine a gray difference value of each pixel point of each radar image according to a relative difference between a gray value of each pixel point and a gray value of each edge pixel point in a preset spatial window centered on each pixel point;

[0179] segment each radar image into a plurality of image regions based on the gray difference values of the pixel points in each radar image;

[0180] take an image region with the largest number of edge contours in the plurality of image regions as a pipeline region of each radar image;

[0181] extract a texture feature value of each pixel point in the pipeline region of each radar image using a texture feature extraction algorithm according to the gray difference value of each pixel point in the pipeline region of each radar image.

[0182] The spatial discontinuity of the oil pipeline in each radar image is calculated based on the missing redundancy ratio of each radar image frame and the texture feature value of each pixel in the pipeline area in each radar image frame.

[0183] Based on the traversal search algorithm, each edge pixel point of the pipeline area in each frame of radar image is searched to obtain the length of the oil pipeline in each frame of radar image.

[0184] In some embodiments, the first calculation module 1104 may further be used to:

[0185] Determine, according to a preset spatial window, a set of pixels centered around each pixel in the pipeline area;

[0186] Obtaining a texture difference value for each pixel in the pipeline area based on the sum of the differences between the texture feature value of each pixel in the corresponding pixel set;

[0187] Based on the missing redundancy ratio of each radar image frame and the texture difference value of each pixel in the pipeline area in each radar image frame, the spatial discontinuity of the oil pipeline in each radar image frame is calculated using the following formula:

[0188] ;

[0189] Where, Indicates the Spatial discontinuity of oil pipelines in frame radar images, Indicates the The missing redundancy ratio of the frame radar image, Indicates the The number of pixels in the pipeline area of ​​the frame radar image, Indicates the The sum of the texture difference values ​​of all pixels in the pipeline area in the frame radar image.

[0190] In some embodiments, the first calculation module 1104 may further be used to:

[0191] Based on the traversal search algorithm, the edge pixels of the pipeline area in each frame of radar image are searched to obtain multiple edge chains composed of edge pixels;

[0192] Determine the length of each edge chain according to the number of edge pixels contained in each edge chain;

[0193] The maximum length of multiple edge chains is taken as the length of the oil pipeline in each frame of radar image.

[0194] In some embodiments, the second calculation module 1105 may be specifically used to:

[0195] constructing a feature vector of the oil pipeline in each frame of the radar image according to the spatial discontinuity and the length of the oil pipeline in the frame of the radar image;

[0196] determining a set of radar images centered on each frame of the radar image according to a preset time window;

[0197] determining the completeness of the oil pipeline in each frame of the radar image according to the feature vector of the oil pipeline in the frame of the radar image and the similarity of the feature vectors of the oil pipelines in the set of radar images;

[0198] determining the interpolation order of each frame of the radar image according to the missing redundancy ratio of the frame of the radar image and the spatial discontinuity and the completeness of the oil pipeline in the frame of the radar image.

[0199] In some embodiments, the second calculation module 1105 can be further configured to:

[0200] determining the interpolation order corresponding to each frame of the radar image according to the missing redundancy ratio of the frame of the radar image and the spatial discontinuity and the completeness of the oil pipeline in the frame of the radar image, using the following formula:

[0201] ;

[0202] wherein, denotes the interpolation order corresponding to the i-th frame of the radar image, denotes the missing redundancy ratio of the i-th frame of the radar image, denotes the spatial discontinuity of the oil pipeline in the i-th frame of the radar image, denotes the completeness of the oil pipeline in the i-th frame of the radar image, denotes a preset scaling coefficient, denotes a normalization function, denotes a down-rounding function. In some embodiments, the identification module 1107 can be configured to: splicing the multiple frames of the radar image after the spline interpolation;

[0203] identifying the position and the direction of the oil pipeline according to the spliced radar image.

[0204]

[0205]

[0206] ​​​​The identification device for the oil pipeline provided in the embodiments of the present specification can obtain multiple frames of radar images collected by the unmanned aerial vehicle; each pixel point is divided according to the relative difference of the gray value of each pixel point in each frame of radar image, and multiple missing pixel points and multiple redundant pixel points are obtained; the missing-redundant ratio of each frame of radar image is determined according to the multiple missing pixel points and the multiple redundant pixel points; the spatial discontinuity and the length of the oil pipeline in each frame of radar image are calculated according to the relative difference of the gray value of each pixel point in the pipeline region in each frame of radar image; the interpolation order of the multiple frames of radar images is calculated according to the missing-redundant ratio of the multiple frames of radar images and the spatial discontinuity and the length of the oil pipeline in the multiple frames of radar images; the multiple missing pixel points and the multiple redundant pixel points in each frame of radar image are spline interpolated according to the interpolation order of each frame of radar image; and the oil pipeline is identified according to the multiple frames of radar images after spline interpolation. Compared with the prior art, the missing-redundant ratio of each frame of radar image, the spatial discontinuity and the length of the oil pipeline in each frame of radar image can be determined according to the difference of the pixel gray value in each frame of radar image, and the influence of the synchronization of the flight speed of the unmanned aerial vehicle and the radar collection frequency on the missing and redundancy of the pixel points in the radar image, the continuity of the oil pipeline and the similar length of the oil pipeline is directly related. In addition, through the joint analysis of multiple frames, the influence of noise or shielding in a single frame of image on pipeline identification is reduced, the interpolation order is adaptively selected according to the severity of the asynchronous data, overfitting or underfitting is avoided, the smoothness and calculation efficiency of the interpolation result are improved, the stability of the interpolation is enhanced, and thus the accuracy and efficiency of the oil pipeline identification are greatly improved.

[0207] It should be noted that the units, devices or modules and the like illustrated in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. For the convenience of description, the above device is described as various modules described respectively in terms of functions. Of course, in the implementation of the present specification, the functions of each module can be implemented in the same software and / or hardware, or the modules implementing the same function can be implemented by a combination of multiple sub-modules or sub-units, etc. The above-described device embodiments are only illustrative, for example, the division of the units is only a logical function division, and in actual implementation, another division mode can be used, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed ones can be indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0208] The embodiment of the present specification further provides a computer device for identifying a petroleum pipeline, comprising a processor and a memory for storing processor executable instructions, wherein the processor, when implemented, can execute the following steps according to the instructions: acquiring a plurality of radar images collected by a UAV; dividing each pixel point according to the relative difference of the gray value of each pixel point of each radar image to obtain a plurality of missing pixel points and a plurality of redundant pixel points; determining the missing redundancy ratio of each radar image according to the plurality of missing pixel points and the plurality of redundant pixel points; calculating the spatial discontinuity and length of the petroleum pipeline in each radar image according to the relative difference of the gray value of each pixel point in the pipeline region of each radar image; calculating the interpolation order of the plurality of radar images according to the missing redundancy ratio of the plurality of radar images and the spatial discontinuity and length of the petroleum pipeline in the plurality of radar images; performing spline interpolation on the plurality of missing pixel points and the plurality of redundant pixel points of each radar image according to the interpolation order of each radar image; and identifying the petroleum pipeline according to the plurality of radar images after spline interpolation.

[0209] In order to be able to more accurately complete the above instructions, referring to Figure 12 The embodiment of the present specification further provides another specific computer device 1200, wherein the computer device 1200 comprises a network communication port 1201, a processor 1202 and a memory 1203, and the above structures are connected by internal cables so that each structure can perform specific data interaction.

[0210] The processor 1202 can be specifically used for acquiring a plurality of radar images collected by a UAV; dividing each pixel point according to the relative difference of the gray value of each pixel point of each radar image to obtain a plurality of missing pixel points and a plurality of redundant pixel points; determining the missing redundancy ratio of each radar image according to the plurality of missing pixel points and the plurality of redundant pixel points; calculating the spatial discontinuity and length of the petroleum pipeline in each radar image according to the relative difference of the gray value of each pixel point in the pipeline region of each radar image; calculating the interpolation order of the plurality of radar images according to the missing redundancy ratio of the plurality of radar images and the spatial discontinuity and length of the petroleum pipeline in the plurality of radar images; performing spline interpolation on the plurality of missing pixel points and the plurality of redundant pixel points of each radar image according to the interpolation order of each radar image; and identifying the petroleum pipeline according to the plurality of radar images after spline interpolation.

[0211] The memory 1203 can be specifically used for storing corresponding instruction programs.

[0212] In the embodiment, the network communication port 1201 can be a virtual port bound with different communication protocols, so as to send or receive different data. For example, the network communication port can be a port responsible for web data communication, a port responsible for FTP data communication, or a port responsible for mail data communication. In addition, the network communication port can also be an actual communication interface or a communication chip. For example, it can be a wireless mobile network communication chip such as GSM, CDMA, etc.; it can also be a Wifi chip; it can also be a Bluetooth chip.

[0213] In the embodiment, the processor 1202 can be implemented in any appropriate manner. For example, the processor can take the form of a microprocessor or processor and a computer readable medium storing computer readable program code (e.g., software or firmware) executable by the microprocessor, logic gates, switches, an Application Specific Integrated Circuit (ASIC), a programmable logic controller, and an embedded microcontroller, etc. The present specification does not make any limitation.

[0214] In the embodiment, the memory 1203 includes a volatile memory and a non-volatile memory. The memory 1203 can include multiple levels, and in a digital system, as long as it can save binary data, it can be a memory; in an integrated circuit, a circuit without a physical form having a storage function is also called a memory, such as RAM, FIFO, etc.; in a system, a storage device with a physical form is also called a memory, such as a memory stick, a TF card, etc.

[0215] The embodiments of the present specification also provide a computer program product, including at least one instruction or at least one program, which is loaded and executed by a processor to implement the method as shown in Figure 1 The embodiments of the present specification also provide a computer program product, including at least one instruction or at least one program, which is loaded and executed by a processor to implement the method as shown in

[0216] It should be understood that in various embodiments of the present specification, the size of the sequence number of each process described above does not mean the order of execution, and the execution order of each process should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present specification.

[0217] It should also be understood that in the embodiments of the present specification, the term "and / or" only describes the association relationship of the associated objects, and means that there can be three relationships. For example, A and / or B can mean that A exists alone, A and B exist together, and B exists alone. In addition, the character " / " in the present specification generally means that the front and rear associated objects are in an "or" relationship.

[0218] Those skilled in the art will appreciate that embodiments of the present application can be readily used as software, hardware, or a combination of software and hardware. In a software embodiment, the present application can be implemented with computer programs (also referred to as software instructions, software code, computer code, and the like) that execute on programmable hardware including computer processors, digital signal processors, microprocessors, central processing units, microcontrollers, programmable hardware logic devices, and the like. Generally, the present application can be implemented in hardware, software, or any combination of hardware and software.

[0219] The present application is described in relation to flow diagrams and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the present application. It is to be understood that each block of the flow diagrams and / or block diagrams, and combinations of blocks in the flow diagrams and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processing device or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flow diagrams and / or block diagrams block or blocks. Figure 1 one or more functions specified by one or more blocks Figure 1 one or more functions specified by one or more blocks

[0220] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the function specified in the flow diagrams and / or block diagrams block or blocks. Figure 1 one or more functions specified by one or more blocks Figure 1 one or more functions specified by one or more blocks

[0221] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational tasks to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions that are executed on the computer or other programmable apparatus provide tasks that implement the functions specified in the flow diagrams and / or block diagrams block or blocks. Figure 1 one or more functions specified by one or more blocks Figure 1 one or more functions specified by one or more blocks

[0222] The above description is only specific embodiments of the present application, and is not intended to limit the scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the scope of the present application.

Claims

1. A method of identifying a petroleum pipeline, characterized by, The method comprises: acquiring a plurality of radar images collected by a UAV; dividing each pixel point according to the relative difference of the gray value of each pixel point of each radar image to obtain a plurality of missing pixel points and a plurality of redundant pixel points; determining the missing-redundant ratio of each radar image according to the plurality of missing pixel points and the plurality of redundant pixel points; calculating the spatial discontinuity and length of the oil pipeline in each radar image according to the relative difference of the gray value of each pixel point in the pipeline region of each radar image; The interpolation order of the multi-frame radar images is calculated based on the missing redundancy ratio of the multi-frame radar images and the spatial discontinuity and length of the oil pipeline in the multi-frame radar images, which includes: constructing a feature vector of the oil pipeline in each frame radar image based on the spatial discontinuity and length of the oil pipeline in each frame radar image; determining a radar image set centered on each frame radar image based on a preset time window; determining the integrity of the oil pipeline in each frame radar image based on the similarity between the feature vector of the oil pipeline in each frame radar image and the feature vector of the oil pipeline in each radar image in the radar image set; and determining the interpolation order corresponding to each frame radar image based on the missing redundancy ratio of each frame radar image and the spatial discontinuity and integrity of the oil pipeline in each frame radar image using the following formula: Where, Indicates the The interpolation order corresponding to the frame radar image, Indicates the The missing redundancy ratio of the frame radar image, Indicates the Spatial discontinuity of oil pipelines in frame radar images, Indicates the The integrity of oil pipelines in frame radar images, Indicates the preset zoom factor, represents the normalization function, represents the floor function; spline interpolation is performed on the plurality of missing pixel points and the plurality of redundant pixel points of each radar image according to the interpolation order of each radar image; identifying the oil pipeline according to the plurality of radar images after spline interpolation.

2. The method of claim 1, wherein, The method comprises: determining the gray difference value of each pixel point of each radar image according to the relative difference of the gray value of each pixel point of each radar image and the gray value of each edge pixel point in a preset spatial window centered on each pixel point; acquiring a histogram of the gray difference values of all pixel points of each radar image; determining a first gray difference threshold value according to the trough between the two wave crests with the smallest gray difference values in the histogram; determining a second gray difference threshold value according to the trough between the two wave crests with the largest gray difference values in the histogram; dividing the pixel points with a gray difference value less than the first gray difference threshold value into missing pixel points; dividing the pixel points with a gray difference value greater than the second gray difference threshold value into redundant pixel points.

3. The method of claim 1, wherein, The method comprises: determining the gray difference value of each pixel point of each radar image according to the relative difference of the gray value of each pixel point of each radar image and the gray value of each edge pixel point in a preset spatial window centered on each pixel point; segmenting each radar image into a plurality of image regions based on the gray difference values of the pixel points in each radar image; taking the image region with the most edge contours in the plurality of image regions as the pipeline region of each radar image; extracting the texture feature values of the pixel points in the pipeline region of each radar image using a texture feature extraction algorithm according to the gray difference values of the pixel points in the pipeline region of each radar image; calculating the spatial discontinuity of the oil pipeline in each radar image according to the missing-redundant ratio of each radar image and the texture feature values of the pixel points in the pipeline region of each radar image; searching for each edge pixel point in the pipeline region of each radar image based on a traversal search algorithm to obtain the length of the oil pipeline in each radar image.

4. The method of claim 3, wherein, The method comprises: determining a pixel point set centered on each pixel point in the pipeline region according to a preset spatial window; According to the texture feature value of each pixel point in the pipeline region and the sum of the texture feature values of the pixel points in the corresponding pixel point set, a texture difference value of each pixel point in the pipeline region is obtained; According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: ； wherein represents the first spatial discontinuity of the oil pipeline in the frame radar image, represents the first missing redundancy ratio of the frame radar image, represents the first pixel point number of the pipeline region in the frame radar image, represents the first sum of texture difference values of all pixel points of the pipeline region in the frame radar image.

5. The method of claim 3, wherein, According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula:

6. The method of claim 1, wherein, According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula:

7. An apparatus for identifying a petroleum pipeline, characterized by According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: The second calculation module is configured to calculate an interpolation order of the multiple radar image frames based on the missing redundancy ratio of the multiple radar image frames and the spatial discontinuity and length of the oil pipeline in the multiple radar image frames, including: constructing an oil pipeline feature vector for each radar image frame based on the spatial discontinuity and length of the oil pipeline in each radar image frame; determining a radar image set centered on each radar image frame based on a preset time window; determining the integrity of the oil pipeline in each radar image frame based on the similarity between the oil pipeline feature vector of each radar image frame and the oil pipeline feature vectors of each radar image in the radar image set; and determining the interpolation order corresponding to each radar image frame based on the missing redundancy ratio of each radar image frame and the spatial discontinuity and integrity of the oil pipeline in each radar image frame using the following formula: Where, Indicates the The interpolation order corresponding to the frame radar image, Indicates the The missing redundancy ratio of the frame radar image, Indicates the Spatial discontinuity of oil pipelines in frame radar images, Indicates the The integrity of oil pipelines in frame radar images, Indicates the preset zoom factor, represents the normalization function, represents the floor function; According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula:

8. A computer device, comprising: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil pipeline in each frame of radar image is calculated using the following formula: According to the missing redundancy ratio of each frame of radar image and the texture difference value of each pixel point in the pipeline region of each frame of radar image, the spatial discontinuity of the oil

Images