Remote sensing image data-based mountain area power transmission corridor damage automatic detection method

Abstract

The invention discloses a mountain area power transmission corridor damage automatic detection method based on remote sensing image data, and relates to the field of power transmission corridor damage detection. The invention aims to solve the problem of low detection accuracy of the existing damage detection method for the power transmission corridor in the mountainous area. The method comprises the following steps: obtaining a power transmission line position data list DataList one by one based on a remote sensing image of a mountainous area power transmission corridor area and a list of positions of power transmission lines in the mountainous area power transmission corridor area; establishing a power transmission corridor convergence undirected attribute extraction module OptNoRot, wherein the output of the OptNoRot is a power transmission corridor convergence undirected vector; processing all contents of the DataList by utilizing OptNoRot to obtain a power transmission corridor convergence undirected attribute of each table item of the DataList; a neighborhood convergence undirected attribute abnormal condition detection module OptYC is established, the input of the OptYC is a to-be-analyzed neighborhood list and a vector of a to-be-detected point, and the output of the OptYC is a neighborhood convergence undirected attribute abnormal condition detection result; and detecting the condition that the power transmission corridor is damaged by utilizing OptYC and DataList. The method is used for detecting the power transmission corridor damage condition.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of power transmission corridor damage detection, and particularly relates to a method for automatically detecting damaged power transmission corridors in mountainous areas based on remote sensing image data. BACKGROUND

[0002] As the core transmission carrier of the power system, power transmission lines face severe operation and maintenance challenges in complex mountainous environments. Due to the steep terrain, dense vegetation, and frequent geological disasters in mountainous areas, power transmission corridors are vulnerable to damage factors such as landslides, tree collapse, and illegal construction. The traditional manual inspection mode has low efficiency, long cycle, poor accessibility in high-risk areas, and other problems, which cannot meet the demand of modern smart grid management. Using remote sensing images to automatically detect damaged power transmission corridors in mountainous areas can improve the timeliness of power facility risk warning and reduce maintenance costs. Therefore, developing automatic detection of damaged power transmission corridors in mountainous areas based on remote sensing image data has become a research focus in the field.

[0003] To detect damaged power transmission corridors in mountainous areas based on remote sensing images, the current industry generally uses the following technical means: 1. Manual inspection of high-resolution remote sensing images to find possible damaged points through visual interpretation. This method is relatively reliable, but manual detection is time-consuming and labor-intensive, and is affected by factors such as employee holidays. This method is high in labor cost and difficult to ensure timeliness. 2. Automatic method based on sample library supervised learning method, which trains CNN, Yolo or UNet neural network model by constructing a limited number of damaged or undamaged power transmission corridor samples, and then uses these models to automatically detect in mountainous remote sensing images. However, this method has obvious technical limitations: on the one hand, the geographical environment of mountainous areas is highly heterogeneous, with significant differences in vegetation coverage, geological structure, and weather conditions in different regions. On the other hand, the damage types are diverse, including but not limited to vegetation encroachment, tower tilt, and ground subsidence. The current sample library-based supervised learning method is difficult to cover all potential damage patterns, and when encountering abnormal situations outside the labeled samples, the miss rate is very high. Moreover, the process of manually labeling samples is prone to subjective bias, which significantly reduces the model's generalization ability, resulting in low accuracy of damaged power transmission corridor detection in mountainous areas. SUMMARY

[0004] The present application aims to solve the problem of low detection accuracy in existing damaged power transmission corridor detection methods in mountainous areas, and proposes a method for automatically detecting damaged power transmission corridors in mountainous areas based on remote sensing image data.

[0005] The method for automatically detecting damaged power transmission corridors in mountainous areas based on remote sensing image data specifically comprises:

[0006] S1, obtaining a list of position data of each power transmission line DataList based on a remote sensing image RSImg of a mountainous power transmission corridor region and a list PointList of positions of power transmission lines in the mountainous power transmission corridor region;

[0007] S2, establishing a power transmission corridor convergence undirected attribute extraction module OptNoRot, input of the OptNoRot being image data to be processed OptNoRotInput, and output of the OptNoRot being a power transmission corridor convergence undirected attribute vector OptOutput;

[0008] S3, processing all contents of the DataList obtained in S1 by using the OptNoRot obtained in S2 to obtain a power transmission corridor convergence undirected attribute of each entry of the DataList;

[0009] S4, establishing a neighborhood convergence undirected attribute abnormality detection module OptYC, input of the OptYC being a neighborhood list to be analyzed OptYCInput and a vector of a point to be detected OptYCTestVector, and output of the OptYC being a neighborhood convergence undirected attribute abnormality detection result OptYCResult;

[0010] S5, detecting a damaged condition of the power transmission corridor by using the OptYC and the DataList.

[0011] Further, in the S1, the remote sensing image RSImg of the mountainous power transmission corridor region and the list PointList of positions of power transmission lines in the mountainous power transmission corridor region are used to obtain the list of position data of each power transmission line DataList, and specifically:

[0012] S101, inputting the remote sensing image RSImg of the mountainous power transmission corridor region;

[0013] S102, inputting the list PointList of positions of power transmission lines in the mountainous power transmission corridor region;

[0014] The PointList is a list of position point coordinates of the power transmission lines on the RSImg;

[0015] The PointList includes the following fields: an X-axis coordinate PosX of a position point of the power transmission line on the RSImg, and a Y-axis coordinate of the position point of the power transmission line on the RSImg;

[0016] S103, inputting a preset width ZLKD of the power transmission corridor and a preset neighborhood threshold ZLYZ of the damaged power transmission corridor;

[0017] A default value of the ZLKD is 5;

[0018] A default value of the ZLYZ is 10;

[0019] S104, obtain the spatial resolution RSFBL of RSImg, the number of points to be analyzed ZLDFX and the basic length to be analyzed ZLCD:

[0020] The spatial resolution RSFBL of RSImg is the spatial resolution in the remote sensing image attribute information of RSImg.

[0021] The number of points to be analyzed ZLDFX is the number of elements in PointList.

[0022] The basic length to be analyzed ZLCD = Round (ZLKD / RSFBL).

[0023] Wherein, Round() is the rounding operation.

[0024] S105, determine whether ZLCD is even, if ZLCD is even, then ZLCD = ZLCD + 1, and then execute S106; otherwise, directly execute S106.

[0025] S106, initialize the position data list of each power transmission line DataList = empty list.

[0026] DataList contains the following fields: image information DLImage corresponding to the power transmission line position point, X-axis coordinate DLPosX of the center point of the image corresponding to the power transmission line position point on RSImg, Y-axis coordinate DLPosY of the center point of the image corresponding to the power transmission line position point on RSImg, and power transmission corridor convergence undirected attribute DLFeature of the power transmission line position point.

[0027] S107, initialize process counter JLCounter = 1.

[0028] S108, initialize process first temporary variable CSHTemp1 = JLCounterth element of PointList.

[0029] S109, obtain the second temporary variable CSHTemp2 of the initialization process:

[0030] RSImg is intercepted as CSHTemp2 with CSHTemp1's PosX and PosY as the center and the side length of 5×ZLCD pixels.

[0031] S110, add a new element to DataList.

[0032] The DLImage of the new element = CSHTemp2.

[0033] DLPosX of the new element = PosX of CSHTemp1;

[0034] DLPosY of the new element = PosY of CSHTemp1;

[0035] DLFeature of the new element = a 20-dimensional vector with all values being 0;

[0036] S111, let JLCounter = JLCounter + 1;

[0037] S112, if JLCounter is less than or equal to ZLDFX, go to S108, otherwise go to S113;

[0038] S113, sort all elements in DataList according to DLPosX value from small to large;

[0039] S114, let initialization process sorting counter JLPXCounter = 2;

[0040] S115, get initialization process sorting first temporary variable JLPXTemp1:

[0041] a, get the transmission line position point corresponding image of the JLPXCounter-th to ZLDFX-th element of DataList, and get the center point of each image;

[0042] b, find the center point closest to the center point of the transmission line position point corresponding image of the JLPXCounter-1-th element among the center points obtained in a, and take the current center point position number as the initialization process sorting first temporary variable JLPXTemp1;

[0043] S116, exchange the position of the JLPXTemp1-th element and the JLPXCounter-th element of DataList;

[0044] S117, let JLPXCounter = JLPXCounter + 1;

[0045] S118, if JLPXCounter is less than or equal to ZLDFX-1, go to S115, otherwise output DataList.

[0046] Further, the OptNoRot in S2 is an extraction module of power transmission corridor convergent undirected attribute, and the input of OptNoRot is image data to be processed OptNoRotInput, and the output of OptNoRot is power transmission corridor convergent undirected attribute vector OptOutput, and the specific process is as follows:

[0047] S201, establish a power transmission corridor convergence undirected attribute extraction module OptNoRot, OptNoRot input is the image data to be processed OptNoRotInput;

[0048] S202, cut OptNoRotInput into 5 horizontal and 5 vertical equal-sized image blocks, a total of 25 equal-sized blocks, arrange the blocks from left to right and then from top to bottom to obtain a block image array OptNoRotArray;

[0049] wherein each image block in OptNoRotArray has a size of ZLCD rows x ZLCD columns;

[0050] S203, establish a sub-block attribute list OptNoRotFeatureList = empty list;

[0051] S204, set a sub-block calculation counter OptNoRotCounter = 1;

[0052] S205, a sub-block first temporary variable OptNoRotTemp1 = the OptNoRotCounterth element of OptNoRotArray;

[0053] S206, convert OptNoRotTemp1 to a grayscale image and perform a two-dimensional fast Fourier transform to obtain OptFFT;

[0054] S207, move the zero-frequency component of OptFFT to the center position to obtain a centered spectrum OptShift, and use OptShift to calculate the power transmission corridor amplitude spectrum OptMag;

[0055] S208, add OptMag to OptNoRotFeatureList;

[0056] S209, set OptNoRotCounter = OptNoRotCounter + 1;

[0057] S210, if OptNoRotCounter is less than or equal to 25, go to S205, otherwise go to S211;

[0058] S211, for all elements in OptNoRotFeatureList, subtract the 13th element from the current element, and store the result in the current element position;

[0059] the current element is other than the 13th element;

[0060] S212, establish the convergent undirected attribute analysis list OptAnaList = empty list;

[0061] S213, establish the convergent undirected attribute analysis counter OptAnaCounter = 1;

[0062] Establish the convergent undirected attribute analysis minimum fluctuation vector OptAnaVector;

[0063] The OptAnaVector is a vector of ZLCD elements, and all the internal elements of the OptAnaVector are 0;

[0064] Establish the convergent undirected attribute analysis minimum fluctuation value OptAnaValue = 100000;

[0065] S214, obtain the convergent undirected attribute analysis candidate list OptAnaBX;

[0066] The OptAnaBX is a list of integers, and the content of the OptAnaBX is determined according to the value of the OptAnaCounter;

[0067] S215, obtain the convergent undirected attribute first temporary variable OptAnaTemp1:

[0068] First, determine the content of the OptAnaBX according to the OptAnaCounter;

[0069] Then, take the corresponding entry in the OptNoRotFeatureList according to the content of the OptAnaBX;

[0070] Finally, add the corresponding elements of each entry and divide by 4 to obtain the convergent undirected attribute first temporary variable OptAnaTemp1;

[0071] Finally, add the corresponding elements of each entry and divide by 4 to obtain the convergent undirected attribute first temporary variable OptAnaTemp1;

[0072] S216, obtain the convergent undirected attribute second temporary variable OptAnaTemp2:

[0073] Calculate the mean of OptAnaTemp1 according to the column to form a vector of ZLCD elements as the convergent undirected attribute second temporary variable OptAnaTemp2;

[0074] S217, the convergent undirected attribute third temporary variable OptAnaTemp3 = the standard deviation of OptAnaTemp2;

[0075] S218, if OptAnaTemp3 is less than OptAnaValue, go to S219, otherwise go to S221;

[0076] S219, set OptAnaTemp2 = tanh(ABS(OptAnaTemp2) / Max(ABS(OptAnaTemp2)));

[0077] where Max is to obtain the maximum value in a vector, and tanh is the hyperbolic tangent function;

[0078] S220, set OptAnaVector = OptAnaTemp2;

[0079] set OptAnaValue = OptAnaTemp3;

[0080] S221, set OptAnaCounter = OptAnaCounter + 1;

[0081] S222, if OptAnaCounter is less than or equal to 8, go to S214, otherwise go to S223;

[0082] S223, establish the power transmission corridor convergence undirected eigenvector OptOutput = OptAnaVector;

[0083] S224, return OptOutput as the result of OptNoRot.

[0084] Further, the calculation of the power transmission corridor amplitude spectrum OptMag using OptShift in S207 is specifically:

[0085] OptMag = log(abs(OptShift) + 0.5) + 2.5

[0086] where abs is to calculate the absolute value, and log is to calculate the natural logarithm.

[0087] Further, the content of OptAnaBX in S214 is determined according to the value of OptAnaCounter, specifically:

[0088] If OptAnaCounter is 1, the content of OptAnaBX is 1, 7, 19, 25;

[0089] If OptAnaCounter is 2, the content of OptAnaBX is 2, 8, 18, 24;

[0090] If OptAnaCounter is 3, then the content of OptAnaBX is 3, 8, 18, 23;

[0091] If OptAnaCounter is 4, then the content of OptAnaBX is 4, 8, 18, 22;

[0092] If OptAnaCounter is 5, then the content of OptAnaBX is 5, 9, 17, 21;

[0093] If OptAnaCounter is 6, then the content of OptAnaBX is 6, 12, 14, 20;

[0094] If OptAnaCounter is 7, then the content of OptAnaBX is 11, 12, 14, 15;

[0095] If OptAnaCounter is 8, then the content of OptAnaBX is 16, 12, 14, 10.

[0096] Further, the S3 uses the OptNoRot obtained from S2 to process all the content of DataList obtained from S1, and obtains the power corridor convergent undirected attribute of each table entry of DataList, specifically:

[0097] S301, let ProcessCounter = 1;

[0098] S302, let ProcessTemp1 = the DLImage of the ProcessCounterth element of DataList;

[0099] S303, let ProcessTemp2 = the output OptOutput of OptNoRot;

[0100] wherein the input OptNoRotInput of OptNoRot = ProcessTemp1;

[0101] S304, set the value of the DLFeature of the ProcessCounterth element of DataList to ProcessTemp1;

[0102] S305, let ProcessCounter = ProcessCounter + 1;

[0103] S306, if ProcessCounter is less than or equal to ZLDFX, go to S302, otherwise output OptOutput as the power transmission corridor convergence undirected attribute DLFeature of each table entry of DataList.

[0104] Further, the S4 includes an abnormal case detection module OptYC for establishing neighborhood convergence undirected attributes, the input of OptYC is a neighborhood list to be analyzed OptYCInput and a vector of a point to be detected OptYCTestVector, the output of OptYC is an abnormal case detection result of neighborhood convergence undirected attributes OptYCResult, and the specific process is as follows:

[0105] S401, establish an abnormal case detection module OptYC for establishing neighborhood convergence undirected attributes, the input of OptYC is a neighborhood list to be analyzed OptYCInput and a vector of a point to be detected OptYCTestVector;

[0106] S402, set the abnormal case detection result of neighborhood convergence undirected attributes OptYCResult = 0;

[0107] S403, establish a detection center vector list OptYCTestCenter;

[0108] OptYCTestCenter contains 4 elements, the value of the first element is equal to OptYCTestVector, the value of the second element is equal to the value of the first element of OptYCInput, the value of the third element is equal to the value of the ZLYZth element of OptYCInput, and the value of the fourth element is equal to the value of the ZLYZ+1th element of OptYCInput;

[0109] S404, based on the 4 centers of OptYCTestCenter, perform K-Means clustering on all elements of OptYCInput to obtain the number of each clustered element after clustering OptYCAfterNumber;

[0110] S405, if the value of the first element of OptYCAfterNumber is greater than or equal to 2, go to S407, otherwise go to S406;

[0111] S406, set OptYCResult = 1, and then execute S407;

[0112] S407, return OptYCResult as the result of OptYC.

[0113] Further, the S5 detects the damaged situation of the power transmission corridor by using OptYC and DataList, specifically:

[0114] S501, set the damaged counter of the power transmission corridor JCPHCounter = ZLYZ + 1;

[0115] S502, get the first temporary variable of the damaged power transmission corridor JCPHTemp1;

[0116] S503, take out the values of the DLFeature corresponding to the JCPHCounter-ZLYZ to JCPHCounter-1 elements in the DataList to form a list, and store them in the first to ZLYZ elements of JCPHTemp1;

[0117] S504, take out the values of the DLFeature corresponding to the JCPHCounter+1 to JCPHCounter+ZLYZ elements in the DataList to form a list, and store them in the ZLYZ+1 to 2×ZLYZ elements of JCPHTemp1;

[0118] S505, the second temporary variable of the damaged power transmission corridor JCPHTemp2 = the result of OptYC OptYCResult;

[0119] The input of the OptYC to be analyzed neighborhood list OptYCInput = JCPHTemp1;

[0120] The input of the OptYC to be detected point vector OptYCTestVector = the value of the DLFeature of the JCPHCounter element in the DataList;

[0121] S506, if JCPHTemp2 is equal to 0, go to S510; otherwise, go to S507;

[0122] S507, the third temporary variable of the damaged power transmission corridor JCPHTemp3 = the value of the DLPosX of the JCPHCounter element in the DataList;

[0123] S508, the fourth temporary variable of the damaged power transmission corridor JCPHTemp4 = the value of the DLPosY of the JCPHCounter element in the DataList;

[0124] S509, obtain the damaged detection result of the power transmission corridor by using JCPHTemp3 and JCPHTemp4;

[0125] S510, set JCPHCounter = JCPHCounter + 1;

[0126] S511, if JCPHCounter is less than or equal to ZLDFX-ZLYZ, proceed to S502; otherwise, output the result of the power transmission corridor damage detection.

[0127] Furthermore, the first temporary variable JCPHTemp1 for detecting the damage to the power transmission corridor in S502 is a list containing 2×ZLYZ elements.

[0128] Furthermore, the detection result of the damage to the power transmission corridor in S509 is: damage to the power transmission corridor is detected at the position where the X coordinate of RSImg is JCPHTemp3 and the Y coordinate is JCPHTemp4.

[0129] The beneficial effects of this invention are as follows:

[0130] This invention provides an automatic detection method for damaged power transmission corridors in mountainous areas based on remote sensing image data. The method utilizes a power transmission corridor convergent undirected attribute extraction module to find typical features of power transmission corridors within image blocks. Then, a neighborhood convergent undirected attribute anomaly detection module is used to determine whether there are obvious anomalies between specific locations in the remote sensing image and their neighborhoods. This allows for automatic detection of damage to power transmission corridors in mountainous areas, improving the accuracy of damage detection. The neighborhood convergent undirected attribute anomaly detection module of this invention can automatically determine the damage status of power transmission corridors within a region of a remote sensing image without relying on samples or a training process. Furthermore, the convergent undirected attribute extraction makes the method adaptive to input data. This invention does not have specific requirements regarding the direction, material, or edge material of the power transmission corridor in the remote sensing image, making it adaptable to a wider range of damage situations in mountainous power transmission corridors. This significantly improves the timeliness and accuracy of power facility risk warnings based on remote sensing image data. Attached Figure Description

[0131] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0132] Specific implementation method one: as follows Figure 1 As shown, the specific process of the automatic detection method for damage to mountain power transmission corridors based on remote sensing image data in this embodiment is as follows:

[0133] S1, based on the high-resolution remote sensing image RSImg of the mountain power transmission corridor area and the list PointList of the locations of power transmission lines in the mountain power transmission corridor area, obtain a data list DataList for each power transmission line location, specifically:

[0134] S101, input a high-resolution remote sensing image RSImg of a mountainous power transmission corridor region;

[0135] S102, input a list PointList of positions of power transmission lines in the mountainous power transmission corridor region, the PointList sequentially describes a list of position points of the power transmission lines on the RSImg;

[0136] The PointList includes the following fields: X-axis coordinate PosX of the position point of the power transmission line on the RSImg, Y-axis coordinate of the position point of the power transmission line on the RSImg;

[0137] Wherein, the origin of the coordinate axis is the upper left corner of the image, the X-axis direction is the left-right direction of the image, and the Y direction is the vertical direction of the image;

[0138] S103, input a preset width ZLKD of the power transmission corridor, the default value of ZLKD is 5 meters; input a preset neighborhood threshold ZLYZ of the damaged power transmission corridor, the default value of ZLYZ is 10;

[0139] S104, obtain the spatial resolution RSFBL of the RSImg, the number ZLDFX of points to be analyzed, and the basic length ZLCD to be analyzed, specifically:

[0140] The spatial resolution RSFBL of the RSImg is the spatial resolution in the remote sensing image attribute information of the RSImg;

[0141] The number ZLDFX of points to be analyzed is the number of elements in the PointList;

[0142] The basic length ZLCD to be analyzed is Round(ZLKD / RSFBL);

[0143] Wherein, Round() is a rounding operation;

[0144] S105, determine whether ZLCD is even, if ZLCD is even, then ZLCD=ZLCD+1, and then execute S106; otherwise, directly execute S106;

[0145] S106, initialize the data list DataList of each power transmission line position = an empty list;

[0146] The DataList includes the following fields: image information DLImage corresponding to the position point of the power transmission line, X-axis coordinate DLPosX of the center point of the image corresponding to the position point of the power transmission line on the RSImg, Y-axis coordinate DLPosY of the center point of the image corresponding to the position point of the power transmission line on the RSImg, and power transmission line position point power transmission corridor convergence undirected attribute DLFeature;

[0147] S107, initialize a process counter JLCounter = 1;

[0148] S108, initialize a first temporary variable CSHTemp1 = the JLCounterth element of PointList;

[0149] S109, obtain a second temporary variable CSHTemp2;

[0150] RSImg intercepts an image of 5 x ZLCD pixels in length and width centered at the PosX and PosY of CSHTemp1 as CSHTemp2;

[0151] S110, add a new element to DataList;

[0152] The added new element has DLImage = CSHTemp2, DLPosX = the PosX of CSHTemp1, DLPosY = the PosY of CSHTemp1, and DLFeature = a 20-dimensional vector with all values being 0;

[0153] S111, JLCounter = JLCounter + 1;

[0154] S112, if JLCounter is less than or equal to ZLDFX, go to S108, otherwise go to S113;

[0155] S113, sort all elements in DataList according to the DLPosX value from small to large;

[0156] S114, initialize a sorting counter JLPXCounter = 2;

[0157] S115, obtain a first temporary variable JLPXTemp1;

[0158] a, obtain the transmission line position point corresponding images of the JLPXCounterth to ZLDFXth elements of DataList, and obtain the center points of each image;

[0159] b, find the center point closest to the center point of the transmission line position point corresponding image of the JLPXCounter-1th element among the center points obtained in a, and take the current center point position number as the first temporary variable JLPXTemp1 of the initialization process sorting;

[0160] S116, exchange the positions of the JLPXTemp1th element and the JLPXCounterth element of DataList;

[0161] S117, let JLPXCounter = JLPXCounter + 1;

[0162] S118, if JLPXCounter is less than or equal to ZLDFX-1, then go to S115, otherwise output DataList.

[0163] S2, establish a power transmission corridor convergence undirected attribute extraction module OptNoRot, the input of OptNoRot is the image data to be processed OptNoRotInput, and the output is the power transmission corridor convergence undirected attribute vector OptOutput, which is specifically:

[0164] S201, establish a power transmission corridor convergence undirected attribute extraction module OptNoRot, the input of OptNoRot is the image data to be processed OptNoRotInput;

[0165] S202, cut OptNoRotInput into 5 horizontal and 5 vertical equal-sized image blocks, a total of 25 equal-sized subblocks, which are arranged in an order of first from left to right and then from top to bottom to obtain a subblock image array OptNoRotArray;

[0166] wherein the size of each image block in OptNoRotArray is ZLCD rows ZLCD columns;

[0167] S203, establish a subblock attribute list OptNoRotFeatureList = an empty list;

[0168] S204, let the subblock calculation counter OptNoRotCounter = 1;

[0169] S205, the first temporary variable of the subblock OptNoRotTemp1 = the OptNoRotCounter-th element of OptNoRotArray;

[0170] S206, convert OptNoRotTemp1 into a gray-scale image and perform two-dimensional fast Fourier transform to obtain OptFFT;

[0171] S207, shift the zero-frequency component of OptFFT to the center position to obtain the centered frequency spectrum OptShift, and calculate the power transmission corridor amplitude spectrum OptMag:

[0172] OptMag = log(abs(OptShift) + 0.5) + 2.5

[0173] wherein abs is the absolute value, and log is the natural logarithm.

[0174] S208, add OptMag to OptNoRotFeatureList;

[0175] S209, OptNoRotCounter = OptNoRotCounter + 1;

[0176] S210, if OptNoRotCounter <= 25, go to S205, else go to S211;

[0177] S211, for all elements in OptNoRotFeatureList, except the 13th element, subtract the 13th element from the element, and store the result in the element;

[0178] S212, create a list of convergent undirected attribute analysis OptAnaList = empty list;

[0179] S213, create a counter of convergent undirected attribute analysis OptAnaCounter = 1;

[0180] create a vector of convergent undirected attribute analysis minimum fluctuation OptAnaVector;

[0181] OptAnaVector is a vector of ZLCD elements, all of which are 0;

[0182] create a value of convergent undirected attribute analysis minimum fluctuation OptAnaValue = 100000;

[0183] S214, get a list of convergent undirected attribute analysis candidates OptAnaBX;

[0184] OptAnaBX is a list of integers, the content of OptAnaBX is determined according to the value of OptAnaCounter, specifically:

[0185] If OptAnaCounter is 1, the content of OptAnaBX is 1, 7, 19, 25;

[0186] If OptAnaCounter is 2, the content of OptAnaBX is 2, 8, 18, 24;

[0187] If OptAnaCounter is 3, the content of OptAnaBX is 3, 8, 18, 23;

[0188] If OptAnaCounter is 4, the content of OptAnaBX is 4, 8, 18, 22;

[0189] If OptAnaCounter is 5, the content of OptAnaBX is 5, 9, 17, 21;

[0190] If OptAnaCounter is 6, the content of OptAnaBX is 6, 12, 14, 20;

[0191] If OptAnaCounter is 7, the content of OptAnaBX is 11, 12, 14, 15;

[0192] If OptAnaCounter is 8, the content of OptAnaBX is 16, 12, 14, 10;

[0193] S215, the first temporary variable of convergent undirected attribute OptAnaTemp1 is obtained;

[0194] First, the content of OptAnaBX is determined according to OptAnaCounter;

[0195] Then, the table entry corresponding to the index is taken from OptNoRotFeatureList according to the content of OptAnaBX;

[0196] Finally, each table entry corresponding element is added and then divided by 4 to obtain the first temporary variable of convergent undirected attribute OptAnaTemp1;

[0197] S216, the second temporary variable of convergent undirected attribute OptAnaTemp2 is obtained;

[0198] OptAnaTemp1 is calculated according to column to form a vector of ZLCD elements as the second temporary variable of convergent undirected attribute OptAnaTemp2;

[0199] S217, the third temporary variable of convergent undirected attribute OptAnaTemp3 = standard deviation of OptAnaTemp2;

[0200] S218, if OptAnaTemp3 is less than OptAnaValue, go to S219, otherwise go to S221;

[0201] S219, OptAnaTemp2 = tanh(ABS(OptAnaTemp2) / Max(ABS(OptAnaTemp2)));

[0202] Where Max is to obtain the maximum value in a vector, and tanh is the hyperbolic tangent function;

[0203] S220, set OptAnaVector=OptAnaTemp2;

[0204] Set OptAnaValue = OptAnaTemp3;

[0205] S221, let OptAnaCounter = OptAnaCounter + 1;

[0206] S222, if OptAnaCounter is less than or equal to 8, go to S214, otherwise go to S223;

[0207] S223, establish the convergent undirected kinetic vector OptOutput=OptAnaVector for the power transmission corridor;

[0208] S224 returns OptOutput as the result of OptNoRot.

[0209] S3 uses the OptNoRot obtained in S2 to process all contents of the DataList obtained in S1, obtaining the transmission corridor convergence undirected attribute of each entry in the DataList, specifically:

[0210] S301, Set the ProcessCounter to 1;

[0211] S302, ProcessTemp1, the first temporary variable in the processing procedure, is the DLImage of the ProcessCounter element of DataList;

[0212] S303, the output of the second temporary variable ProcessTemp2=OptNoRot in the processing procedure is OptOutput;

[0213] OptNoRot's input OptNoRotInput=ProcessTemp1;

[0214] S304, Set the value of DLFeature of the ProcessCounter element of DataList to ProcessTemp1;

[0215] S305, let ProcessCounter=ProcessCounter+1;

[0216] S306, if ProcessCounter is less than or equal to ZLDFX, go to S302; otherwise, output OptOutput as the transmission corridor convergent undirected attribute DLFeature for each entry in DataList.

[0217] S4, establish a neighborhood convergence undirected attribute anomaly situation detection module OptYC, the input of OptYC is a neighborhood list to be analyzed OptYCInput and a vector of a point to be detected OptYCTestVector, the output of OptYC is a neighborhood convergence undirected attribute anomaly situation detection result OptYCResult, specifically:

[0218] S401, establish a neighborhood convergence undirected attribute anomaly situation detection module OptYC, the input of OptYC is a neighborhood list to be analyzed OptYCInput and a vector of a point to be detected OptYCTestVector;

[0219] S402, let the neighborhood convergence undirected attribute anomaly situation detection result OptYCResult=0;

[0220] S403, establish a detection center vector list OptYCTestCenter;

[0221] OptYCTestCenter contains 4 elements, the value of the first element is equal to OptYCTestVector, the value of the second element is equal to the value of the first element of OptYCInput, the value of the third element is the value of the ZLYZth element of OptYCInput, and the value of the fourth element is the value of the ZLYZ+1th element of OptYCInput;

[0222] S404, based on the 4 centers of OptYCTestCenter, K-Means clustering is performed on all elements of OptYCInput, and the number of each clustered element after clustering OptYCAfterNumber is obtained;

[0223] S405, if the value of the first element of OptYCAfterNumber is greater than or equal to 2, go to S407, otherwise go to S406;

[0224] S406, let OptYCResult=1, then execute S407;

[0225] S407, return OptYCResult as the result of OptYC;

[0226] S5, use OptYC and DataList to detect the damaged situation of the power transmission corridor, specifically:

[0227] S501, let the damaged counter of the power transmission corridor JCPHCounter=ZLYZ+1;

[0228] S502, detecting the first temporary variable JCPHTemp1 = a list of 2 x ZLYZ elements is established;

[0229] S503, the values of the DLFeature corresponding to the JCPHCounter-ZLYZth to the JCPHCounter-1th element in the DataList are taken out to form a list, which is stored in the first to the ZLYZth element of JCPHTemp1;

[0230] S504, the values of the DLFeature corresponding to the JCPHCounter+1th to the JCPHCounter+ZLYZth element in the DataList are taken out to form a list, which is stored in the ZLYZ+1th to the 2 x ZLYZth element of JCPHTemp1;

[0231] S505, detecting the second temporary variable JCPHTemp2 = the result OptYCResult of the OptYC;

[0232] The input of the OptYC is the list of the analyzed neighborhood OptYCInput = JCPHTemp1;

[0233] The input of the OptYC is the vector of the detected points OptYCTestVector = the value of the DLFeature of the JCPHCounterth element in the DataList;

[0234] S506, if JCPHTemp2 is equal to 0, go to S510, otherwise go to S507;

[0235] S507, detecting the third temporary variable JCPHTemp3 = the value of the DLPosX of the JCPHCounterth element in the DataList;

[0236] S508, detecting the fourth temporary variable JCPHTemp4 = the value of the DLPosY of the JCPHCounterth element in the DataList;

[0237] S509, obtaining the detection result of the damaged power transmission corridor: the position of the RSImg with the X coordinate of JCPHTemp3 and the Y coordinate of JCPHTemp4 detects the existence of the damaged power transmission corridor;

[0238] S510, JCPHCounter = JCPHCounter + 1;

[0239] S511, if the JCPHCounter is less than or equal to ZLDFX-ZLYZ, go to S502, otherwise output the result of the transmission corridor damage detection.

Claims

1. An automatic detection method for damage to power transmission corridors in mountainous areas based on remote sensing image data, characterized in that... The specific process of the method is as follows: S1. Based on the remote sensing image RSImg of the mountain power transmission corridor area and the list PointList of the locations of power transmission lines in the mountain power transmission corridor area, obtain a data list DataList of each power transmission line location; S2, establish the convergent undirected attribute extraction module OptNoRot for power transmission corridors. The input of OptNoRot is the image data to be processed, OptNoRotInput, and the output is the convergent undirected attribute vector of the power transmission corridor, OptOutput. S3, using the OptNoRot obtained in S2 to process all the contents of the DataList obtained in S1, obtains the transmission corridor convergent undirected property of each entry in the DataList; S4. Establish the neighborhood convergence undirected attribute anomaly detection module OptYC. The input of OptYC is the neighborhood list to be analyzed OptYCInput and the vector of the points to be detected OptYCTestVector. The output of OptYC is the neighborhood convergence undirected attribute anomaly detection result OptYCResult. S5 uses OptyC and DataList to detect damage to the power transmission corridor.

2. The automatic detection method for damage to mountain power transmission corridors based on remote sensing image data according to claim 1, characterized in that: In S1, based on the remote sensing image RSImg of the mountain power transmission corridor area and the list PointList of the locations of power transmission lines in the mountain power transmission corridor area, a data list DataList of individual power transmission line locations is obtained, specifically: S101, Input remote sensing image RSImg of the mountain power transmission corridor area; S102, Input a list PointList of the locations of power transmission lines in the mountainous power transmission corridor area; The PointList is a list consisting of the coordinates of the location points of the transmission line on RSImg; The PointList includes the following fields: the X-axis coordinate PosX of the transmission line location point on RSImg, and the Y-axis coordinate of the transmission line location point on RSImg; S103, input the preset width ZLKD of the power transmission corridor and the preset threshold ZLYZ for the destruction of the power transmission corridor; The default value for ZLKD is 5; The default value for ZLYZ is 10; S104, obtain the spatial resolution RSFBL of RSImg, the number of points to be analyzed ZLDFX, and the basic length of the analysis ZLCD: The spatial resolution of RSImg is RSFBL, which is the spatial resolution in the remote sensing image attribute information of RSImg. The number of points to be analyzed, ZLDFX, is equal to the number of elements in PointList. The base length to be analyzed is ZLCD = Round(ZLKD / RSFBL); Round() is the rounding operation; S105, determine whether ZLCD is even. If ZLCD is even, set ZLCD = ZLCD + 1 and then execute S106; otherwise, execute S106 directly. S106, Initialize the data list for each transmission line location, DataList = an empty list; DataList contains the following fields: image information DLImage corresponding to the location point of the transmission line, X-axis coordinate DLPosX of the center point of the image corresponding to the location point of the transmission line on RSImg, Y-axis coordinate DLPosY of the center point of the image corresponding to the location point of the transmission line on RSImg, and undirected attribute DLFeature of the transmission line location point converging with the transmission corridor. S107, set the initialization process counter JLCounter=1; S108, set the first temporary variable CSHTemp1 in the initialization process to the JLCounter element of PointList; S109, retrieve the second temporary variable CSHTemp2 from the initialization process: In RSImg, an image with a side length of 5×ZLCD pixels centered on PosX and PosY of CSHTemp1 is extracted as CSHTemp2. S110, Add a new element to DataList; The new element's DLImage = CSHTemp2; The new element's DLPosX = CSHTemp1's PosX; The new element's DLPosY = CSHTemp1's PosY; The new element's DLFeature is a 20-dimensional vector with all values ​​equal to 0. S111, let JLCounter = JLCounter + 1; S112, if JLCounter is less than or equal to ZLDFX, go to S108; otherwise go to S113. S113, Sort all elements in DataList according to their DLPosX values ​​in ascending order; S114, set the initialization process sorting counter JLPXCounter=2; S115, retrieve the first temporary variable JLPXTemp1 from the initialization process: a. Obtain the images corresponding to the power line location points of the JLPXCounter to ZLDFX elements in the DataList, and obtain the center point of each image; b. Find the center point that is closest to the center point of the image corresponding to the JLPXCounter-1 element's transmission line location point in the center point obtained in a. Use the current center point location number as the first temporary variable JLPXTemp1 for initialization process. S116, swap the positions of the JLPXTemp1-th and JLPXCounter-th elements of DataList; S117, set JLPXCounter = JLPXCounter + 1; S118: If JLPXCounter is less than or equal to ZLDFX-1, go to S115; otherwise, output DataList.

3. The automatic detection method for damage to mountain power transmission corridors based on remote sensing image data according to claim 2, characterized in that: The module OptNoRot in S2, which extracts convergent undirected attributes of the power transmission corridor, takes the image data to be processed (OptNoRotInput) as input and outputs the convergent undirected attribute vector of the power transmission corridor (OptOutput). Specifically: S201, Establish the convergent undirected attribute extraction module OptNoRot for power transmission corridors. The input of OptNoRot is the image data to be processed, OptNoRotInput. S202, the OptNoRotInput is cut into 5 horizontal and 5 vertical image blocks of equal size, resulting in a total of 25 blocks of equal size. The blocks are arranged from left to right and then from top to bottom to obtain the block image array OptNoRotArray. In OptNoRotArray, each image block has a size of ZLCD rows and ZLCD columns; S203, Create a sub-block attribute list OptNoRotFeatureList = an empty list; S204, set the sub-block calculation counter OptNoRotCounter=1; S205, the first temporary variable of the sub-block, OptNoRotTemp1, is the OptNoRotCounter-th element of OptNoRotArray; S206, convert OptNoRotTemp1 to a grayscale image and perform a two-dimensional fast Fourier transform to obtain OptFFT; S207, the zero-frequency component of OptFFT is moved to the center position to obtain the centered spectrum OptShift, and the amplitude spectrum OptMag of the transmission corridor is calculated using OptShift; S208, Add OptMag to OptNoRotFeatureList; S209, set OptNoRotCounter = OptNoRotCounter + 1; S210, if OptNoRotCounter is less than or equal to 25, go to S205; otherwise, go to S211. S211, For all elements in OptNoRotFeatureList, subtract the 13th item from the current item and store the result in the current item position; The current entry refers to all entries other than the 13 entries. S212, Create a convergent undirected attribute analysis list OptAnaList = an empty list; S213, Establish a convergent undirected attribute analysis counter OptAnaCounter=1; Establish the minimum fluctuation vector OptAnaVector for convergent undirected attribute analysis; The OptAnaVector is a vector with ZLCD elements, and all elements inside the OptAnaVector are 0; Establish the minimum fluctuation value OptAnaValue=100000 for convergent undirected attribute analysis; S214, obtain the candidate list OptAnaBX for convergent undirected attribute analysis; OptAnaBX is a list of integers, and the contents of OptAnaBX are determined based on the value of OptAnaCounter. S215, retrieve the first temporary variable OptAnaTemp1 for the convergent undirected attribute: First, determine the content of OptAnaBX based on OptAnaCounter; Then, based on the content of OptAnaBX, retrieve the corresponding item from OptNoRotFeatureList; Finally, sum the corresponding elements of each table entry and divide by 4 to obtain the first temporary variable OptAnaTemp1 of the convergent undirected attribute; Finally, sum the corresponding elements of each table entry and divide by 4 to obtain the first temporary variable OptAnaTemp1 of the convergent undirected attribute; S216, Obtain the second temporary variable OptAnaTemp2 for the convergent undirected attribute: The mean of OptAnaTemp1 is calculated column by column to form a vector of ZLCD elements, which serves as the second temporary variable OptAnaTemp2 for the convergent undirected attribute; S217, the standard deviation of the third temporary variable OptAnaTemp3 = OptAnaTemp2 for convergent undirected attributes; S218, if OptAnaTemp3 is less than OptAnaValue, then go to S219; otherwise go to S221. S219, let OptAnaTemp2=tanh(ABS(OptAnaTemp2) / Max(ABS(OptAnaTemp2))); Where Max is the maximum value in a vector, and tanh is the hyperbolic tangent function; S220, set OptAnaVector=OptAnaTemp2; Set OptAnaValue = OptAnaTemp3; S221, let OptAnaCounter = OptAnaCounter + 1; S222, if OptAnaCounter is less than or equal to 8, go to S214, otherwise go to S223; S223, establish the convergent undirected kinetic vector OptOutput=OptAnaVector for the power transmission corridor; S224 returns OptOutput as the result of OptNoRot.

4. The automatic detection method for damage to mountain power transmission corridors based on remote sensing image data according to claim 3, characterized in that: The calculation of the transmission corridor amplitude spectrum OptMag using OptShift in S207 is specifically as follows: OptMag=log(abs(OptShift)+0.5)+2.5 Where abs is used to calculate the absolute value, and log is used to calculate the natural logarithm.

5. The automatic detection method for damage to mountain power transmission corridors based on remote sensing image data according to claim 4, characterized in that: The content of OptAnaBX in S214 is determined based on the value of OptAnaCounter, specifically as follows: If OptAnaCounter is 1, then the content of OptAnaBX is 1,7,19,25; If OptAnaCounter is 2, then the content of OptAnaBX is 2,8,18,24; If OptAnaCounter is 3, then the content of OptAnaBX is 3,8,18,23; If OptAnaCounter is 4, then the content of OptAnaBX is 4,8,18,22; If OptAnaCounter is 5, then the content of OptAnaBX is 5,9,17,21; If OptAnaCounter is 6, then the content of OptAnaBX is 6, 12, 14, 20; If OptAnaCounter is 7, then the content of OptAnaBX is 11, 12, 14, 15; If OptAnaCounter is 8, then the content of OptAnaBX is 16, 12, 14, 10.

6. The automatic detection method for damage to mountain power transmission corridors based on remote sensing image data according to claim 5, characterized in that: In step S3, the OptNoRot obtained in S2 is used to process all contents of the DataList obtained in S1, and the transmission corridor convergence undirected attribute of each entry in the DataList is obtained, specifically as follows: S301, Set the ProcessCounter to 1; S302, ProcessTemp1, the first temporary variable in the processing procedure, is the DLImage of the ProcessCounter element of DataList; S303, the output of the second temporary variable ProcessTemp2=OptNoRot in the processing procedure is OptOutput; Wherein, the input of OptNoRot is OptNoRotInput=ProcessTemp1; S304, Set the value of DLFeature of the ProcessCounter element of DataList to ProcessTemp1; S305, let ProcessCounter=ProcessCounter+1; S306, if ProcessCounter is less than or equal to ZLDFX, go to S302; otherwise, output OptOutput as the transmission corridor convergent undirected attribute DLFeature for each entry in DataList.

7. The automatic detection method for damage to mountain power transmission corridors based on remote sensing image data according to claim 6, characterized in that: The S4 module for establishing the neighborhood convergence undirected attribute anomaly detection module OptYC takes the following inputs: the neighborhood list to be analyzed (OptYCInput) and the vector of the points to be detected (OptYCTestVector). The output of OptYC is the neighborhood convergence undirected attribute anomaly detection result (OptYCResult), specifically: S401, Establish the neighborhood convergence undirected attribute anomaly detection module OptYC. The input of OptYC is the neighborhood list to be analyzed OptYCInput and the vector of the point to be detected OptYCTestVector. S402, set the neighborhood convergence undirected attribute anomaly detection result OptYCResult=0; S403, Create a vector list of detection centers OptYCTestCenter; OptYCTestCenter contains 4 elements, where the value of the first element is equal to the value of OptYCTestVector, the value of the second element is equal to the value of the first element of OptYCInput, the value of the third element is the value of the ZLYZ-th element of OptYCInput, and the value of the fourth element is the value of the ZLYZ+1-th element of OptYCInput. S404, based on the four centers of OptYCTestCenter, performs K-Means clustering on all elements of OptYCInput to obtain the number of elements in each cluster after clustering, OptYCAfterNumber. S405, if the value of the first element of OptYCAfterNumber is greater than or equal to 2, go to S407; otherwise go to S406. S406, set OptYCResult=1, then execute S407; S407 returns OptYCResult as the result of OptYC.

8. The automatic detection method for damage to mountain power transmission corridors based on remote sensing image data according to claim 7, characterized in that: The detection of power transmission corridor damage using OptYC and DataList in S5 specifically involves: S501, set the counter for detecting damage to the power transmission corridor JCPHCounter = ZLYZ+1; S502, obtain the first temporary variable JCPHTemp1 for detecting damage to the power transmission corridor; S503, retrieve the values ​​of DLFeature corresponding to the JCPHCounter-ZLYZ to JCPHCounter-1 elements in DataList, form a list, and store it in the corresponding positions of the 1st to ZLYZth elements of JCPHTemp1; S504, retrieve the values ​​of DLFeature corresponding to the JCPHCounter+1 to JCPHCounter+ZLYZ elements in DataList, form a list, and store it in the positions corresponding to the ZLYZ+1 to 2×ZLYZ elements of JCPHTemp1; S505, the result of detecting damage to the power transmission corridor, the second temporary variable JCPHTemp2=OptYC, is OptYCResult; The input neighborhood list to be analyzed for OptYC is OptYCInput=JCPHTemp1; The input vector of the point to be detected in OptYC, OptYCTestVector, is the value of the DLFeature of the JCPHCounter element in DataList; S506, if JCPHTemp2 equals 0, then go to S510; Otherwise, proceed to S507; S507, detect the value of DLPosX of the third temporary variable JCPHTemp3 in DataList for the damaged transmission corridor; S508, detect the value of DLPosY of the fourth temporary variable JCPHTemp4 = JCPHCounter element in DataList when the power transmission corridor is damaged; S509, using JCPHTemp3 and JCPHTemp4 to obtain the damage detection results of the power transmission corridor; S510, set JCPHCounter = JCPHCounter + 1; S511, if JCPHCounter is less than or equal to ZLDFX-ZLYZ, proceed to S502; otherwise, output the result of the power transmission corridor damage detection.

9. The automatic detection method for damage to mountain power transmission corridors based on remote sensing image data according to claim 8, characterized in that: The first temporary variable JCPHTemp1 in S502 for detecting damage to the power transmission corridor is a list containing 2 × ZLYZ elements.

10. The automatic detection method for damage to mountain power transmission corridors based on remote sensing image data according to claim 9, characterized in that: The detection result of the damage to the power transmission corridor in S509 is as follows: damage to the power transmission corridor was detected at the position where the X coordinate of RSImg is JCPHTemp3 and the Y coordinate is JCPHTemp4.

Images