DEM window optimization selection method for power grid microtopography recognition

Abstract

The invention discloses a DEM window optimization selection method for power grid microtopography recognition, and the method comprises the steps: S1, inputting the coordinates of a transmission line tower, judging whether the transmission line tower is located on a flat ground or not, and if yes, determining that no microtopography exists; if not, entering the step S2; s2, judging whether the TPI / SD absolute value in the range a around the pole tower is greater than 1 or not, and if so, determining that the terrain is a peak-valley terrain; if not, determining that the terrain is a slope terrain; and S3, calculating TPI values and elevation SD values under different window sizes, analyzing TPI / SD curve changes, finding out the corresponding window size at the inflection point of the curve, and outputting the optimal window identification size under the microtopography. The size of the window can be adaptively adjusted according to the fluctuation speed of the terrain, so that the micro-terrain features are accurately captured.

Description

Technical Field

[0001] The invention relates to the field of electrical engineering, and in particular to a DEM window optimization selection method for power grid micro-topography identification. Background Art

[0002] In the power system, the operational safety of transmission lines is significantly affected by micro-topography and micro-meteorology, especially under extreme weather conditions (such as ice cover, strong wind, etc.). The deterioration of local meteorological conditions caused by the ground terrain is one of the main causes of line failures. The power grid micro-topography recognition technology aims to identify the micro-topography features that may affect the power grid facilities through accurate analysis of the terrain around the transmission line, and provide technical support for the planning, design, maintenance and disaster warning of the transmission line. With the expansion of the transmission line network, the accuracy, real-time and universality of the micro-topography recognition method have become the key points, and how to accurately select the window size of micro-topography recognition to cope with different terrain change speeds and terrain characteristics and improve the recognition accuracy is worthy of attention.

[0003] At present, in the existing technology, the fixed window method can be used to extract the micro-topography elevation value, and the maximum value of the difference degree of the ground feature information in different windows is analyzed as the basic idea, and the mean change point analysis method is used to obtain the optimal neighborhood analysis window; a rotating extraction window can also be used to extract fixed range elevation data, and the rotation range is 180° to obtain the optimal window size; and a fixed window method can also be used, with the central grid as the center of the circle, to traverse the curvature differential of the boundary grid to obtain the optimal window size. Most of the above-mentioned existing technologies rely on the mean change point analysis method or the curvature differential method to determine the optimal window size. These methods cannot be flexibly adjusted when dealing with areas with rapid terrain changes, thereby affecting the accuracy of recognition. Especially in areas with drastic terrain changes such as mountains and canyons, fixed or rotating windows may not accurately match micro-topography features, resulting in a decrease in recognition accuracy.

[0004] Therefore, in order to solve the above problems, a DEM window optimization selection method for power grid micro-topography identification is needed, which can adaptively adjust the window size according to the speed of terrain fluctuations, so as to accurately capture the micro-topography characteristics. Summary of the invention

[0005] In view of this, the purpose of the present invention is to overcome the defects in the prior art and provide a DEM window optimization selection method for power grid micro-topography identification, which can adaptively adjust the window size according to the speed of terrain fluctuations, so as to accurately capture the micro-topography characteristics.

[0006] The DEM window optimization selection method for power grid micro-topography identification of the present invention comprises:

[0007] S1. Input the coordinates of the transmission line tower to determine whether it is located on flat ground. If so, there is no micro-topography; if not, proceed to step S2;

[0008] S2. Determine whether the absolute value of TPI / SD within a range around the tower is greater than 1. If so, it is a peak-valley terrain; if not, it is a slope terrain;

[0009] S3. Calculate the TPI value and elevation SD value under different window sizes, analyze the change of TPI / SD curve, find the corresponding window size at the inflection point of the curve, and output the optimal window recognition size under the micro-topography.

[0010] Further, the TPI value is determined according to the following formula:

[0011]

[0012] Among them, Z c Represents the elevation value of the center grid of the analysis window. is the average elevation value within the analysis window, Z i is the elevation value at the i-th grid in the window, R is the analysis window size, N R Represents the number of grids within the analysis window size R.

[0013] Further, the elevation SD value is determined according to the following formula:

[0014]

[0015] Furthermore, the radius size of the topographic point topographic position description window is set to 80 grids, and the area with a TPI / SD absolute value greater than 1 with a window radius of 80 grids is described as a peak and valley class, and the area with a TPI / SD absolute value less than 1 with a window radius of 80 grids is described as a slope class.

[0016] Furthermore, for peak and valley areas, by calculating the TPI values ​​and elevation SD values ​​under different window sizes, the changes in the TPI / SD curve are analyzed. When the absolute value of TPI / SD is the largest, the elevation difference between the central area of ​​the current window and other areas in the analysis window is the largest. The window size corresponding to the maximum value of the TPI / SD curve is selected as the micro-topography recognition range, and the local elevation map is cropped, and then input into the trained convolutional neural network model for micro-topography recognition.

[0017] Furthermore, for slope areas, the local terrain flipping method is used to vertically flip the terrain above the point according to the elevation value of the point, and construct it into a peak and valley area. Then, the optimal micro-topography recognition window is calculated according to the peak and valley area, and the local elevation map is cropped and then input into the trained convolutional neural network model for micro-topography recognition.

[0018] Furthermore, before analyzing the change of the TPI / SD curve, the TPI / SD curve is smoothed.

[0019] The beneficial effects of the present invention are as follows: the DEM window optimization selection method for power grid micro-topography identification disclosed by the present invention can adaptively optimize the identification window on the basis of establishing the identification model and the power grid micro-topography intelligent identification software, realize the optimization of the input level of the identification model, improve the universality of the micro-topography identification program, and make it suitable for the micro-topography identification of power transmission lines in various regions across the country. The present invention can flexibly adjust the window according to the rate of terrain change, avoiding the problem that the fixed window size cannot adapt to complex terrain changes. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The present invention will be further described below in conjunction with the accompanying drawings and embodiments:

[0021] Figure 1 (a) is a schematic diagram of the best recognition window selection for terrain with fast terrain changes according to the present invention;

[0022] Figure 1 (b) is a schematic diagram of the best recognition window selection for terrain with slow terrain changes according to the present invention;

[0023] Figure 2 A schematic diagram of describing terrain position based on TPI of the present invention;

[0024] Figure 3 A schematic diagram of describing terrain position based on TPI / SD of the present invention;

[0025] Figure 4 Schematic diagram of the relationship between the optimal window size and |TPI / SD| of the present invention;

[0026] Figure 5 A schematic diagram of local terrain reversal in a slope-type area of ​​the present invention;

[0027] Figure 6 This is a flow chart of the TPI-based adaptive window adjustment of the present invention. DETAILED DESCRIPTION

[0028] The present invention is further described below in conjunction with the accompanying drawings, as shown in the drawings:

[0029] This embodiment discloses a DEM window optimization selection method for power grid micro-topography identification, comprising the following steps:

[0030] S1. Input the coordinates of the transmission line tower to determine whether it is located on flat ground. If so, there is no micro-topography; if not, proceed to step S2;

[0031] S2. Determine whether the absolute value of TPI / SD within a range a around the tower is greater than 1. If so, it is a peak-valley type terrain; if not, it is a slope type terrain; wherein the value of a can be 2 km.

[0032] S3. Calculate the TPI value and elevation SD value under different window sizes, analyze the change of TPI / SD curve, find the corresponding window size at the inflection point of the curve, and output the optimal window recognition size under the micro-topography.

[0033] Since the terrain changes at different speeds in different regions, it is easy to cause the problem of including multiple micro-topography in a single recognition window or incomplete inclusion of a single micro-topography feature. However, the present invention realizes the selection of the optimal window size for micro-topography recognition that is adaptive to the speed of terrain change by analyzing the absolute value of TPI / SD of the grid DEM data in different window sizes centered on the tower. Figure 1 Schematic diagram for selecting the best identification window for different terrain change speeds.

[0034] In this embodiment, TPI (Topographic Position Index) is an index used to describe the topographic position, which is often used in terrain analysis and geomorphology research to quantify the relative height of the position of a topographic point relative to the surrounding terrain. TPI can be used to describe whether a topographic point is located at a top of a mountain, in a valley, or on a hillside. A positive value indicates that the point is located near a highland or a top of a mountain, a negative value indicates that the point is located near a lowland or a valley, and a value close to zero indicates that the point is located on a hillside. Figure 2 As shown, the calculation formula is as follows:

[0035]

[0036] Among them, Z i is the elevation value at the i-th grid in the window, R is the analysis window size, N R Represents the number of grids within the analysis window size R, Z c Represents the elevation value of the center grid of the analysis window. It is the average elevation value within the analysis window.

[0037] In order to more intuitively reflect the elevation difference between the central area and other areas in the analysis window, the SD standard deviation is introduced and compared with the TPI value. The calculation formula is as follows:

[0038]

[0039] Where SD represents the standard deviation of elevation within the analysis window, and TPI represents the difference between the central area elevation value and the average value within the analysis window.

[0040] The present invention uses TPI / SD to describe the topographic position of the topographic point at the tower to avoid the constraint of fixed thresholds and ensure that the algorithm can still extract the corresponding topographic feature information in different terrain scenarios. Since there is no micro-topography when the transmission line is located on flat ground, the elevation difference at the topographic point is firstly judged to exclude the flat ground area, and the topographic position of the topographic point with an elevation difference greater than 200m is described, which is divided into two types of topographic positions: peak and valley type and slope type. Figure 3 shown.

[0041] Among them, according to the "330kV~750kV Overhead Transmission Line Survey Specifications", the digital elevation model (DEM) search range of the newly built transmission line domain is set to 2 kilometers and the DEM accuracy is not less than 25m. The DEM elevation data accuracy used in the present invention is 12.5m. Therefore, the present invention sets the window radius size of the terrain point terrain position description to 80 grids, and describes the area with a TPI / SD absolute value greater than 1 with a window radius of 80 grids as a peak and valley type, and describes the area with a TPI / SD absolute value less than 1 with a window radius of 80 grids as a slope type.

[0042] In this embodiment, for the peak-valley area, by calculating the TPI value and the elevation SD value under different window sizes, the change of the TPI / SD curve is analyzed. When the absolute value of TPI / SD is the largest, the elevation difference between the center area of ​​the current window and other areas in the analysis window is the largest, such as Figure 4 As shown in the figure, the window size corresponding to the maximum value of the TPI / SD curve is selected as the micro-topography recognition range, and the local elevation map is cropped, and then the trained convolutional neural network model is input for micro-topography recognition.

[0043] In practical applications, the actual terrain is far more complex than the ideal terrain in the figure. Although the introduction of the SD value eliminates some disturbances, the TPI / SD curve will still fall into the local minimum for some larger local terrains. Therefore, in order to eliminate the TPI / SD curve disturbance caused by local terrain, the present invention smoothes the TPI / SD curve.

[0044] In this embodiment, for slope areas, since the TPI / SD curve changes under different window sizes always change between (-1, 1), analyzing the stationary point of the TPI / SD curve is basically meaningless for identifying the window size selection. Therefore, the present invention adopts a local terrain flipping method to vertically flip the terrain above the point according to the elevation value of the point as the axis, such as Figure 5 As shown in the figure, it is constructed into peak and valley areas, and then the optimal micro-topography recognition window is calculated according to the peak and valley areas, and the local elevation map is cropped, and then the trained convolutional neural network model is input for micro-topography recognition.

[0045] The adaptive window optimization method proposed in this invention provides a more accurate and flexible solution for power grid micro-terrain identification technology. By combining TPI / SD analysis with dynamic window adjustment, it can solve the problems of insufficient terrain adaptability and incomplete identification in the existing technology, and improve the accuracy and universality of micro-terrain identification. With the increasing demand for intelligent and automated management of power grids, this technology has broad application prospects.

[0046] In order to better understand the present invention, the following verification analysis is now performed on the present invention:

[0047] Since the actual terrain is much more complex than the ideal terrain in practical applications, in order to verify the effectiveness of the adaptive window adjustment proposed in the present invention, a 220kV line tower 1 to 12 is selected as a test line for effectiveness verification. Among them, towers 4 to 6 are in the heavy ice area, which is located in a typical alpine watershed micro-topography, and the terrain of the line changes complexly, and the fixed recognition window size is not good. Among them, the terrain of towers 10 to 12 changes faster, and a larger recognition window size is required, while the terrain of towers 4 to 6 changes slowly, and a smaller recognition window size is required.

[0048] Table 1 shows the TPI / SD values ​​of each tower number in the line and the identified terrain position and the actual terrain position. By comparing the identified terrain position and the actual terrain position of each tower in the table, except for towers No. 3 and No. 10, which are misidentified, the other towers are correctly identified. Table 1 also shows the optimal window size for each tower and the actual optimal window size. It can be seen that there is a large difference between towers No. 3 and No. 4, and the window sizes for the other towers are basically consistent with the actual optimal window sizes. In summary, the adaptive window adjustment proposed by the present invention has high accuracy.

[0049] Table 1

[0050]

[0051] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

Claims

1. A DEM window optimization selection method for power grid micro-topography identification, characterized by: include: S1. Input the coordinates of the transmission line tower to determine whether it is located on flat ground. If so, there is no micro-topography; If not, proceed to step S2; S2. Determine whether the absolute value of TPI / SD within a range around the tower is greater than 1. If so, it is a peak-valley terrain; if not, it is a slope terrain; S3. Calculate the TPI value and elevation SD value under different window sizes, analyze the change of TPI / SD curve, find the corresponding window size at the inflection point of the curve, and output the optimal window recognition size under the micro-topography.

2. The DEM window optimization selection method for power grid micro-topography identification according to claim 1 is characterized in that: The TPI value is determined according to the following formula: Among them, Z c Represents the elevation value of the center grid of the analysis window. is the average elevation value within the analysis window, Z i is the elevation value at the i-th grid in the window, R is the analysis window size, N R Represents the number of grids within the analysis window size R.

3. The DEM window optimization selection method for power grid micro-topography identification according to claim 1 is characterized in that: The elevation SD value is determined according to the following formula:

4. The DEM window optimization selection method for power grid micro-topography identification according to claim 2 is characterized in that: The radius size of the window describing the terrain position of the terrain point was set to 80 grids, and the area with the absolute value of TPI / SD greater than 1 with the window radius of 80 grids was described as the peak and valley class, and the area with the absolute value of TPI / SD less than 1 with the window radius of 80 grids was described as the slope class.

5. The DEM window optimization selection method for power grid micro-topography identification according to claim 1 is characterized in that: For peak and valley areas, by calculating the TPI values ​​and elevation SD values ​​under different window sizes, the changes in the TPI / SD curve are analyzed. When the absolute value of TPI / SD is the largest, the elevation difference between the central area of ​​the current window and other areas in the analysis window is the largest. The window size corresponding to the maximum value of the TPI / SD curve is selected as the micro-topography recognition range, and the local elevation map is cropped, and then input into the trained convolutional neural network model for micro-topography recognition.

6. The DEM window optimization selection method for power grid micro-topography identification according to claim 1 is characterized in that: For slope areas, the local terrain flipping method is used to flip the terrain above the point vertically along the elevation value of the point, and construct it into a peak-valley area. Then, the optimal micro-topography recognition window is calculated according to the peak-valley area, and the local elevation map is cropped. The trained convolutional neural network model is then input for micro-topography recognition.

7. The DEM window optimization selection method for power grid micro-topography identification according to claim 1 is characterized in that: Before analyzing the changes in the TPI / SD curve, the TPI / SD curve was smoothed.

Images