Method, system and device for extracting ridge line reflecting landform features and medium

Abstract

The invention provides a method, a system and equipment for extracting a ridge line reflecting landform characteristics and a medium. The method comprises the following steps: S1, extracting elevation data of a digital elevation model for researching a real terrain in a mountain region range; s2, performing double-precision resampling on the elevation data, and arranging the elevation data into a matrix only with the elevation data; s3, performing wavelet decomposition on the matrix of the elevation data by adopting different wavelet functions and decomposition layers, and reconstructing a low-frequency term in a wavelet decomposition result; s4, comprehensively evaluating a low-frequency item reconstruction result according to a signal reconstruction error and a noise reduction degree, and determining an optimal wavelet function and a decomposition layer number; s5, performing filtering and noise reduction on the matrix of the elevation data by adopting the optimal wavelet function and the decomposition layer number; and S6, extracting a terrain ridge line from the elevation data matrix after noise reduction by adopting an eight-neighborhood relative elevation discrimination method. The method provided by the invention is suitable for better understanding the landform of the mountainous area in the design of the power transmission line in the mountainous area.

Description

Technical Field

[0001] The present invention belongs to the technical field of digital terrain analysis, and in particular relates to a ridgeline extraction method, system, equipment and medium that reflect landform characteristics. Background Art

[0002] There are more and more cases of building transmission lines in mountainous areas, and the specific terrain will cause the flow field to accelerate, which will threaten the safe operation of the transmission line. Therefore, it is necessary to pay special attention to the geomorphic features. However, the real terrain is tortuous and has too many local details, which makes it difficult for designers to grasp the geomorphic features such as the shape and direction of the real terrain in engineering practice. Therefore, it is necessary to extract ridgelines to reflect the geomorphic features, which will help designers to strengthen their understanding of mountainous landforms.

[0003] From a hydrological perspective, ridgeline extraction is essentially the extraction of watershed lines. As the starting point of water flow, the watershed line has zero runoff accumulation. It is only necessary to calculate the surface runoff and extract the zero-value runoff accumulation grid to obtain the watershed line, that is, the ridgeline in the hydrological sense. However, this method is mainly suitable for large-scale terrain analysis, and it may be difficult to accurately describe the details of small-scale terrain. At the same time, simply analyzing the area with zero runoff accumulation is likely to cause the loss and discontinuity of the ridgeline.

[0004] Another method of digital simulation of water flow on a three-dimensional terrain surface is to more easily identify the ridgelines in the terrain through the inverse DEM model, create a drainage network based on the calculation results of the accumulated flow volume, and select the drainage network with a larger flow rate as the ridgeline by setting the flow rate. However, this method is mainly implemented through the hydrological analysis module provided by ArcGIS software, which limits the flexibility of use. At the same time, the flow rate needs to be constantly adjusted during use to highlight the ridgelines required for the study. When a very dense ridge network is required, a smaller flow value can be set, but the primary and secondary ridgeline networks are mixed and need to be further marked and identified; when only the main ridge network is required, a larger flow value can be set, but the ridgeline is usually disconnected and needs to be refined according to the terrain. Summary of the invention

[0005] The first purpose of the present invention is to provide a ridgeline extraction method that reflects landform characteristics in order to address the deficiencies of the prior art, in the hope of simplifying terrain elevation information, highlighting features such as the shape and direction of the terrain, and making different terrains more recognizable.

[0006] To this end, the above-mentioned purpose of the present invention is achieved through the following technical solutions:

[0007] A ridgeline extraction method reflecting landform characteristics mainly includes the following steps:

[0008] S1. Determine the mountainous area to be studied and extract the digital elevation model (DEM) elevation data of the real terrain within the mountainous area to be studied;

[0009] S2, resampling the DEM elevation data in step S1 with 2 times the precision and arranging them into a matrix containing only elevation data;

[0010] S3, using different wavelet functions and decomposition levels to perform wavelet decomposition on the elevation data matrix in step S2, and reconstructing the low-frequency terms in the wavelet decomposition result;

[0011] S4, comprehensively evaluating the reconstruction result of the low-frequency term in step S3 according to the signal reconstruction error and the degree of noise reduction, and determining the optimal wavelet function and decomposition level used for wavelet decomposition;

[0012] S5, using the optimal wavelet function and decomposition layer number in step S4, filtering and denoising the elevation data matrix in step S2;

[0013] S6. Using the eight-neighborhood relative elevation discrimination method, extract terrain ridge lines from the elevation data matrix after noise reduction in step S5.

[0014] While adopting the above technical solutions, the present invention may also adopt or combine the following technical solutions:

[0015] As a preferred technical solution of the present invention: in step S2, according to task requirements, the elevation data matrix of the entire area is decomposed into sub-areas of uniform size, and ridge line extraction is implemented in the sub-areas.

[0016] As a preferred technical solution of the present invention: in step S3, two-dimensional discrete wavelet transform is used to decompose the elevation data matrix in both horizontal and vertical directions.

[0017] As a preferred technical solution of the present invention: in step S4, the reconstruction error and the degree of noise reduction are important indicators for measuring whether the terrain noise reduction is successful. The chamfer distance is used to measure the reconstruction error d CD , and its calculation formula is as follows:

[0018]

[0019] In the formula, S 1 and S 2 Respectively represent the elevation point cloud before and after reconstruction; the first item Represents the elevation point cloud S before reconstruction 1 From any point in the reconstructed elevation point cloud S 2 The average of the sum of the least square distances, the second term Represents the reconstructed elevation point cloud S 2 From any point in the image to the elevation point cloud S before reconstruction1 The average of the sum of the least square distances;

[0020] Secondly, the terrain roughness index TRI is used to measure the degree of noise reduction to prevent the reconstruction error from being too small, resulting in incomplete terrain noise reduction. The calculation formula is as follows:

[0021]

[0022] Where: N is the window length, Z 0 is the elevation of the center, Z k is the elevation of the neighboring location.

[0023] As a preferred technical solution of the present invention: in step S4, based on the primary noise reduction criterion of reconstruction error, the db8 wavelet is selected as the optimal wavelet function for terrain noise reduction; when the noise reduction degree and the reconstruction error are balanced, the optimal number of wavelet decompositions can be determined; it should be pointed out that the optimal number of wavelet decompositions is related to the size of the analysis area in step S2.

[0024] As a preferred technical solution of the present invention: in step S6, the relative elevations of the eight neighborhoods around the center point are judged, and when at least five neighborhoods are lower than the center point, it can be confirmed that the center point is the ridge line position.

[0025] The second object of the present invention is to provide a ridgeline extraction system reflecting landform characteristics, comprising the following modules:

[0026] - an elevation data extraction module, the elevation data extraction module is used to extract the digital elevation model elevation data of the real terrain within the research mountain area according to the determined research mountain area;

[0027] - a matrix arrangement module, which is used to resample the elevation data extracted by the elevation data extraction module with multiple times of precision and arrange them into a matrix containing only elevation data;

[0028] - a low-frequency term reconstruction module, which is used to use different wavelet functions and decomposition levels to perform wavelet decomposition on the matrix of elevation data arranged by the matrix arrangement module, and reconstruct the low-frequency terms in the wavelet decomposition result;

[0029] - An optimal wavelet function and decomposition layer number determination module, which is used to comprehensively evaluate the low-frequency term reconstruction result of the low-frequency term reconstruction module according to the signal reconstruction error and the degree of noise reduction, and determine the optimal wavelet function and decomposition layer number used for wavelet decomposition;

[0030] - A filtering and denoising module, wherein the filtering and denoising module is used to filter and denoise the matrix of the elevation data arranged by the matrix arrangement module by using the optimal wavelet function and the number of decomposition layers determined by the optimal wavelet function and the number of decomposition layers determination module;

[0031] - A terrain ridgeline extraction module, which is used to extract terrain ridgelines from the elevation data matrix after denoising by the filtering and denoising module by using an eight-neighborhood relative elevation discrimination method.

[0032] The third object of the present invention is to provide an electronic device, comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus, and wherein:

[0033] a memory for storing a computer program,

[0034] A processor is used to execute the computer program stored in the memory to implement the steps of the ridgeline extraction method reflecting the landform characteristics as described above.

[0035] Another object of the present invention is to provide a computer-readable storage medium, characterized in that: the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the ridge line extraction method reflecting the landform characteristics as described above are implemented.

[0036] The present invention provides a ridgeline extraction method, system, equipment and medium reflecting landform characteristics, aiming to simplify terrain elevation information, highlight the shape and direction of terrain and other features, so as to improve the recognition of terrain. The method performs wavelet decomposition on the elevation data matrix arranged after resampling, and reconstructs low-frequency items; evaluates the reconstruction result through signal reconstruction error and noise reduction degree, determines the optimal wavelet function and decomposition layer number, and uses the optimal wavelet function and decomposition layer number to filter and reduce noise on the elevation data matrix; finally, an eight-neighborhood relative elevation discrimination method is used to extract terrain ridgelines.

[0037] Compared with the prior art, it has the following beneficial effects:

[0038] (1) The method provided by the present invention is suitable for better understanding the topography of mountainous areas in the design of mountainous power transmission lines.

[0039] (2) Use wavelet decomposition, especially two-dimensional discrete wavelet decomposition, to perform high-frequency filtering on real complex terrain elevation data, filter out the details represented by high-frequency items, and reconstruct low-frequency trend items, thereby achieving the effect of denoising the terrain details and highlighting the main shape of the terrain.

[0040] (3) After denoising the DEM elevation data of the real terrain, the continuous ridge lines of the terrain are further extracted to highlight the main features of the terrain and improve the terrain recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 The present invention provides a flow chart of a ridgeline extraction method that reflects landform features.

[0042] Figure 2 It is an elevation image of 5km×5km terrain.

[0043] Figure 3 This is a schematic diagram of the principle of two-dimensional discrete wavelet decomposition.

[0044] Figure 4a Reconstruction error map of low-frequency terms of Db N wavelet decomposition.

[0045] Figure 4b Reconstruction error map of low-frequency terms of Sym N wavelet decomposition.

[0046] Figure 4c Reconstruction error map of low-frequency terms of CoifN wavelet decomposition.

[0047] Figure 4d Reconstruction error map of low-frequency terms of BiorN.M wavelet decomposition.

[0048] Figure 5 Reconstruction error map of low-frequency terms for different wavelet functions at different decomposition levels.

[0049] Figure 6 This is a diagram showing the comprehensive impact of the number of decomposition layers on the reconstruction results of low-frequency terms.

[0050] Figure 7 The elevation image is reconstructed from the low-frequency terms after the fifth decomposition of the 5km×5km terrain.

[0051] Figure 8 This is the principle diagram of the eight-neighborhood relative elevation discrimination method.

[0052] Fig. 9 This is the ridgeline extraction result of 5km×5km terrain before wavelet denoising.

[0053] Fig.10 This is the ridgeline extraction result after 5km×5km terrain wavelet denoising. DETAILED DESCRIPTION

[0054] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0055] As introduced in the background technology, the prior art generally uses hydrological methods to extract ridgelines, which easily leads to the loss and discontinuity of ridgelines, resulting in the main geomorphic features of the terrain not being prominent enough. In order to solve the above technical problems, the present invention provides a ridgeline extraction method that reflects geomorphic features, and its flow chart is as follows: Figure 1 shown.

[0056] The mountainous area to be studied was determined, and the digital elevation model (DEM) elevation data of the real terrain within the mountainous area was extracted. The data came from the ASTER dataset.

[0057] According to the research object, the elevation data is resampled with 2 times the accuracy and arranged into a matrix with only elevation data. Then, according to the task requirements, the elevation data matrix of the entire area can be divided into 5km×5km, and the window is slid horizontally and vertically by 2km to decompose it into sub-areas of uniform size. Among them, the elevation image of any 5km×5km terrain is as follows Figure 2 shown.

[0058] The elevation data matrix of the sub-region is decomposed by multi-scale two-dimensional discrete wavelet using different wavelet functions and decomposition layers. The signal passes through a low-pass filter and a high-pass filter associated with the wavelet basis to divide the signal frequency into two equal parts. Each time, only the low-frequency part is further decomposed, and the low-frequency terms in the wavelet decomposition result are reconstructed. The decomposition process based on the Mallat tower algorithm is as follows: Figure 3 As shown in Figure 2, terrain elevation data is an approximate representation of the surface by assigning height values ​​to regularly distributed grid cells on a two-dimensional plane. Its scale subspace V j and the wavelet subspace W j The orthogonal decomposition in Hilbert space can be expressed as follows:

[0059]

[0060] Where: V j-1 is the approximate coefficient, is the horizontal detail coefficient, is the vertical detail factor, is the diagonal detail coefficient; Represents the vector direct sum operation of two orthogonal subspaces.

[0061] Furthermore, the 2-scale 2D discrete wavelet decomposition can be expressed as a further decomposition of the approximate term on the right side of the above equation. The multi-scale 2D discrete wavelet decomposition is similar. The 2-scale 2D discrete wavelet orthogonal decomposition can be expressed as follows:

[0062]

[0063] Reconstruction error and denoising degree are important indicators to measure whether terrain denoising is successful. The chamfer distance is used to measure the reconstruction error, and its calculation formula is as follows:

[0064]

[0065] In the formula, S 1 and S 2 Respectively represent the elevation point cloud before and after reconstruction; the first item Represents the elevation point cloud S before reconstruction 1 From any point in the reconstructed elevation point cloud S 2 The average of the sum of the least square distances, the second term Represents the reconstructed elevation point cloud S 2 From any point in the image to the elevation point cloud S before reconstruction 1 The average of the sums of the least squared distances.

[0066] The terrain roughness index is further used to measure the degree of noise reduction, which can prevent the reconstruction error from being too small and the terrain noise reduction from being incomplete. The calculation formula is as follows:

[0067]

[0068] Where: N is the window length, Z 0 is the elevation of the center, Z k is the elevation of the neighboring location.

[0069] right Figure 2 The terrain shown in the figure is decomposed into two-dimensional discrete wavelets using different wavelet functions. The reconstruction error results of the low-frequency terms are shown in Figure 4a to Figure 4d . It can be seen from the figure that the reconstruction error distribution of the approximate terms of different wavelet functions is similar, and they all decrease rapidly with the increase of the vanishing distance, and then tend to be stable. According to the characteristics of the error distribution, db8, sym7, coif4 and bior3.5 are selected as the optimal wavelet functions to be selected. Figure 5 It can be obtained that the reconstruction results of different wavelet functions are very different, and the reconstruction errors of approximate terms at different decomposition levels are very close. Considering that the db8 wavelet has stronger biorthogonality and symmetry, it is suitable for signal compression and denoising. Based on the reconstruction error, the primary denoising criterion, the db8 wavelet is selected as the optimal wavelet function for terrain denoising.

[0070] Using db8 wavelet function Figure 2 The terrain elevation shown in is decomposed at multiple levels and the approximation is reconstructed. Figure 6 It can be obtained that when the degree of noise reduction (roughness index) and the reconstruction error (chamfer distance) reach a balance, the optimal number of wavelet decompositions can be determined.

[0071] In summary, the db8 wavelet function is used to Figure 2 The 5km×5km terrain is decomposed 5 times, and the results are as follows Figure 7 shown.

[0072] The eight-neighborhood relative elevation discrimination method is further used to extract the terrain ridge lines one by one from the elevation submatrix. The algorithm principle diagram is shown in the figure below: Figure 8 As shown. Figure 2 The ridge lines are extracted from the original terrain. The results are shown in Fig. 9 .right Figure 7 The ridgeline is extracted from the denoised terrain. The results are shown in Fig.10 .Depend on Fig. 9 and Fig.10 It can be seen that the ridgeline extraction results after wavelet filtering and denoising are relatively continuous and concise, the position and shape of the terrain can be intuitively obtained, and the main features of the terrain are highlighted. The results show that the terrain features after denoising are more distinct, and different terrains have a high degree of recognition.

[0073] The present invention also provides a ridgeline extraction system reflecting landform features, comprising the following modules:

[0074] - an elevation data extraction module, the elevation data extraction module is used to extract the digital elevation model elevation data of the real terrain within the research mountain area according to the determined research mountain area;

[0075] - a matrix arrangement module, which is used to resample the elevation data extracted by the elevation data extraction module with multiple times of precision and arrange them into a matrix containing only elevation data;

[0076] - a low-frequency term reconstruction module, which is used to use different wavelet functions and decomposition levels to perform wavelet decomposition on the matrix of elevation data arranged by the matrix arrangement module, and reconstruct the low-frequency terms in the wavelet decomposition result;

[0077] - An optimal wavelet function and decomposition layer number determination module, which is used to comprehensively evaluate the low-frequency term reconstruction result of the low-frequency term reconstruction module according to the signal reconstruction error and the degree of noise reduction, and determine the optimal wavelet function and decomposition layer number used for wavelet decomposition;

[0078] - A filtering and denoising module, wherein the filtering and denoising module is used to filter and denoise the matrix of the elevation data arranged by the matrix arrangement module by using the optimal wavelet function and the number of decomposition layers determined by the optimal wavelet function and the number of decomposition layers determination module;

[0079] - A terrain ridgeline extraction module, which is used to extract terrain ridgelines from the elevation data matrix after denoising by the filtering and denoising module by using an eight-neighborhood relative elevation discrimination method.

[0080] The present invention also provides an electronic device, which includes a processor, a communication interface, a memory and a communication bus. The processor, the communication interface and the memory communicate with each other through the communication bus. The memory is used to store computer programs. The processor is used to execute the computer programs stored in the memory to implement the steps of the ridge line extraction method reflecting the landform characteristics as described above.

[0081] The present invention also provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, the steps of the ridge line extraction method reflecting the landform characteristics as described above are implemented.

[0082] The above-mentioned specific implementation methods are used to explain the present invention and are only preferred embodiments of the present invention, rather than limiting the present invention. Any modifications, equivalent substitutions, improvements, etc. made to the present invention within the spirit of the present invention and the protection scope of the claims shall fall within the protection scope of the present invention.

Claims

1. A ridgeline extraction method reflecting landform features, characterized in that: The method comprises the following steps: S1. Determine the mountainous area to be studied and extract the digital elevation model elevation data of the real terrain within the mountainous area to be studied; S2, resampling the elevation data in step S1 with 2 times the precision and arranging them into a matrix containing only elevation data; S3, using different wavelet functions and decomposition levels to perform wavelet decomposition on the matrix of the elevation data in step S2, and reconstructing the low-frequency terms in the wavelet decomposition result; S4, comprehensively evaluating the reconstruction result of the low-frequency term in step S3 according to the signal reconstruction error and the degree of noise reduction, and determining the optimal wavelet function and decomposition level used for wavelet decomposition; S5, using the optimal wavelet function and decomposition layer number in step S4, filtering and denoising the matrix of the elevation data in step S2; S6. Using the eight-neighborhood relative elevation discrimination method, extract terrain ridge lines from the elevation data matrix after noise reduction in step S5.

2. The ridgeline extraction method reflecting landform characteristics according to claim 1 is characterized in that: In step S2, according to task requirements, the elevation data matrix of the entire area is decomposed into sub-areas of uniform size, and ridge line extraction is implemented in the sub-areas.

3. The ridgeline extraction method reflecting landform features according to claim 1, characterized in that: In step S3, a two-dimensional discrete wavelet transform is used to decompose the elevation data matrix in both horizontal and vertical directions.

4. The method for extracting ridgelines reflecting landform features according to claim 1, characterized in that: In step S4, the reconstruction error and the degree of denoising are important indicators for measuring whether the terrain denoising is successful. The chamfer distance is used to measure the reconstruction error d CD , and its calculation formula is as follows: Where S1 and S2 represent the elevation point clouds before and after reconstruction respectively; the first term It represents the average value of the sum of the minimum square distances from any point in the elevation point cloud S1 before reconstruction to the elevation point cloud S2 after reconstruction. The second term It represents the average value of the sum of the minimum square distances from any point in the reconstructed elevation point cloud S2 to the elevation point cloud S1 before reconstruction; Secondly, the terrain roughness index TRI is used to measure the degree of noise reduction to prevent the reconstruction error from being too small, resulting in incomplete terrain noise reduction. The calculation formula is as follows: Where: N is the window length, Z0 is the elevation of the center position, Z k is the elevation of the neighboring location.

5. The method for extracting ridgelines reflecting landform features according to claim 1, characterized in that: In step S4, based on the reconstruction error, which is the primary denoising criterion, the db8 wavelet is selected as the optimal wavelet function for terrain denoising. When the denoising degree and the reconstruction error are balanced, the optimal number of wavelet decompositions can be determined. It should be noted that the optimal number of wavelet decompositions is related to the size of the analysis area in step S2.

6. The method for extracting ridgelines reflecting landform features according to claim 1, characterized in that: In step S6, the relative elevations of the eight neighborhoods around the center point are determined. When at least five neighborhoods are lower than the center point, the center point can be confirmed to be the ridgeline position.

7. A ridgeline extraction system reflecting landform features, characterized in that: The system includes the following modules: - an elevation data extraction module, the elevation data extraction module is used to extract the digital elevation model elevation data of the real terrain within the research mountain area according to the determined research mountain area; - a matrix arrangement module, which is used to resample the elevation data extracted by the elevation data extraction module with multiple times of precision and arrange them into a matrix containing only elevation data; - a low-frequency term reconstruction module, which is used to use different wavelet functions and decomposition levels to perform wavelet decomposition on the matrix of elevation data arranged by the matrix arrangement module, and reconstruct the low-frequency terms in the wavelet decomposition result; - An optimal wavelet function and decomposition layer number determination module, which is used to comprehensively evaluate the low-frequency term reconstruction result of the low-frequency term reconstruction module according to the signal reconstruction error and the degree of noise reduction, and determine the optimal wavelet function and decomposition layer number used for wavelet decomposition; - A filtering and denoising module, wherein the filtering and denoising module is used to filter and denoise the matrix of the elevation data arranged by the matrix arrangement module by using the optimal wavelet function and the number of decomposition layers determined by the optimal wavelet function and the number of decomposition layers determination module; - A terrain ridgeline extraction module, which is used to extract terrain ridgelines from the elevation data matrix after denoising by the filtering and denoising module by using an eight-neighborhood relative elevation discrimination method.

8. An electronic device, comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus, characterized in that: a memory for storing a computer program, A processor, wherein the processor is used to execute a computer program stored in a memory to implement the steps of a ridgeline extraction method reflecting landform features as described in any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the ridgeline extraction method reflecting landform features as described in any one of claims 1 to 7 are implemented.

Images