DEM flying point detection method based on connected domain judgment

Abstract

The present invention provides a DEM flying point detection method based on connected domain judgment. Using a breadth-first algorithm, starting from the top left corner of the map, the method traverses each grid point along the grid until the minimum elevation difference between all grid points and adjacent surrounding grid points is calculated. This method then performs connected domain statistical analysis and flying point calculation based on the calculated minimum grid difference, compared with a set threshold. The connected domain with the largest area, i.e., the data normal value, is excluded, and all grid points in the connected domains other than this are considered flying points. The present invention uses an algorithm to check flying point determination on batches of maps, not only ensuring the accuracy of flying point determination results, but also improving inspection efficiency and saving a significant amount of manpower and time. In actual application scenarios, the method has also been used on a small scale and has been unanimously recognized by users.

Description

Technical Field

[0001] The present invention relates to the field of geographic information surveying and mapping, and is used for quality inspection and acceptance of digital elevation model (DEM) data. It is used to automatically perform flying point detection statistical analysis with high precision and high efficiency, and meet the high-precision quality control requirements of DEM data quality inspection and acceptance. Background Art

[0002] Traditional flying spot detection methods mainly include manual visual inspection, shading map inspection, contour line fitting inspection and neighborhood statistical inspection.

[0003] 1) Manual visual inspection: This involves visually inspecting topographic maps or contour maps generated from DEM data to identify elevation anomalies. This method is inefficient for large-scale DEM data and is susceptible to subjective factors, resulting in unstable detection results and impractical for large-scale application.

[0004] 2) Shaded Image Inspection: This method uses shading technology to highlight three-dimensional data features in a two-dimensional perspective, overlaying DOM and DLG data. Using methods such as "rolling curtains" and "transparency," visual inspection is used to identify terrain transition anomalies in the DEM data. This method is inefficient, insensitive to subtle anomalies, prone to missed detections, and requires a combination of manual and visual inspection, making it unsuitable for large-scale application.

[0005] 3) Contour Fit Check: Contour lines generated from the DEM data are fitted and compared with contour lines on the original map or other terrain representations. The degree of conformity of the contour lines is checked to determine whether the DEM data contains elevation anomalies. This method requires the original map as a reference, making it difficult to obtain high-quality original data in some areas. In areas with complex and variable terrain, contour fit checks may be affected by factors such as topography and vegetation cover, leading to misjudgments.

[0006] 4) Neighborhood Statistics Check: Calculates statistics such as the mean, median, and standard deviation for each point's neighborhood, and identifies outliers based on a set threshold. This method is suitable for preliminary screening of large-scale DEM data, but detection results are significantly affected by factors such as terrain complexity and data source quality. It is sensitive to the assumption of normality in DEM data, making it difficult to distinguish true terrain features from elevation outliers. Furthermore, this method cannot detect consecutive outliers, making it prone to missed detections.

[0007] In summary, traditional flying point detection methods play a certain role in DEM data quality control, but they have many limitations, such as low detection efficiency, poor reliability, insensitivity to subtle anomalies, and missed consecutive flying points. Therefore, exploring an efficient, accurate, and practical flying point detection method is of great significance for improving DEM data quality and ensuring the accuracy of subsequent applications. Summary of the Invention

[0008] To overcome the shortcomings of the prior art, the present invention provides a DEM flying point detection method based on connected domain judgment. Drawing on the concept of connected components in graph theory, a DEM flying point detection method based on connected domain judgment is proposed, drawing on the insights of traditional flying point detection methods. This method uses a breadth-first algorithm, starting from the upper left corner of the map, to traverse each grid point along the grid until the minimum elevation difference between all grid points and their adjacent surrounding grid points is calculated. Because each grid point needs to be traversed, the minimum elevation difference between each grid point must be calculated based on the original grid elevation value. This calculated minimum grid difference is then compared with a set threshold to perform connected domain statistical analysis and flying point calculation. The connected domain with the largest area, i.e., the data normal value, is excluded. All grid points in the connected domains other than this connected domain are then considered flying points. This method effectively identifies flying points in DEM data, improving detection efficiency and accuracy, and providing a new technical approach for DEM data preprocessing and quality control.

[0009] The specific steps of the technical solution adopted by the present invention to solve its technical problem are:

[0010] Step 1: Calculate the minimum elevation difference of 8 neighborhoods;

[0011] Calculate the minimum elevation difference of the 8 neighborhoods of the grid point. The connectivity determination principle is: take point P as c As the center, calculate P c The 8 adjacent points The elevation difference, 1≤x≤8, if P c and If the difference is less than or equal to the threshold Δ, then P c and Under the threshold Δ, it is connected. Otherwise, if P c and If the difference is greater than the threshold Δ, it is considered that P c and It is disconnected under the threshold Δ;

[0012] Step 2: Solve the connected domain based on the threshold;

[0013] Using the DEM data as the grid point elevation difference matrix and the threshold Δ as the judgment value, according to the connectivity judgment principle of step 1, each connected domain in the grid point elevation difference matrix is ​​solved using a breadth-first or depth-first approach, and the number of connected domains under the threshold Δ and the number of points contained in each connected domain are counted;

[0014] Step 3: Baseline threshold selection;

[0015] Starting from the threshold Δ=2, we get the number of connected domains and the number of points contained in each connected domain when the threshold is Δ. Under a certain threshold Δ, if the number of grid points contained in the connected domain with the largest area is less than a%, then the threshold Δ is increased by 1. Repeat steps 1 to 2 until the number of grid points contained in the connected domain with the largest area accounts for more than a% of the entire map. The threshold at this time is the benchmark threshold Δ. min , except for the points included in this connected domain, all other points are considered as abnormal points;

[0016] Step 4: Statistical analysis of outliers;

[0017] Based on the reference threshold Δ min At the beginning, the threshold is increased by 1 each time, and steps 1 to 2 are repeated. The number of connected domains and the number of abnormal points under each threshold are counted until there is only one connected domain in the entire image. At this time, the corresponding threshold is Δmax;

[0018] Step 5: Flying point determination;

[0019] The statistical threshold is Δ min ~Δ max When the range is set, the inspection results of the number of connected domains and the number of abnormal points corresponding to each threshold are obtained, and the proportion of each number of connected domains and the number of abnormal points to the total number of points is statistically analyzed to determine the flying points.

[0020] The percentage a% is set to 99%.

[0021] The specific judgment principles in step five are:

[0022] The specific method for determining flying points based on the ratio of abnormal points to the total number of points is as follows:

[0023] According to the threshold Δ min ~Δ max The ratio of abnormal points to the total number of points corresponding to each threshold is calculated. The abnormal points with a ratio less than the specified ratio are determined as flying points. The specified ratio is one ten-thousandth of the number of abnormal points to the total number of points. For example, in Table 1, when the threshold Δ = 10, the ratio of abnormal points to the total number of points is 0.0376‰, which is less than one ten-thousandth, that is, 0.0376‰ < 0.1‰, so the 24 abnormal points under the threshold Δ = 10 are all flying points.

[0024] The specific judgment principles in step five are:

[0025] The specific method is to determine the flying point based on the threshold value set by manual experience. The threshold range for determining abnormal points as flying points is summarized based on manual experience. The threshold range is different according to the different map terrain. The specific terrain type and threshold range are: the elevation difference threshold in plain terrain is 0-2, the elevation difference threshold in hilly terrain is 3-6, the elevation difference threshold in mountainous terrain is 7-25, and the elevation difference threshold in high mountainous terrain is 26-90. Points outside this elevation difference threshold range are flying points.

[0026] An electronic device comprises: one or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs are configured to execute the method described above.

[0027] A computer-readable storage medium stores program code, which can be called by a processor to execute the method described above.

[0028] The beneficial effect of the present invention lies in its applicability to all existing terrain types. For example, in flat terrain, traditional manual empirical inspections have difficulty detecting subtle differences in flat terrain. However, the eight-neighborhood minimum elevation difference in this algorithm can calculate the outliers within the minimum range of eight neighborhoods, thereby greatly improving inspection accuracy. Furthermore, during the experimental process, for batch maps of the same terrain, such as 128 experimental data sheets, after setting a specified threshold or specifying the proportion of outliers to the total number of points, the algorithm inspected each batch of maps one by one, ultimately obtaining a statistical analysis result of the number of flying points. Using the algorithm to inspect flying points in batches of maps not only ensures the accuracy of the flying point determination results, but also improves inspection efficiency, saving a significant amount of manpower and time. In actual application scenarios, it has also been used on a small scale and has been unanimously recognized by users. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 To calculate the minimum elevation difference diagram of 8 neighborhoods, Figure 1 (a) is to calculate the minimum elevation difference map of 8 neighborhoods, Figure 1 (b) is the 8-neighborhood connectivity graph.

[0030] Figure 2 This is a schematic diagram of the minimum elevation difference matrix connected domain algorithm. Figure 2 (a) is the elevation value map of DEM data, Figure 2 (b) is the minimum elevation difference matrix diagram, Figure 2 (c) is a schematic diagram of the connected domain.

[0031] Figure 3 The flowchart for determining flying points based on the set threshold and the proportion of abnormal points.

[0032] Figure 4 This is the test effect diagram of abnormal points in high mountains.

[0033] Figure 5 This is a graph showing the relationship between the high mountain threshold and the ratio of outlier points to the total number of points.

[0034] Figure 6 This is the test effect diagram of mountain anomaly points.

[0035] Figure 7 This is a graph showing the relationship between the mountain threshold and the ratio of outliers to the total number of points.

[0036] Figure 8 This is the test effect diagram of hilly outlier points.

[0037] Figure 9 The relationship between the hill threshold and the ratio of outliers to the total number of points is shown in Figure 2.

[0038] Figure 10 This is the test effect diagram of flat ground anomaly points.

[0039] Figure 11 This is a graph showing the relationship between the flat ground threshold and the ratio of outliers to the total number of points. DETAILED DESCRIPTION

[0040] The present invention will be further described below with reference to the accompanying drawings and examples.

[0041] The specific steps of the DEM flying point detection method based on 8-neighborhood connected domain are as follows:

[0042] Step 1: Calculate the minimum elevation difference of the 8 neighborhoods

[0043] Calculate the minimum elevation difference of the 8-neighborhood grid point. The connectivity determination principle is: Figure 1 As shown in (a), point P c As the center, calculate P c The 8 adjacent points The elevation difference, 1≤x≤8, if P c and If the difference is less than or equal to the threshold Δ, then P c and Under the threshold Δ, it is connected. Otherwise, if P c and If the difference is greater than the threshold Δ, it is considered that P c and It is disconnected under the threshold Δ, such as Figure 1 In (b), green is used to represent the center point P c is connected, red indicates disconnected, assuming P c Under the threshold Δ is connected, and then find P c and The minimum value of the elevation difference between the three points;

[0044] Step 2: Solve the connected domain based on the threshold

[0045] Take the DEM data as the grid point elevation difference matrix, take the threshold Δ as the judgment value, and solve each connected domain in the grid point elevation difference matrix using the breadth-first or depth-first method according to the connectivity judgment principle in step 1, and count the number of connected domains under the threshold Δ and the number of points contained in each connected domain. Figure 2 (a) shows the DEM data. The threshold Δ=5 is set to solve the connected domain of the DEM data. The solution process is as follows: Figure 2 As shown in (b), Figure 2 (c) The final connected domain result graph is given. When the threshold is 5, there are 3 connected domains. Connected domain 1 (green area) includes 9 grid points, connected domain 2 (purple area) includes 13 grid points, and connected domain 3 (orange area) includes 3 grid points.

[0046] Step 3: Baseline threshold selection

[0047] Starting from the threshold Δ=2, we get the number of connected domains and the number of points contained in each connected domain when the threshold is Δ. Under a certain threshold Δ, if the number of grid points contained in the connected domain with the largest area is less than 99%, then the threshold Δ is increased by 1. Repeat steps 1 to 2 until the number of grid points contained in the connected domain with the largest area accounts for more than 99% of the entire map. The threshold at this time is the benchmark threshold Δ. min , except for the points included in this connected domain, all other points are considered as abnormal points;

[0048] Step 4: Statistical analysis of outliers

[0049] Based on the reference threshold Δ min At the beginning, the threshold is increased by 1 each time, and steps 1 to 2 are repeated. The number of connected domains and the number of abnormal points under each threshold are counted until there is only one connected domain in the entire image. At this time, the corresponding threshold is Δmax;

[0050] Step 5: Flying point determination

[0051] The statistical threshold is Δ min ~Δ max When the range is set, the inspection results of the number of connected domains and the number of abnormal points corresponding to each threshold are obtained, and the proportion of each number of connected domains and the number of abnormal points to the total number of points is statistically analyzed to determine the flying points.

[0052] The specific judgment principle is to choose one of the following two principles:

[0053] 1) Flying point determination based on the ratio of abnormal points to the total number of points. The specific determination method is as follows:

[0054] According to the threshold Δ min ~Δ max The ratio of abnormal points to the total number of points corresponding to each threshold is calculated. The abnormal points with a ratio less than the specified ratio are determined as flying points. The specified ratio is one ten-thousandth of the number of abnormal points to the total number of points. For example, in Table 1, when the threshold Δ = 10, the ratio of abnormal points to the total number of points is 0.0376‰, which is less than one ten-thousandth, that is, 0.0376‰ < 0.1‰, so the 24 abnormal points under the threshold Δ = 10 are all flying points.

[0055] 2) Flying point determination based on threshold setting based on manual experience. The specific method is based on manual experience, and the threshold range for determining abnormal points as flying points is summarized. The threshold range summarized varies depending on the terrain of the map. The specific terrain type and threshold range are: the elevation difference threshold in plain terrain is 0-2, the elevation difference threshold in hilly terrain is 3-6, the elevation difference threshold in mountainous terrain is 7-25, and the elevation difference threshold in high mountainous terrain is 26-90. Points outside this elevation difference threshold range are considered flying points.

[0056] The reference range of the threshold is set according to the terrain. For example, the artificial reference threshold range for mountainous terrain in Table 1 is 6-25. When the threshold Δ is set to 9, all 85 abnormal points under the current threshold are determined to be flying points.

[0057] To verify the effectiveness and accuracy of the proposed method, a test was conducted. DEM data was selected, an elevation difference threshold was set, and the program was run for detection. The results showed that the new method can detect flying points in DEM data, and the detection results are consistent with those of manual inspection.

[0058] Table 1 Statistics of the ratio of mountain threshold to abnormal points and total number of points

[0059]

[0060] As shown in Table 1, the threshold value of the DEM data test for abnormal points in this area is 6-14, which is consistent with the artificial threshold experience of 6-25 for the elevation value threshold difference in mountainous terrain; when the proportion of abnormal points is less than 0.1‰, it is determined to be a flying point, that is, when the threshold Δ=10, the abnormal points account for 0.04‰ of the total number of points, and all abnormal points within this range are determined to be flying points. The specific abnormal point test effect of mountainous terrain is shown in the figure below. Figure 6 The statistical analysis is shown in Figure 2 Figure 7 shown.

[0061] Table 2 Statistics of the ratio of hill threshold to abnormal points and total number of points

[0062]

[0063]

[0064] As shown in Table 2, the threshold value of the DEM data test for abnormal points in this area is 2-5, which is consistent with the artificial threshold experience of 2-6 for the elevation value threshold difference in hilly terrain. When the abnormal point ratio is less than 0.1‰, it is determined to be a flying point. That is, when the threshold Δ=3, the abnormal points account for 0.05‰ of the total number of points, and all abnormal points within this range are determined to be flying points. The specific abnormal point test effect diagram of hilly terrain is shown in the figure below. Figure 8 The statistical analysis is shown in Figure 2 Figure 9 shown.

[0065] Table 3 Statistics of the ratio of flat ground threshold to abnormal points and total number of points

[0066]

[0067] As shown in Table 3, the threshold value of the DEM data test for abnormal points in this area is 2-3. The artificial threshold experience of the elevation value difference in flat terrain is 0-2. At this time, the abnormal point with a threshold of 3 is determined to be a flying point. When the abnormal point ratio is less than 0.1‰, it is determined to be a flying point. That is, when the threshold Δ=2, the abnormal point accounts for 0.05‰ of the total number of points, and all abnormal points within this range are determined to be flying points. The specific abnormal point test effect of flat terrain is shown in the figure below. Figure 10 The statistical analysis is shown in Figure 2 Figure 11 shown.

[0068] Table 4 Statistics of the ratio of high mountain threshold to abnormal points and total number of points

[0069]

[0070]

[0071] As shown in Table 4, 1) when calculating the benchmark connected domain, if the threshold Δ=8, the connected domain that meets the maximum number of grid points accounts for more than 99% of the total number of points, and this is used as the benchmark connected domain.

[0072] 2) Check the threshold value set by manual experience and determine that the points beyond the threshold reference value range are flying points. For example, the elevation difference threshold value in high mountainous areas is generally 26-90. If the threshold Δ=26 is set, all abnormal points within this range are determined to be flying points.

[0073] 3) According to the DEM data quality inspection requirements, if the abnormal point is less than 0.1‰, it is determined to be a flying point. For example, when the threshold Δ=21, the abnormal point accounts for 0.08‰ of the total number of points, then all abnormal points within this range are determined to be flying points. Figure 4 The statistical analysis is shown in Figure 2 Figure 5 shown.

[0074] The above test results demonstrate that the present invention is applicable to all existing terrain types. For example, in flat terrain, traditional manual inspections struggle to detect subtle differences. However, the algorithm's eight-neighborhood minimum elevation difference can calculate outliers within the minimum eight-neighborhood range, significantly improving inspection accuracy. Furthermore, during testing, for batches of maps of the same terrain, such as 128 experimental data sheets, after setting a specified threshold or a specified ratio of outliers to the total number of points, the algorithm then inspected each batch of maps individually, ultimately generating a statistical analysis of the number of flying points. This algorithm's inspection of flying point determinations on batches of maps not only ensures the accuracy of flying point determination results but also improves inspection efficiency, saving significant manpower and time. The algorithm has also been used on a small scale in practical application scenarios, receiving unanimous approval from users.

[0075] When checking for flying points in a DEM image, depending on the elevation difference threshold, the area connected by the flying points can be considered a connected domain. Multiple flying points form multiple connected domains, each of which can be represented as an independent object or region. In software image processing, flying point analysis can be performed on the image to calculate features such as elevation difference, number of flying points, and flying point ratio, allowing for further processing and analysis.

[0076] Specifically, the connected domain analysis in this experiment includes the following:

[0077] 1: Image conversion: Convert DEM data into a visual grayscale image, such as Figure 4 、 Figure 6 、 Figure 8 、 Figure 10 Grayscale image of

[0078] 2: Outlier Marking: Traverse each pixel in the image and mark the pixels that are judged as outliers with special colors, such as Figure 4 、 Figure 6 The red part in the effect diagram represents the abnormal point connected domain formed by the abnormal points;

[0079] 3: Flying point rate analysis: For each connected domain, its features such as the number of connected domains and the number of abnormal points can be calculated, and a comprehensive analysis of the flying point rate, which is the ratio of abnormal points to the total number of points under different thresholds, can be performed.

[0080] Comprehensive test results, such as Figure 5 The relationship diagram of the high mountain threshold and the ratio of abnormal points to the total number of points is shown in Figure 2. The threshold Δ is between 25 and 90, or the ratio of abnormal points to the total number of points is less than 0.1‰, which are all considered as flying points. Figure 6 The relationship diagram between the mountain threshold and the ratio of abnormal points to the total number of points is shown in Figure 2. The threshold Δ=6~25 or the ratio of abnormal points to the total number of points is less than 0.1‰, which are all considered as flying points. Figure 7In the relationship diagram between the hill threshold and the ratio of abnormal points to the total number of points, the reference threshold Δ=2~6 or the ratio of abnormal points to the total number of points is less than 0.1‰, which is determined as a flying point; Figure 8 This is a relationship diagram between the flat ground threshold and the ratio of outliers to the total number of points. Points within the reference threshold range of Δ=2 or with a ratio of outliers to the total number of points less than 0.1‰ are all determined to be flying points.

Claims

1. A DEM flying point detection method based on connected domain judgment, characterized in that The steps include: Step 1: Calculate the minimum elevation difference of 8 neighborhoods; Calculate the minimum elevation difference of the 8 neighborhoods of the grid point. The connectivity determination principle is: take point P as c As the center, calculate P c The 8 adjacent points The elevation difference, 1≤x≤8, if P c and If the difference is less than or equal to the threshold Δ, then P c and Under the threshold Δ, it is connected. Otherwise, if P c and If the difference is greater than the threshold Δ, it is considered that P c and It is disconnected under the threshold Δ; Step 2: Solve the connected domain based on the threshold; Using the DEM data as the grid point elevation difference matrix and the threshold Δ as the judgment value, according to the connectivity judgment principle of step 1, each connected domain in the grid point elevation difference matrix is ​​solved using a breadth-first or depth-first approach, and the number of connected domains under the threshold Δ and the number of points contained in each connected domain are counted; Step 3: Baseline threshold selection; Starting from the threshold Δ=2, we get the number of connected domains and the number of points contained in each connected domain when the threshold is Δ. Under a certain threshold Δ, if the number of grid points contained in the connected domain with the largest area is less than a%, then the threshold Δ is increased by 1. Repeat steps 1 to 2 until the number of grid points contained in the connected domain with the largest area accounts for more than a% of the entire map. The threshold at this time is the benchmark threshold Δ. min , except for the points included in this connected domain, all other points are considered as abnormal points; Step 4: Statistical analysis of outliers; Based on the reference threshold Δ min At the beginning, the threshold is increased by 1 each time, and steps 1 to 2 are repeated. The number of connected domains and the number of abnormal points under each threshold are counted until there is only one connected domain in the entire image. At this time, the corresponding threshold is Δmax; Step 5: Flying point determination; The statistical threshold is Δ min ~Δ max When the range is set, the inspection results of the number of connected domains and the number of abnormal points corresponding to each threshold are obtained, and the proportion of each number of connected domains and the number of abnormal points to the total number of points is statistically analyzed to determine the flying points.

2. The DEM flying point detection method based on connected domain judgment according to claim 1 is characterized in that: The percentage a% is set to 99%.

3. The DEM flying point detection method based on connected domain judgment according to claim 1 is characterized in that: The specific judgment principles in step five are: The specific method for determining flying points based on the ratio of abnormal points to the total number of points is as follows: According to the threshold Δ min ~Δ max The ratio of abnormal points corresponding to each threshold to the total number of points is calculated, and abnormal points with a ratio less than the specified ratio are determined as flying points. The specified ratio is set to one ten-thousandth of the number of abnormal points to the total number of points.

4. The DEM flying point detection method based on connected domain judgment according to claim 1 is characterized in that: The specific judgment principles in step five are: The specific method of flying point determination based on threshold setting based on manual experience is to determine the threshold range for determining abnormal points as flying points based on manual experience. According to the different map terrains, the specific terrain types and threshold ranges are as follows: the elevation difference threshold in plain terrain is 0-2, the elevation difference threshold in hilly terrain is 3-6, the elevation difference threshold in mountainous terrain is 7-25, and the elevation difference threshold in high mountainous terrain is 26-90. Points outside this elevation difference threshold range are flying points.

5. An electronic device, characterized in that: include: one or more processors; Memory; One or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs are configured to perform the method according to any one of claims 1 to 4.

6. A computer-readable storage medium, characterized in that The computer-readable storage medium stores program code, and the program code can be called by a processor to execute the method according to any one of claims 1 to 4.

Images