Submerged analysis method and system based on image recognition

Abstract

The invention provides a submerging analysis method and system based on image recognition, and belongs to the technical field of image recognition, and the method specifically comprises the steps: carrying out the determination of comparison processing schemes in different sub-regions according to the analysis results of the elevation values in different sub-regions in a recommended analysis region through a submerging threshold value inputted by a user, based on an analysis result obtained by the comparison processing scheme, when it is determined that a submerging threshold value belongs to a risk threshold value, determining an amplification strategy of the recommended analysis region according to the comparison processing scheme of the sub-region and the ground feature element data in the adjacent region, and performing amplification processing on the recommended analysis region according to the amplification strategy to obtain an amplified region; and according to a comparative analysis result of the amplified region, carrying out generation processing of colors of pixel values to obtain an analysis result, generating a grid file based on the analysis result, outputting a drawing result as a high-resolution image file, and embedding geographic space coordinates, so that the reliability and comprehensiveness of submerging analysis processing are improved.

Description

Technical Field

[0001] This invention belongs to the field of image recognition technology, and in particular relates to a flooding analysis method and system based on image recognition. Background Technology

[0002] In the survey and design of water conservancy projects, inundation analysis and planning are crucial steps. Traditional inundation analysis methods require high accuracy of digital elevation model (DEM) information. For example, the invention patent application CN202411935367.2, "Early Warning Method for Water Inundation Area Based on Image Recognition," presents a similar technical solution. However, this solution lacks customization options and struggles to meet the practical needs of survey and design. Furthermore, the high-precision DEM data is classified and cannot meet the demands of complex field and office work environments.

[0003] Therefore, there is an urgent need for a flooding analysis method and system based on image recognition. Summary of the Invention

[0004] To achieve the objectives of this invention, the following technical solution is adopted: Specifically, this application provides a flooding analysis method based on image recognition, which includes: After loading the raster data and vector layers of the digital elevation model into the system, S1 allows users to select ground features through an interactive interface. Based on the elevation data of the area where the ground features are located, the system determines the recommended analysis area and the geometry of the clipping area for the ground features. S2 calculates the minimum bounding rectangle of the clipping area geometry and clips the elevation data to the bounding box range; S3 uses the user-input flooding threshold and determines the comparison processing scheme for different sub-regions based on the analysis results of the elevation values ​​in different sub-regions of the recommended analysis area. Based on the analysis results obtained from the comparison processing scheme, if it is determined that the flooding threshold belongs to the risk threshold, proceed to the next step. S4 uses the comparison processing scheme of the sub-region and the feature data of the adjacent region to determine the expansion strategy of the recommended analysis region. The expansion processing of the recommended analysis region is performed using the expansion strategy to obtain the expanded region. Based on the comparison analysis results of the expanded region, the color generation processing of the pixel values ​​is performed to obtain the analysis results. Based on the analysis results, a raster file is generated, and the drawing results are output as a high-resolution image file with embedded geospatial coordinates.

[0005] The beneficial effects of this invention are as follows: Based on the user-input flooding threshold, the system determines comparative processing schemes for different sub-regions according to the analysis results of elevation values ​​in different sub-regions within the recommended analysis area. This enables the determination of comparative processing schemes from the perspective of flooding risk in different sub-regions, thereby improving the efficiency of flooding risk identification and processing, while also ensuring the reliability of flooding analysis and processing in different sub-regions when the flooding risk is high.

[0006] By using the comparative processing scheme of sub-regions and the feature data of adjacent regions, the expansion strategy of the recommended analysis area is determined. This avoids the technical problem of poor reliability of the identification and processing results when the inundation risk is high, which is caused by performing inundation analysis and processing solely in the recommended analysis area. This further improves the reliability and comprehensiveness of the identification and processing of inundation risk.

[0007] Furthermore, the digital elevation model is a model that achieves digital simulation of ground topography and features using limited terrain elevation data.

[0008] Furthermore, the aforementioned land features are objects represented by land feature symbols on topographic maps, specifically including residential areas, industrial and mining enterprise buildings, public facilities, independent land features, roads and their ancillary facilities, pipelines, water systems and their ancillary facilities.

[0009] Furthermore, the elevation data includes the elevation values ​​of different pixels in the digital elevation model.

[0010] Furthermore, the method for determining the recommended analysis area for the aforementioned land feature elements is as follows: Based on the elevation data of the area where the feature is located, determine the elevation values ​​of different pixels in the area where the feature is located; Based on the elevation values ​​of different pixels, determine the deviation between the elevation values ​​of different pixels; Based on the aforementioned deviations, a recommended analysis area for the aforementioned land feature is determined.

[0011] Furthermore, determining that the flooding threshold belongs to the risk threshold specifically includes: Based on the analysis results obtained from the aforementioned comparative processing scheme, the area of ​​the flooded region in different sub-regions under the flooding threshold is determined; Based on the area proportion of the flooded area in different sub-regions, the flood risk coefficient of different sub-regions is determined; Based on the flooding risk coefficient, determine whether the flooding threshold belongs to the risk threshold.

[0012] In a second aspect, the present invention provides a computer system comprising: a memory and a processor connected in communication, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the above-described image recognition-based flooding analysis method when running the computer program.

[0013] Other features and advantages will be set forth in the following description, and the objects and other advantages of the invention are realized and obtained through the structures particularly pointed out in the description and the drawings.

[0014] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0015] The above and other features and advantages of the present invention will become more apparent from a detailed description of exemplary embodiments thereof with reference to the accompanying drawings.

[0016] Figure 1 This is a flowchart of a flooding analysis method based on image recognition; Figure 2 This is a flowchart illustrating the method for determining the recommended analysis area for geographic features; Figure 3 This is a flowchart illustrating the method for determining the comparison processing scheme for sub-regions; Figure 4 A flowchart for determining whether a flooding threshold falls under a risk threshold; Figure 5 It is a framework diagram of a computer system. Detailed Implementation

[0017] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0018] Example 1 like Figure 1 As shown, this application provides a flooding analysis method based on image recognition, specifically including: After loading the raster data and vector layers of the digital elevation model into the system, S1 allows users to select ground features through an interactive interface. Based on the elevation data of the area where the ground features are located, the system determines the recommended analysis area and the geometry of the clipping area for the ground features. Furthermore, the digital elevation model is a model that achieves digital simulation of ground topography and features using limited terrain elevation data.

[0019] Furthermore, the aforementioned land features are objects represented by land feature symbols on topographic maps, specifically including residential areas, industrial and mining enterprise buildings, public facilities, independent land features, roads and their ancillary facilities, pipelines, water systems and their ancillary facilities.

[0020] Furthermore, the elevation data includes the elevation values ​​of different pixels in the digital elevation model.

[0021] Specifically, such as Figure 2 As shown, the method for determining the recommended analysis area for the aforementioned land feature is as follows: Based on the elevation data of the area where the feature is located, determine the elevation values ​​of different pixels in the area where the feature is located; Elevation data: The ground elevation value represented by each pixel in a region, which is the basic data for describing terrain undulation.

[0022] By using the elevation values ​​of different pixels, the deviation rate of different pixels is determined, and based on the deviation rate of different pixels, the distribution deviation pixels among the pixels are determined. Deviation rate: The degree of deviation between the elevation value of a pixel and the average elevation value of all pixels within a certain range (such as a 3x3 window). Deviation rate = |(elevation value of the pixel - average elevation value of surrounding pixels)| / average elevation value of surrounding pixels. It quantifies the difference between the point and the surrounding terrain. The higher the deviation rate, the more unusual the terrain of the point is, which may be a special terrain such as a ridge, valley, steep slope, isolated peak, or depression.

[0023] Pixels with distribution deviation: Pixels with a deviation rate greater than a certain threshold (such as 2%) are "outliers" in the terrain, which are locations where the terrain has changed significantly.

[0024] Based on the distribution data of the pixels with the distribution deviation, the recommended analysis area for the ground features is determined.

[0025] Recommended analysis area: The area formed by the spatial aggregation of "distribution deviation pixels" is a region with drastic terrain changes and is therefore considered a recommended analysis area.

[0026] It is understood that the distribution deviation pixels are pixels with a deviation rate greater than a preset deviation rate threshold, and in one possible embodiment, pixels with a deviation rate greater than 2%.

[0027] Specifically, based on the distribution data of the distribution deviation pixels, the recommended analysis area for the land feature is determined, including: Using the aforementioned land feature as the center, the area where the land feature is located is divided into multiple sub-regions according to unit area; It should be noted that each of the sub-regions is a unit area centered on the aforementioned land feature.

[0028] Sub-region division: Using the ground features as the center, the entire region is divided into multiple concentric rings (fans or rings) of equal area to analyze the distribution of terrain anomalies (distribution deviation pixels) at different distances.

[0029] Based on the distribution data of feature pixels with distribution deviation in different sub-regions, the number of feature pixels with distribution deviation in different sub-regions is determined.

[0030] The number of pixels with distribution deviation meets the requirement: within a certain sub-region, the proportion of pixels with distribution deviation to the total number of pixels in that sub-region is greater than a preset proportion threshold (e.g., 0.2). If the proportion is greater than the preset proportion threshold, it means that this sub-region is a "terrain anomaly area".

[0031] Preset percentage threshold: A critical proportion for determining whether "topographic anomaly areas" are prevalent. If the proportion of the number of sub-regions of "topographic anomaly areas" exceeds this threshold, it indicates that the anomaly is a common phenomenon and the analysis scope needs to be expanded; if it does not exceed this threshold, it indicates that the anomaly is a local phenomenon and the analysis scope can be narrowed.

[0032] Preset scaling factor: Definition: A coefficient used to calculate the radius of the recommended analysis area. When anomalies are prevalent, the system recommends a circular area centered on the ground features with a radius equal to the unit length × preset scaling factor as the key analysis area.

[0033] Case 1: When the number of distribution deviation pixels in different sub-regions does not meet the requirements, the recommended analysis area for the feature is determined to be the area where the feature is located.

[0034] Case 2: When there is a sub-region where the number of pixels with distribution deviation meets the requirements, and the proportion of the number of sub-regions where the number of pixels with distribution deviation meets the requirements in the region where the feature is located is greater than a preset proportion threshold, then the region with the feature as the center and the product of the unit length and the preset proportion factor as the radius is taken as the recommended analysis region of the feature. In one possible embodiment, the preset proportion factor can be selected as 0.6.

[0035] Case 3: Alternatively, it can be understood that if the proportion of the number of sub-regions with a sufficient number of pixels with distribution deviation in the area where the feature is located is not greater than a preset proportion threshold, then the recommended analysis area for the feature is determined to be the area where the feature is located.

[0036] Step 1: Obtaining data Specifically, the surrounding 2000-meter range is divided into two concentric ring sub-regions (for simplification): Sub-region A (inner ring): 0 - 1000 meters range, Sub-region B (outer ring): 1000 - 2000 meters range.

[0037] Table 1 shows the total number of pixels and the number of "distribution deviation pixels" in each sub-region.

[0038] Step 2: Apply decision rules; Case 1 Judgment: Condition: Does the number of deviation points in all sub-regions not meet the requirement? Judgment: No, because sub-region A meets the requirement. Proceed to the next level.

[0039] Scenario 2 judgment: Operation: Calculate the percentage of sub-regions that meet the requirements for the number of deviation points. Number of sub-regions that meet the requirements: 1 (sub-region A). Total number of sub-regions: 2. Percentage = 1 / 2 = 0.5. Condition: Since this percentage (0.5) is not greater than the preset percentage threshold (0.5), proceed to the last layer.

[0040] Case 3 Decision: Condition: Since the proportion (0.5) is not greater than the preset proportion threshold (0.5), the decision is to determine the recommended analysis area as "the area where the land feature is located", that is, the entire 2000-meter radius study area by default.

[0041] The wisdom of this method lies in its ability to dynamically adjust the analysis scope based on the spatial distribution pattern of terrain anomalies: If the anomaly is very widespread, it may mean that the geological conditions of the entire area are complex. The system will narrow the scope (unit length × preset scaling factor) and recommend a core area for in-depth analysis to avoid losing focus due to an overly large scope. If the anomaly is local or moderate, the system will maintain or use the original large scope for analysis because either only the core area has a problem (requiring a large scope for comparison) or the problem is not concentrated and requires a global perspective.

[0042] S2 calculates the minimum bounding rectangle of the clipping area geometry and clips the elevation data to the bounding box range; S3 uses the user-input inundation threshold and determines different comparison processing schemes for different sub-regions within the recommended analysis area based on the analysis results of elevation values ​​in different sub-regions. Based on the analysis results obtained from the comparison processing schemes, if the inundation threshold is determined to be a risk threshold, an expansion strategy for the recommended analysis area is determined using the comparison processing schemes for the sub-regions and the feature data in adjacent areas. The expanded area is obtained by performing expansion processing on the recommended analysis area using the expansion strategy. Based on the comparison analysis results of the expanded area, color generation processing of pixel values ​​is performed to obtain the analysis results. A raster file is generated based on the analysis results, and the rendering results are output as a high-resolution image file with embedded geospatial coordinates.

[0043] When faced with multiple sub-regions at risk of inundation, the system intelligently selects inundation risk simulation strategies with varying levels of precision and computational cost based on their distance from ground features and the degree of internal topographic relief.

[0044] Specifically, such as Figure 3 As shown, the method for determining the comparison processing scheme for the sub-region is as follows: Based on the analysis results of the elevation values ​​in the sub-region, the average elevation value of different pixels in the sub-region is determined and used as the reference elevation value of the sub-region. Reference elevation value: The average elevation value of all pixels within a sub-region, representing the overall altitude of that sub-region.

[0045] Based on the elevation data of different pixels in different sub-regions and the reference elevation value of the sub-regions, the reference deviation rate of different pixels in the sub-regions is determined; Reference deviation rate: The degree of deviation between the elevation value of a single pixel within a sub-region and the reference elevation value of that sub-region. The calculation formula is: Reference deviation rate = |(pixel elevation - reference elevation value)| / reference elevation value. It measures the degree of terrain undulation within the sub-region. The larger the value, the more fragmented and undulating the terrain.

[0046] Based on the deviation between the reference elevation value and the inundation threshold in the sub-region, the inundation risk sub-regions are determined, and the inundation risk sub-regions are sorted from largest to smallest according to the interval distance of the inundation risk sub-regions to obtain the sorting result; Sub-regions with flood risk that have an absolute difference between their interval distance and the next flood risk sub-region in the sorting result and a preset distance threshold are designated as interval regions. The comparison processing scheme for the sub-regions is determined based on the deviation between the interval distance of the sub-regions and the interval distance of the interval regions, as well as the reference deviation rate of different pixels.

[0047] Sub-regions at risk of inundation: Sub-regions whose reference elevation values ​​are below a set inundation threshold (such as flood level). Interval distance: The shortest distance between sub-regions at risk of inundation and ground features (targets that need to be protected).

[0048] Interval Zone: In a list of flood risk sub-regions sorted by interval distance from largest to smallest, the first sub-region whose interval distance difference with the next sub-region is exceptionally large. It serves as a dividing point to distinguish between "far-risk zones" and "near-risk zones".

[0049] Background: To protect an important substation (geographic feature), it is necessary to simulate the surrounding area that may be at risk of flooding. The known flood inundation threshold is 100 meters. System parameters: preset distance threshold: 500 meters (used to identify "interval zones"), preset reference deviation rate threshold: 0.1 (i.e., 10%).

[0050] The system identified five flood-risk sub-regions (all with reference elevations < 100 meters), calculated their distances from the substation, and then sorted them from largest to smallest: Table 2 Sorting Results

[0051] Calculate the difference in distance between adjacent areas: Difference 1 (ZY): 2500 - 1800 = 700 meters, Difference 2 (YX): 1800 - 400 = 1400 meters, Difference 3 (XW): 400 - 350 = 50 meters, Difference 4 (WV): 350 - 300 = 50 meters; Identifying the "interval zone": The difference 2 (1400 meters) > the preset distance threshold (500 meters). Therefore, region Y is determined as the "interval zone". The boundary is as follows: regions Z and Y belong to the "far risk zone", while regions X, W, and V belong to the "near risk zone".

[0052] In one of the cases: if the interval distance of the sub-region is greater than the interval distance of the interval region, then because the interval distance between the flood risk area and the ground features is longer, the comparison processing scheme of the sub-region is determined to be that no flood risk simulation processing is required.

[0053] In the above steps, for the distant risk areas (regions Z and Y), the decision logic is as follows: the distances between these areas and the substation (2500 meters and 1800 meters, respectively) are both greater than the distance between these areas and region Y (1800 meters). Due to the great distance, the flood threat is relatively indirect and has a lower priority. The final solution is to determine their comparative treatment as "no need for flood risk simulation processing". The system will save computing resources and will not simulate these areas for the time being.

[0054] In another case: if the interval distance of the sub-region is not greater than the interval distance of the interval region, then the reference deviation rate of different pixels in the sub-region is determined. When the average value of the reference deviation rate of different pixels in the sub-region is greater than the preset reference deviation rate threshold, then when the sub-region is subjected to flooding risk simulation processing, the flooding risk simulation processing is performed by pixel-by-pixel comparison.

[0055] In the above steps, for the near-risk zones (regions X, W, and V), the distance between these zones and the substation is no greater than the distance between these zones and the substation. Therefore, it is necessary to further determine the simulation accuracy based on the degree of terrain undulation within each zone. The system calculated the average reference deviation rate within each zone.

[0056] Region X: The terrain in this region is highly undulating, with an average reference deviation rate of 0.15. Judgment: 0.15 > preset reference deviation rate threshold (0.1). Decision: The terrain is complex, requiring high-precision simulation. The comparative processing scheme is determined to be "simulation processing of flooding risk through pixel-by-pixel comparison." This means the system will calculate water flow pixel by pixel, achieving the highest accuracy but also incurring the greatest computational burden.

[0057] In another possible scenario, when the average reference deviation rate of different pixels in the sub-region is not greater than a preset reference deviation rate threshold, it is determined that when the sub-region is subjected to flooding risk simulation processing, the flooding risk simulation processing is performed simultaneously based on the target number of pixels.

[0058] Region W: The terrain within this region is relatively flat, with an average reference deviation rate of 0.08. Judgment: 0.08 ≤ preset reference deviation rate threshold (0.1). Decision: The terrain is flat, making a more efficient method suitable. The comparative processing scheme is determined to be "based on the target number of pixels, simultaneously performing simulation processing of flooding risk." This is a batch processing method, for example, calculating 4 or 9 pixels as a unit, significantly improving simulation speed while ensuring a certain level of accuracy.

[0059] Region V: The terrain inside this region is also relatively flat, and the average reference deviation rate is 0.06. Judgment: 0.06 ≤ preset reference deviation rate threshold (0.1). Decision: Same as Region W, adopt "based on the target number of pixels, and simultaneously carry out the simulation processing of flooding risk".

[0060] Through this method, the system achieves intelligent and hierarchical risk simulation resource allocation: remote areas are directly ignored to save resources; nearby areas with complex terrain are simulated pixel by pixel with high precision and high consumption to ensure the accuracy of core risk point assessment; nearby areas with flat terrain are simulated in batches with medium precision and high efficiency to quickly complete large-scale assessments.

[0061] This strategy ensures that limited computing resources are used effectively, comprehensively assessing risks while maximizing overall analysis efficiency.

[0062] Specifically, such as Figure 4 As shown, determining that the flooding threshold belongs to the risk threshold specifically includes: Based on the analysis results obtained from the aforementioned comparative processing scheme, the area of ​​the flooded region in different sub-regions under the flooding threshold is determined; Based on the area proportion of the flooded area in different sub-regions, the flood risk coefficient of different sub-regions is determined; Area percentage of flooded region: The percentage of the total area of ​​pixels that are simulated to be flooded within a sub-region, which directly reflects the severity of the disaster in that sub-region under the current flooding threshold.

[0063] Inundation Risk Coefficient: A composite risk indicator that integrates the "degree of damage to the sub-region itself" and the "impact of the surrounding areas." It considers not only the area proportion of the sub-region itself but also the inundation situation of its adjacent sub-regions. If both the sub-region and its surrounding areas are severely affected, it indicates that the risk is contiguous and systemic, and its risk coefficient is higher, measuring the true risk level of a sub-region within the overall context.

[0064] Based on the flooding risk coefficient, determine whether the flooding threshold belongs to the risk threshold.

[0065] It is understandable that when the average flooding risk coefficient in different sub-regions is greater than the preset risk coefficient threshold, the flooding threshold is determined to be a risk threshold.

[0066] It should be noted that the adjacent region is a sub-region whose distance from the sub-region meets the requirements, that is, a sub-region whose pixels overlap with the boundary of the sub-region.

[0067] When it is necessary to evaluate the threshold of "Option C: water level reaches 2 meters" to determine whether it falls under the "risk threshold", the preset risk coefficient threshold is 0.6. Step 1: Simulate, analyze, and calculate the flooding risk coefficient for each sub-region; The system simulated flooding in multiple sub-regions divided by a grid and calculated the flooding risk coefficient for each sub-region. For simplicity, we focus on four adjacent sub-regions.

[0068] Table 3 Calculation results for the four sub-regions

[0069] Calculation Explanation: To simplify the calculation, the risk coefficient in the table above uses the formula: Risk Coefficient = Percentage of its own inundated area × Average percentage of inundated area of ​​adjacent areas. Other formulas can also be used, but the core idea is that the higher the degree of inundation of itself and its surroundings, the greater the risk coefficient.

[0070] Step 2: Calculate the global mean and determine the result; Calculate the global mean: average the risk coefficients of all sub-regions. Assume the average flooding risk coefficient for the entire city is 0.68.

[0071] Application of decision rules: Condition: Global mean (0.3) > preset risk coefficient threshold (0.1), final decision: determine that "flooding threshold of 2 meters" belongs to the risk threshold.

[0072] Below the 2-meter threshold: only scattered, unconnected areas may be flooded, and the disaster is localized and controllable. Once the 2-meter "risk threshold" is reached: the flood will connect into a large area, forming a wide flood zone (such as areas 1, 2, and 4 connected into one). This method elevates the physical results of flood simulation into decision support information with clear management significance, truly realizing the leap from "data" to "insight".

[0073] Specifically, the method for determining the amplification strategy for the recommended analysis region is as follows: The objective is to intelligently determine whether to expand the scope of the analysis after completing the initial flooding risk analysis, in order to ensure the comprehensiveness and reliability of the risk assessment. Its core considerations are the "credibility of the initial analysis" and the "importance of surrounding features."

[0074] Based on the comparison processing scheme of the sub-regions, the sub-regions to be compared and analyzed are determined and used as the analysis sub-regions. Based on the comparison processing scheme of the analysis sub-regions, the identification reliability type of the recommended analysis region is determined. Analysis sub-regions: those sub-regions that have undergone flood risk simulation processing, identify reliability type: reliable analysis region: preliminary analysis results have high credibility, deviation analysis region: preliminary analysis results have low credibility, and the scope needs to be expanded for verification.

[0075] Based on the distribution data of ground features in the adjacent areas of the analysis sub-region, determine the number of ground features in the adjacent areas; The expansion strategy for the recommended analysis area is determined based on the number of land features in the adjacent areas and the reliability of the identification of the recommended analysis area.

[0076] Expansion strategy: Decide whether to include adjacent regions of the analyzed sub-region in the next step of flooding risk identification and processing.

[0077] Preset quantity range: The range of the number of pixel-by-pixel analysis sub-regions used to determine the "reliable analysis area". For example, at least 3 and at most 6 sub-regions use high-precision pixel-by-pixel analysis. Preset risk coefficient threshold: 0.7.

[0078] Case 1: The reliability type of the recommended analysis region is determined based on the number of sub-regions analyzed by performing flood risk simulation processing in a pixel-by-pixel comparison manner. In one possible embodiment, if the number of sub-regions analyzed by performing flood risk simulation processing in a pixel-by-pixel comparison manner is within a preset range, then the reliability type of the recommended analysis region is determined to be a reliable analysis region; otherwise, it belongs to a deviation analysis region.

[0079] The system checks how many sub-regions within the "New Urban Area" use a high-precision "pixel-by-pixel comparison" simulation scheme. Assuming the check results are as follows: Scenario A (Reliable Analysis Area): 4 sub-regions used pixel-by-pixel analysis. Judgment: 4 is greater than 3. Conclusion: Reliable type identification = Reliable analysis area.

[0080] Scenario B (deviation analysis area): Only 2 sub-regions used pixel-by-pixel analysis (or all used batch processing), judgment: 2 is not greater than 3, conclusion: reliable identification type = deviation analysis area.

[0081] It should be noted that when the analysis area is within the deviation analysis area, the adjacent areas of the analysis sub-region need to be expanded to achieve the identification and processing of flooding risk.

[0082] Scenario 1: When it falls within the "deviation analysis area" (Scenario B), the decision logic is as follows: Due to insufficient credibility of the preliminary analysis (too few high-precision analysis areas), the results may be inaccurate. To obtain a reliable global assessment, the analysis scope must be expanded for cross-validation. The final expansion strategy is to expand all adjacent areas of the "new urban area" and include them all in the next round of flood risk identification.

[0083] It should also be noted that when the sub-region is not part of the deviation analysis area, if the flooding risk coefficient of the analysis sub-region is less than the preset risk coefficient threshold, then the adjacent areas of the analysis sub-region do not need to be amplified.

[0084] Case 2: If the flooding risk coefficient of the analyzed sub-region is not less than the preset risk coefficient threshold, and if there are ground features in the adjacent areas of the analyzed sub-region, then it is determined that the adjacent areas need to be amplified.

[0085] Case 3: If there are no land features in the adjacent area, then it is determined that the adjacent area does not need to be expanded.

[0086] Scenario 2: When it belongs to the "reliable analysis area" (Scenario A), the system enters a more refined decision-making level, requiring the examination of the specific risk value of each analysis sub-region and the situation of its adjacent regions. We assume that there are three analysis sub-regions A1, A2, and A3 within the new urban area.

[0087] Table 4. Amplification processing strategies for different regions

[0088] For sub-region A1: the risk coefficient (0.5) is less than the threshold (0.7), indicating that the flood risk itself is low. Therefore, regardless of what is in the adjacent regions, there is no need for amplification analysis for the time being.

[0089] For sub-region A2: the risk coefficient (0.8) is very high, and there are important residential areas adjacent to it. This means that flooding is very likely to spread from A2 to the residential areas, causing serious consequences. Therefore, an expanded analysis of this adjacent residential area is necessary.

[0090] For sub-region A3: the risk coefficient (0.9) is very high, but its adjacent area is wasteland with no features requiring special protection. Therefore, even if the flood spreads, the damage will be limited, and from a management cost perspective, expansion can be temporarily suspended.

[0091] This amplification strategy determination method embodies a "cost-benefit analysis based on risk and value": First, assess the quality (credibility) of your own work: If the quality is poor, amplify all areas to make up for the lack of confidence. Under the premise that the quality of your own work is high: Focus only on high-risk areas: For low-risk areas, do not waste resources and focus only on valuable spread directions: For high-risk areas, only amplify the analysis in the direction where there are important features (potential disaster-bearing bodies).

[0092] This approach ensures that limited computing and analytical resources are always allocated to the most dangerous and valuable aspects, avoiding indiscriminate expansion and achieving precise and intelligent risk management.

[0093] Example 2 Secondly, such as Figure 5 As shown, the present invention provides a computer system comprising: a memory and a processor connected in communication, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the above-described image recognition-based flooding analysis method when running the computer program.

[0094] Furthermore, the method for determining the comparison processing scheme for the sub-region is as follows: Based on the analysis results of the elevation values ​​in the sub-region, the average elevation value of different pixels in the sub-region is determined and used as the reference elevation value of the sub-region. Based on the elevation data of different pixels in different sub-regions and the reference elevation value of the sub-regions, the reference deviation rate of different pixels in the sub-regions is determined; Based on the reference deviation rate of different pixels in the sub-region and the deviation between the reference elevation value and the flooding threshold, a comparison processing scheme for the sub-region is determined.

[0095] It is understood that the absolute value of the difference between the elevation value of the pixel and the reference elevation value of the sub-region, and the ratio of the difference to the reference elevation value of the sub-region, are used as the reference deviation rate of the pixel.

[0096] It should be noted that when the reference elevation values ​​of different sub-regions are all greater than the flooding threshold, different sub-regions are at risk of flooding during the simulation. Therefore, the comparison processing scheme for the sub-regions is determined to be based on the target number of pixels, while simultaneously performing simulation processing of flooding risk.

[0097] In one possible embodiment, the number of targets is 10, thereby improving the efficiency of flood risk analysis and processing when the flood risk is relatively low.

[0098] Furthermore, when there is a sub-region where the reference elevation value is not greater than the inundation threshold, the sub-region where the reference elevation value is not greater than the inundation threshold is regarded as the inundation risk sub-region. The minimum distance between the pixel of the inundation risk sub-region and the center of the circle is taken as the interval distance, that is, the minimum distance to the ground feature is taken as the interval distance. When the minimum interval distance of the inundation risk sub-region meets the requirements, that is, the distance to the ground feature is very long, the comparison processing scheme of the sub-region is determined to be based on the target number of pixels, and the inundation risk simulation processing is performed simultaneously.

[0099] It should also be noted that when the minimum value of the interval distance of the flood risk sub-regions does not meet the requirements, based on the interval distance of different flood risk sub-regions, if the average value of the interval distance of different flood risk sub-regions is less than the preset interval distance threshold, then when the flood risk simulation processing is carried out, the flood risk simulation processing is carried out by comparing pixels one by one.

[0100] Furthermore, when the average interval distance of different flood risk sub-regions is not less than a preset interval distance threshold, the flood risk sub-regions are sorted from largest to smallest according to their interval distances to obtain a sorting result. The flood risk sub-regions whose absolute value of the difference between their interval distance and the next flood risk sub-region in the sorting result is greater than the preset distance threshold are taken as interval regions. If the interval distance of the sub-region is greater than the interval distance of the interval region, then since the interval distance between the flood risk region and the ground feature is relatively long, the comparison processing scheme for the sub-region is determined to be that no flood risk simulation processing is required.

[0101] Additionally, it should be noted that if the interval distance of the sub-region is not greater than the interval distance of the interval region, then the reference deviation rate of different pixels in the sub-region is determined. When the average value of the reference deviation rate of different pixels in the sub-region is greater than the preset reference deviation rate threshold, then the simulation of flooding risk is performed by comparing pixels one by one.

[0102] Furthermore, when the average reference deviation rate of different pixels in the sub-region is not greater than the preset reference deviation rate threshold, it is determined that when the sub-region is subjected to flooding risk simulation processing, the flooding risk simulation processing is performed simultaneously based on the target number of pixels.

[0103] Example 3 Furthermore, the method for determining the amplification strategy for the recommended analysis region is as follows: Based on the comparison processing scheme of the sub-regions, the sub-regions to be compared and analyzed are determined and used as the analysis sub-regions; Based on the distribution data of ground features in the adjacent areas of the analysis sub-region, determine the number of ground features in the adjacent areas; The expansion strategy for the recommended analysis area is determined based on the number of ground features in the adjacent areas and the comparison processing scheme of the sub-regions.

[0104] Furthermore, when the number of the analysis sub-regions is less than the preset sub-region number threshold, since the number of analysis sub-regions is small, in order to fully determine the true situation of the flooding analysis, the recommended analysis region amplification strategy is determined to be that all adjacent regions of the analysis sub-regions need to be amplified.

[0105] Additionally, it should be noted that when the number of the analysis sub-regions is not less than a preset sub-region number threshold, the comparison processing reliability coefficient of the recommended analysis region is determined based on the comparison processing schemes of different analysis sub-regions. When the comparison processing reliability coefficient of the recommended analysis region is less than a preset reliability coefficient threshold, the amplification strategy of the recommended analysis region is determined to be that all adjacent regions of the analysis sub-region need to be amplified.

[0106] In one possible embodiment, the reliability coefficient of the comparison processing is determined based on the sum of preset weight coefficients corresponding to the comparison processing schemes of different analysis sub-regions. The preset weight coefficients are determined according to the comparison processing scheme. When the flooding risk is simulated by comparing pixels one by one, the preset weight coefficient is 0.1, and in other cases it is 0.05.

[0107] It is understood that when the reliability coefficient of the comparison processing of the recommended analysis area is not less than the preset reliability coefficient threshold, if the flooding risk coefficient of the analysis sub-region is less than the preset risk coefficient threshold, then it is determined that the adjacent areas of the analysis sub-region do not need to be amplified.

[0108] It should also be noted that if the flooding risk coefficient of the analyzed sub-region is greater than the preset risk coefficient threshold, and if the deviation rate between the reference elevation value of the adjacent region and the reference elevation value of the analyzed sub-region is within a preset range or if there are ground features in the adjacent region, then it is determined that the adjacent region needs to be amplified.

[0109] It should be noted that if the deviation rate between the reference elevation value of the adjacent area of ​​the analysis sub-region and the reference elevation value of the analysis sub-region is not within the preset range and there are no ground features in the adjacent area, then it is determined that the adjacent area does not need to be amplified.

[0110] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments of apparatus, devices, and non-volatile computer storage media are basically similar to the method embodiments, so the descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0111] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0112] The above description is merely one or more embodiments of this specification and is not intended to limit this specification. Various modifications and variations can be made to the one or more embodiments of this specification by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of one or more embodiments of this specification should be included within the scope of the claims of this specification.

Claims

1. A flooding analysis method based on image recognition, characterized in that, Specifically, it includes: After the system loads the raster data and vector layers of the digital elevation model, the user selects ground features through the interactive interface, and determines the recommended analysis area and the geometry of the clipping area of ​​the ground features based on the elevation data of the area where the ground features are located. Calculate the minimum bounding rectangle of the clipping region geometry and clip the elevation data to the bounding box range; Based on the user-input flooding threshold, and according to the analysis results of elevation values ​​in different sub-regions of the recommended analysis area, a comparative processing scheme for different sub-regions is determined. Based on the analysis results obtained from the comparative processing scheme, if it is determined that the flooding threshold belongs to the risk threshold, proceed to the next step. Based on the comparison processing scheme of the sub-region and the feature data of the adjacent regions, the expansion strategy of the recommended analysis region is determined. The expansion processing of the recommended analysis region is performed using the expansion strategy to obtain the expanded region. Based on the comparison analysis results of the expanded region, the color generation processing of pixel values ​​is performed to obtain the analysis results. Based on the analysis results, a raster file is generated, and the drawing results are output as a high-resolution image file with embedded geospatial coordinates.

2. The flooding analysis method based on image recognition as described in claim 1, characterized in that, The digital elevation model is a model that uses limited terrain elevation data to digitally simulate ground topography and features.

3. The flooding analysis method based on image recognition as described in claim 1, characterized in that, The aforementioned land features are objects represented by land feature symbols on topographic maps, specifically including residential areas, industrial and mining enterprise buildings, public facilities, independent land features, roads and their ancillary facilities, pipelines, water systems and their ancillary facilities.

4. The flooding analysis method based on image recognition as described in claim 1, characterized in that, The elevation data includes the elevation values ​​of different pixels in the digital elevation model.

5. The flooding analysis method based on image recognition as described in claim 1, characterized in that, The method for determining the recommended analysis area for the aforementioned land feature elements is as follows: Based on the elevation data of the area where the feature is located, determine the elevation values ​​of different pixels in the area where the feature is located; Based on the elevation values ​​of different pixels, determine the deviation between the elevation values ​​of different pixels; Based on the aforementioned deviations, a recommended analysis area for the aforementioned land feature is determined.

6. The flooding analysis method based on image recognition as described in claim 5, characterized in that, Based on the aforementioned deviations, a recommended analysis area for the land feature is determined, specifically including: Based on the aforementioned deviation, the average elevation value of different pixels in the area where the feature is located is determined and used as a reference elevation value. The absolute value of the difference between the elevation value of the pixel and the reference elevation value, and the ratio of the difference to the reference elevation value, are used as the deviation rate of the pixel. Based on the different pixel deviation rates, the recommended analysis area for the ground features is determined.

7. The flooding analysis method based on image recognition as described in claim 6, characterized in that, When the average deviation rate of different pixels does not meet the requirements, the recommended analysis area for the feature is determined to be the area where the feature is located.

8. The flooding analysis method based on image recognition as described in claim 1, characterized in that, The method for determining the amplification strategy for the recommended analysis region is as follows: Based on the comparison processing scheme of the sub-regions, the sub-regions to be compared and analyzed are determined and used as the analysis sub-regions; Based on the distribution data of ground features in the adjacent areas of the analysis sub-region, determine the number of ground features in the adjacent areas; The expansion strategy for the recommended analysis area is determined based on the number of ground features in the adjacent areas and the comparison processing scheme of the sub-regions.

9. The flooding analysis method based on image recognition as described in claim 8, characterized in that, When the number of the analysis sub-regions is less than the preset sub-region number threshold, the recommended analysis region amplification strategy is determined to be that all adjacent regions of the analysis sub-region need to be amplified.

10. A computer system, comprising: A memory and processor connected in communication, and a computer program stored in the memory and capable of running on the processor, characterized in that, when the processor runs the computer program, it executes a flooding analysis method based on image recognition as described in any one of claims 1-9.

Images