Modeling method and device based on oblique photography and unmanned aerial vehicle measurement and mapping

Abstract

The invention provides a modeling method and device based on oblique photography and unmanned aerial vehicle measurement surveying and mapping. The method comprises the steps that unmanned aerial vehicle flight historical data and measurement area related data are collected to form a basic database; carrying out region division according to landform and ground feature distribution characteristics of a measurement region, and identifying modeling precision influence factors; predicting modeling precision based on the influence factors to obtain a precision reference target; key influence factors of the same type of measurement area schemes are set, historical measurement information is identified, and a precision historical reference value is obtained through data analysis; and determining a modeling precision prediction target according to the precision reference target and the historical reference value. According to the method, the area with the large elasticity of the measurement parameters can be adjusted preferentially, the deviation is controlled within the preset error range, and the modeling precision of different types of to-be-measured measurement areas is predicted. The efficiency and accuracy of oblique photography and unmanned aerial vehicle surveying and mapping can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle surveying and mapping, and more specifically, to a modeling method and device based on oblique photography and unmanned aerial vehicle surveying and mapping. Background Art

[0002] Existing oblique photography and drone surveying and mapping technologies have been widely used in the fields of geographic information systems, urban planning, and environmental monitoring. These technologies can provide high-resolution surface image data. Through the photography equipment carried by drones, large-area geographic information can be quickly obtained, which facilitates the three-dimensional modeling of topography. However, due to the complexity of topography and distribution of objects, the modeling accuracy of different regions is affected by many factors, such as terrain undulation, object density, weather conditions, etc. These factors have different degrees of influence on the modeling accuracy, making it difficult for traditional modeling methods to accurately predict and control the final modeling results.

[0003] In the process of implementing the embodiments of the present invention, the inventors found that there are at least the following problems or defects in the prior art: the existing modeling methods often lack a comprehensive analysis of the topography and distribution characteristics of the survey area, and fail to fully identify and quantify the key factors affecting the modeling accuracy, resulting in inaccurate predictions of modeling accuracy and inability to optimize and adjust according to the characteristics of different regions, thereby affecting the efficiency and quality of modeling. In addition, when processing historical measurement data, the prior art also lacks an effective data analysis method to extract historical reference values ​​of accuracy to guide actual modeling work. These problems limit the application potential of oblique photography and drone surveying and mapping technology in complex survey areas, and a new modeling method is urgently needed to improve the accuracy and reliability of modeling. Summary of the invention

[0004] The present invention provides a modeling method and device based on oblique photography and unmanned aerial vehicle surveying and mapping.

[0005] In a first aspect of the present invention, a modeling method based on oblique photography and drone surveying and mapping is provided, comprising:

[0006] Collect UAV flight history data and survey area related information to form a basic database, wherein the historical data includes flight logs, image data and corresponding survey area geographic information;

[0007] Divide the survey area into regions according to the topography and distribution characteristics of the survey area, and identify the factors affecting the modeling accuracy for different regions of the survey area;

[0008] Based on the influencing factors of the modeling accuracy, the modeling accuracy of oblique photography and UAV surveying and mapping is predicted to obtain an accuracy reference target;

[0009] The key influencing factors of the same type of survey area scheme are set based on the influencing factors of the modeling accuracy, the same type of historical measurement information is identified based on the key influencing factors, and then the accuracy historical reference values ​​of the same type of historical measurements are obtained through data analysis;

[0010] And determine the modeling accuracy prediction target based on the accuracy reference target and the accuracy historical reference value.

[0011] Furthermore, after determining the modeling accuracy prediction target, it further includes:

[0012] Combined with regional decomposition, the design accuracy of different regions is compared with the modeling accuracy prediction target, the deviation and impact ratio are determined, and the regions with greater measurement parameter elasticity are adjusted first to control the deviation within a preset error range;

[0013] And the modeling accuracy of different types of areas to be measured is predicted by using the modeling method based on oblique photography and UAV surveying and mapping.

[0014] Furthermore, the survey area is divided into regions according to the topography and distribution characteristics of the survey area, and the factors affecting the modeling accuracy are identified for different regions of the survey area, including:

[0015] Determine the topography and distribution characteristics of the survey area based on the image data and the geographic information of the survey area;

[0016] Decomposing the geographical part of the survey area into mountainous areas, plains, cities, water areas and special terrain areas according to the topography and distribution characteristics of the landforms in the survey area;

[0017] And identify the factors affecting the modeling accuracy in the mountainous areas, plains, cities, waters and special terrain areas of the survey area, the factors affecting the modeling accuracy include terrain undulation, density of land objects, weather conditions, light intensity, wind speed, air quality, surface material, height and area of ​​land objects.

[0018] Furthermore, based on the factors affecting modeling accuracy, the modeling accuracy of oblique photography and UAV surveying and mapping is predicted, and the accuracy reference targets include:

[0019] The modeling accuracy of the mountainous area, plain, city, water area and special terrain area in the survey area is predicted separately to obtain the modeling accuracy of different areas;

[0020] And the modeling accuracies of the different areas are superimposed to obtain the total modeling accuracy of the survey area as the accuracy reference target.

[0021] Further, identifying the same type of historical measurement information based on the key influencing factors includes:

[0022] Based on the key influencing factors, all sample data are searched, and the improved K-means algorithm is applied to find the number of clusters with the smallest sum of squared error (SSE) value and the corresponding initial cluster centers;

[0023] After the initial clustering center is determined, the category of the sample data and the center of each category are finally determined according to the improved density peak clustering algorithm, wherein the key influencing factors of the mountainous area include terrain undulation, height of the object and surface material;

[0024] The key factors affecting the plain include the density of landforms and the surface material;

[0025] The key influencing factors of the city include land feature density, land feature height and surface material;

[0026] Key factors affecting the water area include surface texture and wind speed;

[0027] And the key influencing factors of the special terrain area include terrain undulation, surface material and special terrain attributes.

[0028] Furthermore, the categories of sample data and the centers of each category are finally determined according to the improved density peak clustering algorithm, further including:

[0029] The objective function of the improved density peak clustering algorithm is:

[0030] max{G(I)}

[0031] stG=λ i -ω i

[0032] Where m is the local neighborhood parameter, λ i Represents the local data association, ω i Represents the shortest path distance to a point with a higher degree of association, I is the set of sample points, i = 1, 2, ..., n, n is the number of data;

[0033] The local data correlation λ of each data point is calculated in the iterative process i and the shortest path distance to a point with higher association ω i , and obtain G = λ i -ω i , and then finally get the optimal number of clusters, cluster prototypes and the type of sample data with the maximum value of G;

[0034] Given the local neighborhood parameter m, set the iteration stop threshold δ, initialize the sample point attributes, and set the iteration counter t = 0. If there exists i such that λ i >0, then

[0035]

[0036] If there exists i such that λ i =0, then ω i =max{ω j}, j = 1, 2, ..., n:

[0037] like Then the algorithm stops and outputs the partitioning result and cluster prototype, otherwise let t = t + 1 and go to the previous step;

[0038] Determine the best classification tree C m , calculate the value of G(C), when the condition is met: (G(C m )=max{G(C)}, then we think that C m is the optimal number of classifications, and the output division result and the clustering prototype are obtained.

[0039] Furthermore, the historical reference values ​​of accuracy of the same type of historical measurements obtained through data analysis include:

[0040] Calculate the median M of the sample data d :Sort the sample data from small to large. If the number of data n is an odd number, the median

[0041] If n is even, then

[0042]

[0043] where x (k) Indicates the kth data after sorting;

[0044] Calculate the mode M o , the mode is the data value that appears most frequently in the sample data;

[0045] Calculate the range R:

[0046] R = x max -x min

[0047] where x max and x min are the maximum and minimum values ​​in the sample data respectively;

[0048] Calculate the mean absolute deviation x:

[0049]

[0050] in is the average value of the sample data; the Grubbs criterion is used to eliminate outliers. Suppose a set of measurement data is x1, x2, …, x n, whose arithmetic mean is The standard deviation is s, find the suspicious value x (k) (k=1 or n), calculate If T>G(α,n)(G(α,n)) is the Grubbs coefficient, which is obtained by looking up the Grubbs table based on the given significance level α and the number of data n, then x (k) outliers and remove them, and perform outlier analysis on the sample data;

[0051] The reasonable interval L of accuracy is calculated by the following formula:

[0052]

[0053] Where k1 is a coefficient determined according to the distribution characteristics of the measured data; the sample data after outlier processing is x′1, x′2,…, x′ n′ , then the average Finally, the accuracy historical reference value CKJD of the same type of historical measurement is calculated by the following formula;

[0054] Among them, C represents the category of historical measurement data, K represents the key factors affecting the measurement accuracy of this category, J represents the judgment and processing process of the data, and D represents the valid data after processing.

[0055] Further, determining the modeling accuracy prediction target according to the accuracy reference target and the accuracy historical reference value includes:

[0056] Assume that the accuracy reference target is CKJC, and the weighting coefficient β ranges from [0,1];

[0057] The modeling accuracy prediction target MJ is determined by the following formula:

[0058] MJ=βCKJD+(1-β)CKJC

[0059] Furthermore, based on the influencing factors of the modeling accuracy, the modeling accuracy of oblique photography and UAV surveying and mapping is predicted, and the accuracy reference targets include:

[0060] The method comprises predicting the modeling accuracy of oblique photography and UAV surveying and mapping based on the influencing factors of the modeling accuracy to obtain the accuracy reference target, including predicting the modeling accuracy of oblique photography and UAV surveying and mapping based on the influencing factors of the modeling accuracy using a neural network model optimized based on a particle swarm algorithm to obtain the accuracy reference target, wherein the neural network model maps the input vector to a high-dimensional feature space through the activation function of the neuron, constructs an optimal decision boundary in the high-dimensional feature space, and the output of the neural network is a weighted combination of the outputs of each neuron, wherein each neuron corresponds to an input feature.

[0061] In a second aspect of the present invention, a modeling device based on oblique photography and drone surveying and mapping is provided, comprising:

[0062] A data acquisition module is used to collect the UAV flight history data and survey area related information to form a basic database, wherein the historical data includes flight logs, image data and corresponding survey area geographic information;

[0063] An influencing factor identification module is used to divide the survey area into regions according to the topography and distribution characteristics of the survey area, and identify the influencing factors of modeling accuracy for different regions of the survey area;

[0064] A target prediction module predicts the modeling accuracy of oblique photography and UAV surveying and mapping based on the influencing factors of the modeling accuracy to obtain an accuracy reference target;

[0065] A historical accuracy acquisition module, used to set key influencing factors of the same type of measurement area based on the influencing factors of the modeling accuracy, identify the same type of historical measurement information based on the key influencing factors, and then obtain the accuracy historical reference value of the same type of historical measurement through data analysis;

[0066] And a prediction target acquisition module is used to determine the modeling accuracy prediction target based on the accuracy reference target and the accuracy historical reference value.

[0067] According to the above-mentioned embodiments of the present invention, at least the following beneficial effects are achieved: the modeling method based on oblique photography and UAV surveying and mapping to which the present invention belongs can effectively improve the prediction accuracy of modeling accuracy. By collecting historical UAV flight data and relevant information of the survey area, a basic database is formed, and regional division is performed according to the topography and distribution characteristics of the survey area to identify the key factors affecting the modeling accuracy. This method can predict the modeling accuracy of different areas and obtain the accuracy reference target, and then obtain the accuracy historical reference value of the same type of historical measurement through data analysis. Combined with these reference values, the modeling accuracy prediction target can be determined, so that the modeling accuracy can be optimized and adjusted before the actual measurement to ensure the reliability and accuracy of the measurement results.

[0068] In addition, this method can also make more refined predictions on modeling accuracy through improved clustering algorithms and neural network models. This method can identify historical measurement information of the same type, and calculate statistical parameters such as the median, mode, range, and mean absolute deviation of sample data. After removing outliers, the reasonable range of accuracy can be calculated. Through these refined processing, the modeling accuracy of different types of measurement areas can be more accurately predicted, providing more precise guidance for actual measurement work, thereby improving overall work efficiency and quality of results. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] The above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood by reading the detailed description below with reference to the accompanying drawings. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, in which:

[0070] Figure 1 A schematic flow chart of a modeling method based on oblique photography and UAV surveying and mapping provided by an embodiment of the present invention;

[0071] Figure 2 A schematic diagram of the structure of a modeling device based on oblique photography and drone surveying and mapping provided by an embodiment of the present invention;

[0072] Figure 3 The schematic diagram schematically shows the structure of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0073] The principles and spirit of the present invention will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are provided only to enable those skilled in the art to better understand and implement the present invention, and are not intended to limit the scope of the present invention in any way. On the contrary, these embodiments are provided to make the present invention more thorough and complete, and to fully convey the scope of the present invention to those skilled in the art.

[0074] Those skilled in the art know that the embodiments of the present invention can be implemented as a system, device, apparatus, method or computer program product. Therefore, the present invention can be specifically implemented in the following forms, namely: complete hardware, complete software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0075] It should be noted that any number of elements in the drawings is for illustration rather than limitation, and any naming is only for distinction and does not have any limiting meaning.

[0076] Reference below Figure 1 , Figure 1The following is a flow chart of a modeling method based on oblique photography and UAV surveying and mapping provided by an embodiment of the present invention. Figure 1 As shown, a modeling method 100 based on oblique photography and drone surveying and mapping includes:

[0077] Step 101, collecting UAV flight history data and survey area related information to form a basic database, wherein the history data includes flight logs, image data and corresponding survey area geographic information;

[0078] Step 102, dividing the survey area into regions according to the topography and distribution characteristics of the survey area, and identifying the factors affecting the modeling accuracy for different regions of the survey area;

[0079] Step 103, predicting the modeling accuracy of oblique photography and UAV surveying and mapping based on the influencing factors of the modeling accuracy to obtain an accuracy reference target;

[0080] Step 104, setting key influencing factors of the same type of survey area scheme based on the influencing factors of the modeling accuracy, identifying the same type of historical measurement information based on the key influencing factors, and then obtaining the accuracy historical reference value of the same type of historical measurement through data analysis;

[0081] Step 105, and determining a modeling accuracy prediction target based on the accuracy reference target and the accuracy historical reference value.

[0082] It should be noted that this method first involves collecting UAV flight history data and survey area related information to form a basic database. The basic database here refers to a collection of UAV flight logs, image data, and corresponding survey area geographic information, which provides necessary information support for subsequent modeling accuracy prediction.

[0083] Specifically, the construction of the basic database includes but is not limited to the flight log information of the drone, such as flight altitude, speed, route, etc., as well as the captured high-resolution image data and geographic information system (GIS) data of the survey area. These data can be collected in real time during the flight by the sensors and photography equipment carried by the drone, and integrated into the database through post-processing.

[0084] More specifically, for flight log information, specific parameter thresholds can be set, for example, the flight altitude and speed of the drone in different terrains to ensure the quality and consistency of the data. For image data, cameras with different resolutions can be used to meet the needs of survey areas of different sizes. GIS data includes information such as the terrain undulations and distribution of objects in the survey area, which can be managed and analyzed through existing geographic information system software.

[0085] Furthermore, according to actual needs, other sensors carried by drones, such as multispectral sensors or LiDAR, can be used to obtain richer geographic information.

[0086] In some embodiments, after determining the modeling accuracy prediction target, the method further includes:

[0087] Combined with regional decomposition, the design accuracy of different regions is compared with the modeling accuracy prediction target, the deviation and impact ratio are determined, and the regions with greater measurement parameter elasticity are adjusted first to control the deviation within a preset error range;

[0088] And the modeling accuracy of different types of areas to be measured is predicted by using the modeling method based on oblique photography and UAV surveying and mapping.

[0089] It should be noted that after determining the prediction target of modeling accuracy, this embodiment further includes combining regional decomposition to compare the design accuracy of different regions with the prediction target to determine their deviation and impact ratio. Regional decomposition here refers to dividing the survey area into several regions according to the characteristics of topography and landform distribution, so as to evaluate and adjust the modeling accuracy of each region separately.

[0090] Specifically, regional decomposition can be performed based on the specific characteristics of the survey area, such as mountainous areas, plains, cities, water areas, and special terrain areas. The design accuracy of each area refers to the measurement accuracy expected to be achieved in the area, which can be set according to the complexity and importance of the area. The prediction target refers to the modeling accuracy target predicted based on the influencing factors. By comparison, it is possible to identify which areas have a design accuracy lower than the prediction target, so that the measurement parameters need to be adjusted to improve the accuracy.

[0091] Preferably, for areas that need to be adjusted first, those areas with greater flexibility in measurement parameters can be selected. For example, in urban areas, due to the influence of high-rise buildings, it may be necessary to adjust the flight altitude and shooting angle of the drone to obtain better image data. In mountainous areas, it may be necessary to increase the flight path to cover more terrain undulations.

[0092] Furthermore, the data quality of these areas can be improved by adjusting the sensor settings of the drone or adding additional sensors. In terms of controlling the deviation, a pre-set error range, such as ±5 cm or ±10 cm, can be set to ensure that the final modeling results meet the project requirements. Through these adjustments, the deviation can be effectively controlled within an acceptable range to ensure that the modeling accuracy meets the project requirements.

[0093] In some embodiments, the survey area is divided into regions according to the topography and distribution characteristics of land objects in the survey area, and factors affecting the modeling accuracy are identified for different regions of the survey area, including:

[0094] Determine the topography and distribution characteristics of the survey area based on the image data and the geographic information of the survey area;

[0095] Decomposing the geographical part of the survey area into mountainous areas, plains, cities, water areas and special terrain areas according to the topography and distribution characteristics of the landforms in the survey area;

[0096] And identify the factors affecting the modeling accuracy in the mountainous areas, plains, cities, waters and special terrain areas of the survey area, the factors affecting the modeling accuracy include terrain undulation, density of land objects, weather conditions, light intensity, wind speed, air quality, surface material, height and area of ​​land objects.

[0097] It should be noted that, according to the topography and distribution characteristics of the survey area, this implementation method involves dividing the survey area into regions and identifying the factors affecting the modeling accuracy. The regional division here refers to dividing the survey area into different geographical areas, such as mountainous areas, plains, cities, water areas and special terrain areas, according to the specific geographical characteristics of the survey area, so as to more accurately manage and predict the modeling accuracy.

[0098] Specifically, regional division can determine the topography and distribution characteristics of the survey area by analyzing image data and geographic information of the survey area. For example, mountainous areas may have complex terrain undulations, urban areas may have high-density buildings, and water areas may need to consider factors such as water surface reflection.

[0099] More specifically, factors affecting modeling accuracy include terrain undulation, object density, weather conditions, light intensity, wind speed, air quality, surface material, object height and area, etc. These factors can be weighted and prioritized according to their potential impact on measurement accuracy.

[0100] Preferably, for each area, the identified factors affecting modeling accuracy can be further refined. For example, in mountainous areas, terrain undulations and surface materials may be the main influencing factors, so high-resolution terrain data and surface cover type data can be used to improve modeling accuracy. In urban areas, object density and object height may be key factors, so the flight frequency and shooting angle of the drone can be increased to ensure accurate capture of buildings. For water areas, wind speed and surface material may be more important, so suitable weather conditions can be selected for data collection, and sensor technology that can penetrate the water surface can be used.

[0101] Furthermore, the data collection strategy and processing algorithm can be adjusted according to the actual measurement needs and available resources to adapt to the specific requirements of different regions.

[0102] In some embodiments, the accuracy of modeling by oblique photography and UAV surveying and mapping is predicted based on the factors affecting modeling accuracy, and the accuracy reference target obtained includes:

[0103] The modeling accuracy of the mountainous area, plain, city, water area and special terrain area in the survey area is predicted separately to obtain the modeling accuracy of different areas;

[0104] And the modeling accuracies of the different areas are superimposed to obtain the total modeling accuracy of the survey area as the accuracy reference target.

[0105] It should be noted that the accuracy of oblique photography and UAV surveying and mapping modeling is predicted based on the factors affecting modeling accuracy to obtain the accuracy reference target. The "accuracy reference target" here refers to an expected accuracy level set for each area after analyzing the influencing factors. This level will serve as the quality standard for subsequent measurement and modeling work.

[0106] Specifically, the process of predicting modeling accuracy involves predicting the modeling accuracy of different regions separately. For example, the modeling accuracy of mountainous areas, plains, cities, water areas and special terrain areas may vary due to differences in topography and distribution of objects.

[0107] More specifically, when predicting, statistical analysis methods can be used to combine historical data and field measurement data to estimate the modeling accuracy of each area. The overall modeling accuracy is obtained by superimposing the modeling accuracy of different areas and comprehensively considering the characteristics and importance of each area to obtain the accuracy reference target of the entire survey area.

[0108] Preferably, multiple methods can be used to perform cross-validation for modeling accuracy predictions in different regions to improve the accuracy of the predictions. For example, a machine learning algorithm, such as a random forest or a support vector machine (SVM), can be used to train historical data to predict the modeling accuracy of new data.

[0109] Furthermore, the prediction model can be verified and calibrated through field tests, such as deploying ground control points at different locations in the survey area. When superimposing the modeling accuracy of different regions, a weighted average method can be used, where the weight can be determined based on the importance of the region and the difficulty of measurement. This method can ensure that the final accuracy reference target is both scientific and practical, and can provide effective guidance for actual measurement and modeling work.

[0110] In some embodiments, identifying the same type of historical measurement information based on the key influencing factors includes:

[0111] Based on the key influencing factors, all sample data are searched, and the improved K-means algorithm is applied to find the number of clusters with the smallest sum of squared error (SSE) value and the corresponding initial cluster centers;

[0112] After the initial clustering center is determined, the category of the sample data and the center of each category are finally determined according to the improved density peak clustering algorithm, wherein the key influencing factors of the mountainous area include terrain undulation, height of the object and surface material;

[0113] The key factors affecting the plain include the density of landforms and the surface material;

[0114] The key influencing factors of the city include land feature density, land feature height and surface material;

[0115] Key factors affecting the water area include surface texture and wind speed;

[0116] And the key influencing factors of the special terrain area include terrain undulation, surface material and special terrain attributes.

[0117] It should be noted that identifying the same type of historical measurement information based on key influencing factors involves using the improved K-means algorithm and density peak clustering algorithm to classify the sample data. The key influencing factors here refer to those factors that have a significant impact on the modeling accuracy, such as terrain undulation, land object density, etc., while the sample data refers to the collection of historical measurement data, which will be used to predict the modeling accuracy of the current survey area.

[0118] Specifically, the process of identifying the same type of historical measurement information includes searching all sample data, applying the improved K-means algorithm to determine the optimal number of clusters and initial cluster centers. The improved density peak clustering algorithm will be used to finally determine the categories of sample data and the centers of each category.

[0119] More specifically, in this process, the key influencing factors of mountainous areas, plains, cities, water areas and special terrain areas will be considered separately. For example, mountainous areas may consider terrain undulation, object height and surface material, while urban areas may consider object density, object height and surface material.

[0120] Preferably, for the improved K-means algorithm, the number of clusters with the smallest sum of squared errors (SSE) can be set as the optimization target to determine the optimal number of clusters and initial cluster centers. For the density peak clustering algorithm, the local neighborhood parameters and the iteration stop threshold can be adjusted to optimize the clustering results.

[0121] Furthermore, a variety of data preprocessing techniques, such as data standardization or normalization, can be used to improve the effect of clustering algorithms. When determining cluster centers, a variety of distance metrics, such as Euclidean distance or Manhattan distance, can be considered to accommodate different types of data distribution characteristics. Refinements and alternatives to these steps can improve the accuracy of historical measurement information identification, thereby providing a more reliable reference for modeling accuracy prediction.

[0122] In some embodiments, finally determining the category of sample data and the center of each category according to the improved density peak clustering algorithm further includes:

[0123] The objective function of the improved density peak clustering algorithm is:

[0124] max{G(I)}

[0125] stG=λ i -ω i

[0126] Where m is the local neighborhood parameter, λ i Represents the local data association, ω i Represents the shortest path distance to a point with a higher degree of association, I is the set of sample points, i = 1, 2, ..., n, n is the number of data;

[0127] The local data correlation λ of each data point is calculated in the iterative process i and the shortest path distance to a point with higher association ω i , and obtain G = λ i -ω i , and then finally get the optimal number of clusters, cluster prototypes and the type of sample data with the maximum value of G;

[0128] Given the local neighborhood parameter m, set the iteration stop threshold δ, initialize the sample point attributes, and set the iteration counter t = 0. If there exists i such that λ i >0, then

[0129]

[0130] If there exists i such that λ i =0, then ω i =max{ω j}, j = 1, 2, ..., n:

[0131] like Then the algorithm stops and outputs the partitioning result and cluster prototype, otherwise let t = t + 1 and go to the previous step;

[0132] Determine the best classification tree C m, calculate the value of G(C), when the condition is met: (G(C m )=max{G(C)}, then we think that C m is the optimal number of classifications, and the output division result and the clustering prototype are obtained.

[0133] It should be noted that the process of finally determining the categories of sample data and the centers of each category according to the improved density peak clustering algorithm involves the objective function and iterative calculation of the algorithm. The improved density peak clustering algorithm here is a clustering technology that determines the clustering of data by calculating the local data association of data points and the shortest path distance to points with higher association.

[0134] Specifically, the objective function of the improved density peak clustering algorithm is defined as: the local neighborhood parameter λ represents the local data association degree ρ represents the shortest path distance d to a point with a higher association degree, and the Δ value of each data point i in the sample point set X is calculated as Δi=ρi-di. In the iterative process, ρ and d are calculated for each data point, and the Δ value is obtained. Then, the optimal number of clusters, cluster prototypes, and the type of sample data are finally obtained with the maximum Δ value. The local neighborhood parameter λ and the iteration stop threshold ε are key parameters in the algorithm and need to be set according to the characteristics of the data and the clustering requirements.

[0135] Preferably, the selection of the local neighborhood parameter λ can be dynamically adjusted according to the distribution density of the data points to ensure that each data point has enough neighbors for comparison. The iteration stop threshold ε can be set to a small positive number to ensure that the algorithm can converge within a reasonable number of iterations.

[0136] Furthermore, for the attribute initialization of data points, the output of the K-means algorithm can be used as the initial cluster center to speed up the convergence of the clustering algorithm. In the iterative process, a variety of distance measurement methods can be used to calculate the distance between data points, such as Euclidean distance or Manhattan distance, to adapt to different data characteristics. These refined operation steps and alternatives can improve the accuracy and efficiency of the clustering algorithm, thereby more accurately determining the categories and category centers of sample data.

[0137] In some embodiments, obtaining the accuracy historical reference value of the same type of historical measurement through data analysis includes:

[0138] Calculate the median M of the sample data d :Sort the sample data from small to large. If the number of data n is an odd number, the median

[0139] If n is even, then

[0140]

[0141] where x (k) Indicates the kth data after sorting;

[0142] Calculate the mode M o , the mode is the data value that appears most frequently in the sample data;

[0143] Calculate the range R:

[0144] R = x max -x min

[0145] where x max and x min are the maximum and minimum values ​​in the sample data respectively;

[0146] Calculate the mean absolute deviation MAD:

[0147]

[0148] in is the average value of the sample data; the Grubbs criterion is used to eliminate outliers. Suppose a set of measurement data is x1, x2, …, x n , whose arithmetic mean is The standard deviation is s, find the suspicious value x (k) (k=1 or n), calculate If T>G(α,n)(G(α,n)) is the Grubbs coefficient, which is obtained by looking up the Grubbs table based on the given significance level α and the number of data n, then x (k) outliers and remove them, and perform outlier analysis on the sample data;

[0149] The reasonable interval L of accuracy is calculated by the following formula:

[0150]

[0151] Where k1 is a coefficient determined according to the distribution characteristics of the measured data; the sample data after outlier processing is x′1, x′2,…, x′ n′ , then the average Finally, the accuracy historical reference value CKJD of the same type of historical measurement is calculated by the following formula;

[0152] Among them, C represents the category of historical measurement data, K represents the key factors affecting the measurement accuracy of this category, J represents the judgment and processing process of the data, and D represents the valid data after processing.

[0153] It should be noted that the historical reference values ​​of accuracy of the same type of historical measurements are obtained through data analysis, which involves calculating statistical parameters such as the median, mode, range, and mean absolute deviation of sample data. The historical reference values ​​of accuracy here refer to a series of statistics calculated based on historical data to evaluate and predict the modeling accuracy of the current measurement area.

[0154] Specifically, when calculating the median of sample data, the data needs to be sorted from small to large. If the number of data n is an odd number, the median is the value in the middle; if n is an even number, the median is the average of the two middle values. The mode refers to the value that appears most frequently in the sample data. The range is the difference between the maximum and minimum values ​​in the sample data, reflecting the degree of dispersion of the data. The mean absolute deviation is the average of the absolute values ​​of the differences between each value of the sample data and its mean, which is used to measure the central tendency of the data. The Grubbs criterion is a method for identifying and eliminating outliers, and outliers are determined by calculating the Grubbs coefficient.

[0155] Preferably, for the calculation of the median, a fast selection algorithm can be used to efficiently find the median in the sorted data set. When determining the mode, a threshold can be set to determine whether a value is the mode. For example, if the frequency of occurrence of a value exceeds a certain proportion of the total data volume, it can be considered to be the mode.

[0156] Furthermore, when calculating the range, the distribution of the data can be considered to determine whether the data needs to be standardized. For the calculation of the mean absolute deviation, the weighted mean absolute deviation can be used, where the weight can be assigned according to the importance or reliability of the data. When applying the Grubbs criterion, the corresponding Grubbs coefficient can be obtained by looking up the table according to the significance level and number of the data to more accurately identify outliers. These refined operating steps and alternatives can improve the accuracy and reliability of the accuracy analysis of historical measurement data.

[0157] In some embodiments, determining the modeling accuracy prediction target according to the accuracy reference target and the accuracy historical reference value includes:

[0158] Assume that the accuracy reference target is CKJC, and the weighting coefficient β ranges from [0,1];

[0159] The modeling accuracy prediction target MJ is determined by the following formula:

[0160] MJ=βCKJD+(1-β)CKJC

[0161] It should be noted that the process of determining the modeling accuracy prediction target based on the accuracy reference target and the accuracy historical reference value involves weighted averaging of the two reference values. The modeling accuracy prediction target here refers to an expected accuracy value comprehensively derived based on the prediction accuracy of the current survey area and the accuracy of historical data, which is used to guide the actual modeling work.

[0162] Specifically, in this process, both the accuracy reference target and the accuracy historical reference value are considered as the calculation objects of the weighted coefficient. The value range of the weighted coefficient α is [0,1], which represents the weight of the historical reference value in the prediction target, while 1-α represents the weight of the accuracy reference target. In this way, the impact of historical data and prediction data on the final prediction target can be balanced, making the prediction target more in line with the actual situation.

[0163] Preferably, the specific value of the weighting coefficient α can be dynamically adjusted according to the reliability of the historical data and the accuracy of the prediction model. For example, if the historical data is very rich and accurate, the value of α can be appropriately increased so that the historical reference value occupies a larger proportion in the prediction target. Conversely, if the prediction model has been verified to have a high accuracy, the value of α can be appropriately reduced so that the accuracy reference target has a greater impact.

[0164] Furthermore, other factors, such as the complexity of the survey area, weather conditions, etc., can be introduced as the basis for adjusting the weighting coefficient. These detailed operation steps and alternatives can make the modeling accuracy prediction target more accurate and adaptable.

[0165] In some embodiments, the modeling accuracy of oblique photography and UAV surveying and mapping is predicted based on the influencing factors of the modeling accuracy, and the accuracy reference target is obtained, including:

[0166] The method comprises predicting the modeling accuracy of oblique photography and UAV surveying and mapping based on the influencing factors of the modeling accuracy to obtain the accuracy reference target, including predicting the modeling accuracy of oblique photography and UAV surveying and mapping based on the influencing factors of the modeling accuracy using a neural network model optimized based on a particle swarm algorithm to obtain the accuracy reference target, wherein the neural network model maps the input vector to a high-dimensional feature space through the activation function of the neuron, constructs an optimal decision boundary in the high-dimensional feature space, and the output of the neural network is a weighted combination of the outputs of each neuron, wherein each neuron corresponds to an input feature.

[0167] It should be noted that based on the factors affecting modeling accuracy, a neural network model optimized by particle swarm algorithm is used to predict the modeling accuracy of oblique photography and UAV surveying and mapping to obtain the accuracy reference target. The neural network model optimized by particle swarm algorithm here is a neural network combined with particle swarm optimization (PSO) technology to improve the accuracy of model prediction.

[0168] Specifically, the particle swarm algorithm is an optimization algorithm that simulates the behavior of biological groups such as bird flocks or fish schools, and is used to find the optimal solution for the weights and biases of the neural network. The neural network model maps the input vector to a high-dimensional feature space through the activation function of the neuron and constructs the optimal decision boundary in this space. The output of the neural network is a weighted combination of the outputs of each neuron. Each neuron corresponds to an input feature, which includes factors affecting the modeling accuracy such as terrain undulation and land object density.

[0169] Preferably, for the structure of the neural network model, different numbers of layers and neurons in each layer can be selected to adapt to modeling tasks of different complexity. For example, for simple terrain, fewer layers and neurons may be required, while for complex urban areas, deeper networks and more neurons may be required. The parameter settings of the particle swarm algorithm, such as the number of particles, the maximum number of iterations, and the learning factor, also need to be adjusted according to the specific problem.

[0170] Furthermore, techniques such as cross-validation can be used to evaluate the generalization ability of the model and adjust the model parameters accordingly. These refined operation steps and alternatives can improve the accuracy and robustness of the model's predictive modeling accuracy.

[0171] The above-mentioned embodiments of the present invention have the following beneficial effects: the modeling method described in the present invention can form a basic database by collecting drone flight history data and survey area related information, and divide the area according to the topography and distribution characteristics of the survey area, and identify the key factors affecting the modeling accuracy. This method can predict the factors affecting the modeling accuracy, obtain the accuracy reference target, and provide a scientific basis for subsequent measurement work. At the same time, by separately predicting and superimposing the modeling accuracy of different areas, the total modeling accuracy of the survey area can be obtained, providing an effective means for accurately controlling the modeling quality.

[0172] In addition, the implementation steps of the modeling method include adjusting the measurement parameters in combination with regional decomposition, using improved clustering algorithms and neural network models to predict modeling accuracy, etc. These steps can improve the accuracy of modeling accuracy prediction, optimize the measurement scheme, and ensure that the design accuracy of different regions matches the prediction target. In addition, the historical measurement accuracy reference value obtained through data analysis can provide an important reference for determining the modeling accuracy prediction target, making the prediction results closer to the actual measurement situation, thereby improving the efficiency and quality of modeling work.

[0173] like Figure 2 As shown, in some embodiments, a modeling device 200 based on oblique photography and drone surveying and mapping comprises:

[0174] The data acquisition module 201 is used to collect the UAV flight history data and survey area related information to form a basic database, wherein the history data includes flight logs, image data and corresponding survey area geographic information;

[0175] An influencing factor identification module 202 is used to divide the survey area into regions according to the topography and distribution characteristics of the survey area, and identify the influencing factors of the modeling accuracy for different regions of the survey area;

[0176] A target prediction module 203 predicts the modeling accuracy of oblique photography and UAV surveying and mapping based on the influencing factors of the modeling accuracy to obtain an accuracy reference target;

[0177] A historical accuracy acquisition module 204 is used to set key influencing factors of the same type of measurement area based on the influencing factors of the modeling accuracy, identify the same type of historical measurement information based on the key influencing factors, and then obtain the accuracy historical reference value of the same type of historical measurement through data analysis;

[0178] And a prediction target acquisition module 205 is used to determine a modeling accuracy prediction target based on the accuracy reference target and the accuracy historical reference value.

[0179] It can be understood that the modules recorded in the modeling device 200 based on oblique photography and drone surveying and mapping are similar to those in the reference Figure 1 The steps in the modeling method based on oblique photography and drone surveying and mapping described above correspond to each other. Therefore, the operations, features and beneficial effects described above for the modeling method based on oblique photography and drone surveying and mapping are also applicable to the modeling device 200 based on oblique photography and drone surveying and mapping and the modules contained therein, and will not be repeated here.

[0180] Reference below Figure 3 , which shows a schematic diagram of a structure 300 of an electronic device suitable for implementing some embodiments of the present invention. The electronic devices in some embodiments of the present invention may include, but are not limited to, mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 3The terminal device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0181] like Figure 3 As shown, the electronic device 300 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a storage device 308 into a random access memory (RAM) 303. In the RAM 303, various programs and data required for the operation of the electronic device 300 are also stored. The processing device 301, the ROM 302, and the RAM 303 are connected to each other via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0182] Typically, the following devices may be connected to the I / O interface 305: input devices 306 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; output devices 307 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 308 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 309. The communication devices 309 may allow the electronic device 300 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 3 The electronic device 300 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead. Figure 3 Each block shown in the figure may represent one device, or may represent multiple devices as required.

[0183] Furthermore, the storage medium of the embodiment of the present application stores program instructions that can implement all the above methods, wherein the program instructions can be stored in the above storage medium in the form of a software product, including several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or terminal devices such as a computer, a server, a mobile phone, and a tablet.

[0184] The above descriptions are only some preferred embodiments of the present invention and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present invention is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above inventive concept. For example, the above features are replaced with (but not limited to) technical features with similar functions disclosed in the embodiments of the present invention to form a technical solution.

Claims

1. A modeling method based on oblique photography and drone surveying and mapping, characterized in that: include: Collect UAV flight history data and survey area related information to form a basic database, wherein the historical data includes flight logs, image data and corresponding survey area geographic information; Divide the survey area into regions according to the topography and distribution characteristics of the survey area, and identify the factors affecting the modeling accuracy for different regions of the survey area; Based on the influencing factors of the modeling accuracy, the modeling accuracy of oblique photography and UAV surveying and mapping is predicted to obtain an accuracy reference target; The key influencing factors of the same type of survey area scheme are set based on the influencing factors of the modeling accuracy, the same type of historical measurement information is identified based on the key influencing factors, and then the accuracy historical reference values ​​of the same type of historical measurements are obtained through data analysis; And determine the modeling accuracy prediction target based on the accuracy reference target and the accuracy historical reference value.

2. The modeling method based on oblique photography and drone surveying and mapping according to claim 1 is characterized in that: After determining the modeling accuracy prediction target, it further includes: Combined with regional decomposition, the design accuracy of different regions is compared with the modeling accuracy prediction target, the deviation and impact ratio are determined, and the regions with greater measurement parameter elasticity are adjusted first to control the deviation within a preset error range; And the modeling accuracy of different types of areas to be measured is predicted by using the modeling method based on oblique photography and UAV surveying and mapping.

3. The modeling method based on oblique photography and drone surveying and mapping according to claim 1 is characterized in that: The survey area is divided into regions according to the topography and distribution characteristics of the survey area, and the factors affecting the modeling accuracy are identified for different regions of the survey area, including: Determine the topography and distribution characteristics of the survey area based on the image data and the geographic information of the survey area; Decomposing the geographical part of the survey area into mountainous areas, plains, cities, water areas and special terrain areas according to the topography and distribution characteristics of the landforms in the survey area; And identify the factors affecting the modeling accuracy in the mountainous areas, plains, cities, waters and special terrain areas of the survey area, and the factors affecting the modeling accuracy include terrain undulation, density of land objects, weather conditions, light intensity, wind speed, air quality, surface material, height and area of ​​land objects.

4. The modeling method based on oblique photography and drone surveying and mapping according to claim 3 is characterized in that: Based on the factors affecting modeling accuracy, the modeling accuracy of oblique photography and UAV surveying and mapping is predicted, and the accuracy reference targets include: The modeling accuracy of the mountainous area, plain, city, water area and special terrain area in the survey area is predicted separately to obtain the modeling accuracy of different areas; And the modeling accuracies of the different areas are superimposed to obtain the total modeling accuracy of the survey area as the accuracy reference target.

5. The modeling method based on oblique photography and drone surveying and mapping according to claim 3 is characterized in that: Identifying the same type of historical measurement information based on the key influencing factors includes: Based on the key influencing factors, all sample data are searched, and the improved K-means algorithm is applied to find the number of clusters with the smallest sum of squared error (SSE) value and the corresponding initial cluster centers; After the initial clustering center is determined, the category of the sample data and the center of each category are finally determined according to the improved density peak clustering algorithm, wherein the key influencing factors of the mountainous area include terrain undulation, height of the object and surface material; The key factors affecting the plain include the density of landforms and the surface material; The key influencing factors of the city include land feature density, land feature height and surface material; Key factors affecting the water area include surface texture and wind speed; And the key influencing factors of the special terrain area include terrain undulation, surface material and special terrain attributes.

6. The modeling method based on oblique photography and drone surveying and mapping according to claim 5 is characterized in that: The categories of sample data and the centers of each category are finally determined according to the improved density peak clustering algorithm, which further includes: The objective function of the improved density peak clustering algorithm is: max{G(I)} stG=λ i -oh i Where m is the local neighborhood parameter, λ i Represents the local data association, ω i Represents the shortest path distance to a point with a higher degree of association, I is the set of sample points, i = 1, 2, ..., n, n is the number of data; The local data correlation λ of each data point is calculated in the iterative process i and the shortest path distance to a point with higher association ω i , and obtain G = λ i -ω i , and then finally get the optimal number of clusters, cluster prototypes and the type of sample data with the maximum value of G; Given the local neighborhood parameter m, set the iteration stop threshold δ, initialize the sample point attributes, and set the iteration counter t = 0; if there exists i such that λ i >0, then If there exists i such that λ i =0, then ω i =max{ω j }, j = 1, 2, ..., n: like Then the algorithm stops and outputs the partitioning result and cluster prototype, otherwise let t = t + 1 and go to the previous step; Determine the best classification tree C m , calculate the value of G(C), when the condition (G(C m )=max{G(C)}, then we think that C m is the optimal number of classifications, and the output division result and the clustering prototype are obtained.

7. The modeling method based on oblique photography and drone surveying and mapping according to claim 1 is characterized in that: Historical reference values ​​of accuracy obtained through data analysis for the same type of historical measurements include: Calculate the median M of the sample data d :Sort the sample data from small to large. If the number of data n is an odd number, the median If n is even, then where x (k) Indicates the kth data after sorting; Calculate the mode M o , the mode is the data value that appears most frequently in the sample data; Calculate the range R: R=x max -x min where x max and x min are the maximum and minimum values ​​in the sample data respectively; Calculate the mean absolute deviation MAD: in is the average value of the sample data; the Grubbs criterion is used to eliminate outliers. Suppose a set of measurement data is x1, x2, …, x n , whose arithmetic mean is The standard deviation is s, find the suspicious value x (k) (k=1 or n), calculate If T>G(α,n)(G(α,n)) is the Grubbs coefficient, which is obtained by looking up the Grubbs table based on the given significance level α and the number of data n, then x (k) outliers and remove them, and perform outlier analysis on the sample data; The reasonable interval L of accuracy is calculated by the following formula: Where k1 is a coefficient determined according to the distribution characteristics of the measured data; the sample data after outlier processing is x′1, x′2,…, x′ n′ , then the average Finally, the accuracy historical reference value CKJD of the same type of historical measurement is calculated by the following formula; Among them, C represents the category of historical measurement data, K represents the key factors affecting the measurement accuracy of this category, J represents the judgment and processing process of the data, and D represents the valid data after processing.

8. The modeling method based on oblique photography and drone surveying and mapping according to claim 7 is characterized in that: Determining the modeling accuracy prediction target according to the accuracy reference target and the accuracy historical reference value includes: Assume that the accuracy reference target is CKJC, and the weighting coefficient β ranges from [0,1]; The modeling accuracy prediction target MJ is determined by the following formula: MJ=βCKJD+(1-β)CKJC 9. The modeling method based on oblique photography and drone surveying and mapping according to claim 1 is characterized in that: Based on the influencing factors of modeling accuracy, the modeling accuracy of oblique photography and UAV surveying and mapping is predicted, and the accuracy reference targets include: The method comprises predicting the modeling accuracy of oblique photography and UAV surveying and mapping based on the influencing factors of the modeling accuracy to obtain the accuracy reference target, including predicting the modeling accuracy of oblique photography and UAV surveying and mapping based on the influencing factors of the modeling accuracy using a neural network model optimized based on a particle swarm algorithm to obtain the accuracy reference target, wherein the neural network model maps the input vector to a high-dimensional feature space through the activation function of the neuron, constructs an optimal decision boundary in the high-dimensional feature space, and the output of the neural network is a weighted combination of the outputs of each neuron, wherein each neuron corresponds to an input feature.

10. A modeling device based on oblique photography and drone surveying and mapping, characterized in that: include: A data acquisition module is used to collect the UAV flight history data and survey area related information to form a basic database, wherein the historical data includes flight logs, image data and corresponding survey area geographic information; An influencing factor identification module is used to divide the survey area into regions according to the topography and distribution characteristics of the survey area, and identify the influencing factors of modeling accuracy for different regions of the survey area; A target prediction module predicts the modeling accuracy of oblique photography and UAV surveying and mapping based on the influencing factors of the modeling accuracy to obtain an accuracy reference target; A historical accuracy acquisition module, used to set key influencing factors of the same type of measurement area based on the influencing factors of the modeling accuracy, identify the same type of historical measurement information based on the key influencing factors, and then obtain the accuracy historical reference value of the same type of historical measurement through data analysis; And a prediction target acquisition module is used to determine the modeling accuracy prediction target based on the accuracy reference target and the accuracy historical reference value.

Images