Geological target multi-scale aerial photography acquisition planning method

Abstract

The invention discloses a geological target multi-scale aerial photography acquisition planning method, and relates to the technical field of image recognition. Comprising the following steps: extracting and fitting geometric features by using image data to generate a geometric description matrix of a target area; based on the geometric description matrix and a multi-aerial photography mode matching rule, outputting a mode matching matrix through image analysis, and generating an initial route set; correcting the route in real time, generating a route set adapted to multiple meteorological conditions, and marking the route priority of a high-risk area by analyzing geologic features in the aerial photography image; carrying out sampling point optimization layout on the target area; and driving the unmanned aerial vehicle to complete aerial photography operation, and recording and collecting a data matrix in real time. According to the method, the defects in the aspects of route planning, meteorological adaptability and image data utilization in the prior art are effectively overcome, and a more intelligent and efficient geological target acquisition solution can be provided.

Description

Technical Field

[0001] The invention relates to the technical field of image recognition, and in particular to a multi-scale aerial photography acquisition planning method for geological targets. Background Art

[0002] In geological exploration, aerial images are a key data source, and their applications cover many aspects, such as mineral resource survey, fault and geological structure monitoring, land use and environmental change analysis. With the continuous development of image processing technology, sensor technology and data analysis technology, the quality and analysis accuracy of aerial images have been significantly improved, especially in the fields of multispectral image acquisition, three-dimensional reconstruction and large-scale data processing. However, although the existing technology has made progress in many aspects, it still faces some challenges, especially how to generate accurate aerial photography plans in dynamic environments, how to optimize acquisition routes, and how to effectively integrate multi-source information to improve the accuracy and efficiency of geological target acquisition.

[0003] Most of the existing aerial photography acquisition methods for geological targets rely on fixed route planning or planning based on simple geological models, lacking dynamic adjustment and efficient multi-source data integration. For example, in many traditional aerial photography tasks, route planning is often based on experience or relatively simple geometric parameters, ignoring the diversity and complexity of geological targets. Especially under complex terrain and uncertain meteorological conditions, it is difficult for traditional methods to dynamically adjust routes to cope with environmental changes. In addition, existing methods usually do not fully consider the impact of different meteorological conditions on route selection and data acquisition, resulting in the accuracy and efficiency of aerial photography tasks cannot be effectively guaranteed under severe meteorological conditions. At the same time, although modern remote sensing technology and image processing technology have made significant progress, how to generate the optimal aerial photography path through accurate geometric feature fitting and optimization model is still a technical difficulty. Summary of the invention

[0004] In view of the problems existing in the above background technology, the present invention is proposed.

[0005] Therefore, the problem to be solved by the present invention is how to optimize the layout of flight routes and sampling points under multiple meteorological conditions by means of technical means such as image processing.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides a method for planning multi-scale aerial photography acquisition of geological targets, which includes: S1: using the target geometric parameters in the geological database, calculating the spatial distribution characteristics of the geological targets through a shape parameterization model, and acquiring image data to extract and fit the geometric characteristics to generate a geometric description matrix of the target area; S2: based on the geometric description matrix and multi-aerial photography pattern matching rules, generating an optimal aerial photography model for different target features, and outputting a pattern matching matrix through image analysis, and generating an initial route set; S3: combining weather forecast data with the initial route set, correcting the route in real time, generating a route set that adapts to multiple meteorological conditions, and marking the route priority of high-risk areas by analyzing the geological features in the aerial images; S4: on the basis of the generated route set, optimizing the layout of sampling points in the target area in combination with the acquisition equipment parameters and the geological attributes of the target area; S5: based on the route set and the sampling point layout, driving the drone to complete the aerial photography operation, and recording the acquisition data matrix in real time, including aerial images, aerial photography parameters and corresponding meteorological condition data.

[0008] As a preferred solution of the multi-scale aerial photography acquisition planning method for geological targets of the present invention, S1 includes the following operation steps: extracting basic geometric parameters of the target area from the geological database, combining the edge, contour and texture features extracted from the image data, pre-processing the extracted parameters based on the hierarchical analysis algorithm, and forming a basic parameter matrix P b ; Based on the basic parameter matrix P b , combined with the local geological features extracted from the aerial images, the target area is gridded by regional segmentation to generate a grid unit set G containing local features c ; Combine the geometric features extracted from the aerial image, perform geometric feature fitting on the grid unit set, obtain the shape feature description set of the target area, and calculate the local curvature and direction angle through the shape parameterization model; Based on the shape feature description set, perform cluster analysis on the description set in combination with the target geometric attributes to generate the geometric description matrix M of the target area g .

[0009] As a preferred solution of the multi-scale aerial photography acquisition planning method for geological targets of the present invention, the grid unit set G c Generated by the following formula:

[0010]

[0011] Among them, g i is the area of ​​a single grid unit, A t is the total area of ​​the target area, N g is the number of grid divisions, f im is the adjustment factor based on image data analysis.

[0012] As a preferred solution of the multi-scale aerial photography acquisition planning method for geological targets of the present invention, the geometric feature fitting of the grid unit set includes: geometric feature fitting of the grid unit set, the texture and edge features in the image are used as input to determine the curvature and direction angle of the grid unit:

[0013]

[0014] The local curvature k of the grid cell is i The main direction angle α of the grid cell is calculated by the change of cell height. i Determined by the level change of the unit, F s is the shape feature description set, Δz is the height change, Δx is the horizontal change, Δy is the vertical change, is the adjustment factor related to the height change in the image data, is the adjustment factor related to the direction angle in the image data.

[0015] As a preferred solution of the multi-scale aerial photography acquisition planning method for geological targets of the present invention, the step S2 includes the following steps: extracting the target feature point set T according to the local curvature and azimuth characteristics of the geometric description matrix and the grid unit set of the target area. p ; Based on the extracted target feature point set T p , match the coverage rules of different aerial photography modes, and optimize the aerial photography mode selection through image recognition algorithm to generate the aerial photography mode matching matrix M p ; According to the aerial photography mode matching matrix M p The output result generates the initial route set L s .

[0016] As a preferred solution of the multi-scale aerial photography acquisition planning method for geological targets of the present invention, wherein: the aerial photography pattern matching matrix M p The generation includes: using the target feature point set T p The spatial distribution density and directional characteristics of the aerial photography are used to select the corresponding aerial photography model; the target features extracted from the aerial photography images are combined with the deep learning algorithm to identify the targets and extract the key geological features; based on the image recognition results, the extracted geological features are combined with the geometric description matrix to optimize the selection of the aerial photography mode.

[0017] As a preferred solution of the multi-scale aerial photography acquisition planning method for geological targets of the present invention, S3 includes the following steps: parsing weather forecast data, combining the target area grid unit set G c and the geological features in the aerial images, and calculate the comprehensive weight W of the grid cells affected by weather conditions c(i) and mark the grid cell set of high-impact areas; modify the initial route set L according to the high-impact area set s , adjust the position and interval of route nodes, and generate a route set R1 that adapts to multiple meteorological conditions; according to each route in the revised route, assign route priority according to the comprehensive weight of weather conditions, the proportion of high-impact areas covered by route nodes, and geological characteristics, and mark the route priority in high-risk areas.

[0018] As a preferred solution of the multi-scale aerial photography acquisition planning method for geological targets of the present invention, wherein: the comprehensive weight W of the grid unit affected by weather conditions c The calculation formula of (i) is as follows:

[0019] W c (i) = ω v V i +ω d cos(φ i -φ tar )+ω r P i +ω f W im

[0020] Among them, V i is the wind speed value of the i-th grid unit, φ i is the wind direction angle of the i-th grid unit, φ tar is the target area direction of the i-th grid unit, P i is the rainfall probability of the i-th grid cell, ω v ,ω d ,ω f and ω r is the corresponding weight coefficient, W im is the image feature influence weight of the i-th grid unit; the calculation of the route set R1 adapted to multiple meteorological conditions is as follows:

[0021]

[0022] Among them, R1(k) is the node coordinate of the corrected route k, L s (k) is the node coordinate of route k in the initial route set, G c (k) is the set of grid cells covered by route k, which is used to filter the interaction between the route and the grid. c (i) is the comprehensive weather impact weight of the i-th grid cell, d ik is the distance from the ith grid unit to route k, ε is the distance smoothing factor to avoid division by zero, γ is the route correction intensity factor, is the unit direction vector from the i-th grid unit to the route k, which is used to describe the direction of the route correction; the node coordinates R1(k) of the corrected route k are aggregated to form the route set R1.

[0023] In a second aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of the multi-scale aerial photography acquisition planning method for geological targets as described in the first aspect of the present invention are implemented.

[0024] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, the steps of the multi-scale aerial photography acquisition planning method for geological targets as described in the first aspect of the present invention are implemented.

[0025] The beneficial effects of the present invention are as follows: by extracting target geometric parameters from the geological database, combined with image data extraction and geometric feature fitting, the present invention can more accurately describe the spatial distribution characteristics of the target area; by combining weather forecast data with the initial route set, the route can be adjusted in real time to ensure that the aerial photography task can still be completed efficiently under multiple meteorological conditions. Through multi-aerial photography pattern matching and image analysis, high-risk areas can be marked and the priority of routes can be optimized, thereby further improving the efficiency and safety of aerial photography acquisition.

[0026] At the same time, based on image analysis and the geological properties of the target area, the sampling point layout is optimized to achieve more accurate data collection. In general, the technical solution of the present invention effectively overcomes the deficiencies in route planning, meteorological adaptability and image data utilization in the prior art, and can provide a more intelligent and efficient geological target acquisition solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0028] Figure 1 Flowchart of the method for planning multiscale aerial photogrammetric acquisition for geological targets.

[0029] Figure 2 Survey route geometry models for multiple UAVs. DETAILED DESCRIPTION

[0030] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0031] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0032] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0033] Example 1

[0034] Reference Figure 1-2 , which is the first embodiment of the present invention, and provides a method for planning multi-scale aerial photography acquisition of geological targets, such as Figure 1 As shown, the following steps are included:

[0035] S1: Using the target geometric parameters in the geological database (such as fault length, formation boundary dip, and ore layer width), the spatial distribution characteristics of the geological target are calculated through the shape parameterization model, and image data is collected to extract and fit the geometric features to generate the geometric description matrix M of the target area. g .

[0036] S1.1: Extract the basic geometric parameters of the target area (such as fault length, formation boundary dip, and ore layer width) from the geological database, combine the edge, contour, and texture features extracted from the image data, and pre-process the extracted parameters based on the hierarchical analysis algorithm to form a basic parameter matrix P b .

[0037] Through image data analysis, the geometric feature extraction of geological targets is enhanced to ensure a more accurate description of spatial distribution.

[0038] S1.2: Based on the basic parameter matrix P b , combined with the local geological features extracted from the aerial images, the target area is gridded by regional segmentation to generate a grid unit set G containing local features c , where each grid cell contains geological attribute data.

[0039] Specifically, the spatial variation of geological features is considered in the division process in combination with image features to improve the accuracy of grid division.

[0040] Among them, the grid unit set G containing local features c It can be generated by the following formula:

[0041]

[0042] The size of the grid unit is determined by the target area and the number of grids. i is the area of ​​a single grid unit, A t is the total area of ​​the target area, N g is the number of grid divisions, h im It is an adjustment factor based on image data analysis, reflecting the impact of image characteristics (such as texture density, edge complexity, etc.) on mesh division.

[0043] S1.3: Combine the geometric features extracted from the aerial image to fit the geometric features of the grid unit set, obtain the shape feature description set of the target area, and calculate the local curvature and direction angle through the shape parameterization model.

[0044] Specifically, geometric features are fitted to the grid cell set, and the texture and edge features in the image are used as input to determine the curvature and orientation angle of the grid cell:

[0045]

[0046] The local curvature k of the grid cell is i The main direction angle α of the grid cell is calculated by the change of cell height. i Determined by the level change of the unit, F s is the shape feature description set, Δz is the height change, Δx is the horizontal change, Δy is the vertical change, is the adjustment factor related to the height change in the image data, is the adjustment factor related to the direction angle in the image data.

[0047] It should be noted that the shape feature description set includes the main direction angle, curvature and boundary attributes of each grid unit.

[0048] S1.4: Based on the shape feature description set F s , cluster analysis is performed on the description set in combination with the target geometric attributes to generate the geometric description matrix M of the target area g :

[0049]

[0050] Among them, ω i is the weight of the grid unit, F si is the shape feature description of the i-th grid unit.

[0051] The geometric description matrix contains the geometric classification and boundary information of each feature area.

[0052] Optionally, the results of the clustering analysis can be further optimized through image processing and target detection algorithms (such as target detection methods based on deep learning), so that the geometric description matrix can more accurately reflect the spatial distribution of geological features, especially in complex terrain or areas that are difficult to identify by traditional methods.

[0053] It can be seen that the present invention can accurately locate the spatial distribution of different geological features by decomposing the target area into local grids and fitting and clustering the shapes, providing a reliable geometric basis for the selection of subsequent aerial photography modes.

[0054] S2: Based on the geometric description matrix M g According to the multi-aerial pattern matching rules, the optimal aerial photography model is generated for different target features (such as parallel arrangement, free curve, F-shaped and free bead), and the pattern matching matrix M is output through image analysis. p , and generate the initial route set R0.

[0055] S2.1: Extract the target feature point set based on the local curvature and angular characteristics of the geometric description matrix and the grid cell set of the target area.

[0056] Specifically, in the geometric description matrix M g In the example, we select i Exceeds the threshold and the main direction angle α i The set of grid cells that conform to the change trend of the target area is defined as the target feature point set T p .

[0057] The feature point set is used to locate geometric change information of key target areas.

[0058] S2.2: Based on the extracted target feature point set, match the coverage rules of different aerial photography modes, and further optimize the aerial photography mode selection through image recognition algorithms (such as target detection based on deep learning) to generate the aerial photography mode matching matrix M p .

[0059] Specifically, using the target feature point set T p According to the spatial distribution density and directional characteristics of the drone, the corresponding aerial photography model is selected, such as parallel arrangement, free curve, F-shaped or free beaded drone geometric models. Figure 2 shown.

[0060] Furthermore, combined with the target features extracted from the aerial images, the key geological features in the images are identified through deep convolutional neural networks (CNN) or target detection algorithms (such as YOLO and Faster R-CNN), the selection of aerial photography mode is further optimized, and the matching matrix is ​​adjusted through the image recognition results to adapt to the actual geological characteristics of the target area.

[0061] Specifically, deep learning algorithms (such as CNN) are used to identify targets in aerial images and extract key geological features (such as fault lines, mineral veins, etc.);

[0062] Based on the image recognition results, the extracted geological features are combined with the geometric description matrix to optimize the selection of aerial photography modes, such as parallel arrangement, free curve, and figure-shaped.

[0063] Through image analysis, the pattern matching matrix is ​​further adjusted so that the aerial photography mode can cover key areas more accurately based on the image data, ensuring the comprehensiveness and accuracy of the aerial photography operation.

[0064] S2.3: Generate an initial route set based on the output results of the aerial photography pattern matching matrix.

[0065] Specifically, the aerial photography pattern matching matrix M p The pattern rule and grid unit set G c Perform corresponding mapping to generate the initial route set L s :

[0066]

[0067] Among them, ψ i is the route planning weight, ψ th is the route weight threshold, μ j is the pattern coverage influence coefficient, m is the total number of grid cells in the target area, n is the total number of target feature points extracted, L i A single route in the initial route set.

[0068] The initial route set combines the spatial distribution characteristics of key feature points in the target area and the optimal aerial photography mode to form a diverse route planning scheme.

[0069] It can be seen that the present invention extracts the target feature point set through local curvature and direction angle, and directly extracts the key area information from the geometric description matrix, avoiding the traditional method of relying on manually set area division, improving the extraction efficiency and automation, and proposing aerial photography pattern matching rules based on feature point sets. The target area features are accurately associated with the aerial photography mode through the matching matrix, and the deep learning algorithm is applied to the optimization of the aerial photography pattern matching matrix. The accuracy of aerial photography mode selection is enhanced through image recognition, and the flexibility of multiple mode adaptation is provided; in addition, the generation of the initial route set is combined with the spatial distribution and pattern rules of the target feature points, which effectively avoids the limitations of fixed patterns in conventional route planning and improves the diversity and adaptability of route planning.

[0070] S3: Combine weather forecast data (including wind speed, wind direction, and rainfall probability) with the initial route set R0, correct the route in real time, generate a route set R1 that adapts to multiple meteorological conditions, and mark the route priority in high-risk areas by analyzing the geological features in the aerial images.

[0071] S3.1: Parse weather forecast data and mark high-impact areas.

[0072] Preferably, the wind speed, wind direction, and rainfall probability in the weather forecast data are combined with the target area grid unit set G c and the geological features in the aerial image, and calculate the comprehensive weight W of the unit affected by weather conditions c (i) and mark the set of grid cells in the high-impact area.

[0073] W c (i) = ω v V i +ω d cos(φ i -φ tar )+ω r P i +ω f W im

[0074] Among them, V i is the wind speed value of the i-th grid unit, φ i is the wind direction angle of the i-th grid unit, φ tar is the target area direction of the i-th grid unit, P i is the rainfall probability of the i-th grid cell, ω v ,ω d ,ω f and ω r is the corresponding weight coefficient, W im is the image feature influence weight of the i-th grid unit.

[0075] When Wc (i) When it is greater than the impact threshold, it is considered to be of high impact, and the corresponding grid unit is marked as a high impact area.

[0076] S3.2: Modify the initial route set L according to the high impact area set s , adjust the position and interval of route nodes to generate a route set R1 that adapts to multiple weather conditions.

[0077]

[0078] Among them, R1(k) is the node coordinate of the corrected route k, L s (k) is the node coordinate of route k in the initial route set, G c (k) is the set of grid cells covered by route k, which is used to filter the interaction between the route and the grid. c (i) is the comprehensive weather impact weight of the i-th grid cell, d ik is the distance from the ith grid unit to route k, ε is the distance smoothing factor to avoid division by zero, γ is the route correction intensity factor, is the unit direction vector from the ith grid cell to the route k, used to describe the direction of the route correction.

[0079] By integrating weather weights and distance attenuation effects, the positions of route nodes are dynamically adjusted to achieve route avoidance optimization for high-impact areas.

[0080] The node coordinates R1(k) of the corrected route k are aggregated to form the route set R1.

[0081] S3.3: For each route in the revised route, assign route priority based on the comprehensive weight of weather conditions, the proportion of high-impact areas covered by route nodes, and geological characteristics, and mark the route priority in high-risk areas.

[0082] It should be noted that route adjustments should be made when weather conditions change, especially when planning in complex terrain or dangerous areas. Weighting the impact of geological features through aerial image analysis can ensure that routes can avoid risks to the greatest extent possible when planning in complex terrain or high-risk areas.

[0083] S4: Based on the generated route set, combined with the acquisition equipment parameters and the geological properties of the target area, the sampling points in the target area are optimized to ensure the effectiveness of data collection.

[0084] Specifically, based on the generated route set, the sampling points in the target area are optimized according to the specific parameters of the acquisition equipment (such as sensor type, resolution, effective ranging range, etc.) and the geological properties of the target area (such as geological structure, surface features, terrain complexity, etc.).

[0085] The optimization process takes into account the following factors:

[0086] Uniformity of spatial distribution of sampling points to ensure adequate coverage throughout the target area;

[0087] Select key sampling areas based on geological attributes to improve the pertinence and effectiveness of data collection;

[0088] Determine the spacing and density between sampling points based on equipment performance to avoid data redundancy or collection blind spots.

[0089] Finally, a sampling point layout matrix is ​​formed, which clearly marks the geographical coordinates and distribution of each sampling point.

[0090] S5: Based on the route set and sampling point layout, drive the UAV to complete the aerial photography operation and record the collected data matrix in real time, including aerial images, aerial photography parameters and corresponding meteorological conditions data for subsequent geological analysis.

[0091] Specifically, based on the generated route set and sampling point layout matrix, the drone is dispatched to perform aerial photography operations according to the predetermined route.

[0092] During the operation, the drone collects and records relevant data in real time, including aerial images (image data covering the target area), aerial parameters (such as aerial altitude, speed, inclination, shooting time, etc.) and corresponding meteorological conditions data (such as temperature, humidity, wind speed and other environmental parameters).

[0093] This data is processed and stored in real time through edge computing devices or onboard sensors to ensure data integrity and accuracy.

[0094] This embodiment also provides a computer device, which is suitable for the case of a multi-scale aerial photography acquisition planning method for geological targets, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the multi-scale aerial photography acquisition planning method for geological targets proposed in the above embodiment.

[0095] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0096] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for realizing multi-scale aerial photography acquisition planning of geological targets as proposed in the above embodiment is implemented.

[0097] In summary, the present invention can more accurately describe the spatial distribution characteristics of the target area by extracting target geometric parameters from the geological database, combining image data extraction and geometric feature fitting; by combining weather forecast data with the initial route set, the route can be adjusted in real time to ensure that the aerial photography task can still be completed efficiently under multiple meteorological conditions. Through multi-aerial photography pattern matching and image analysis, high-risk areas can be marked and the priority of routes can be optimized, thereby further improving the efficiency and safety of aerial photography acquisition.

[0098] At the same time, based on image analysis and the geological properties of the target area, the sampling point layout is optimized to achieve more accurate data collection. In general, the technical solution of the present invention effectively overcomes the deficiencies in route planning, meteorological adaptability and image data utilization in the prior art, and can provide a more intelligent and efficient geological target acquisition solution.

[0099] Example 2

[0100] Referring to Table 1, which is the second embodiment of the present invention, this embodiment provides a method for planning multi-scale aerial photography acquisition of geological targets. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through simulation experiments.

[0101] The experiment selected a mountainous area with complex geological structure as the target area, where there are multiple faults, different stratum dip angle changes and diverse mineral layer distribution.

[0102] A detailed geological database of the target area was obtained from the local geological department, which contained basic geometric parameters such as fault length, formation boundary dip, and ore layer width. These data were accumulated through years of geological exploration and mapping.

[0103] The target area was photographed multiple times using high-precision aerial photography equipment, and high-resolution aerial images were obtained as the source of image data.

[0104] The length data of 50 faults in the target area, ranging from 100 to 1000 meters, the dip angle data of stratigraphic boundaries, ranging from 20° to 70°, and the width data of 30 mineral layers, ranging from 5 to 50 meters, were extracted from the geological database.

[0105] The aerial images are analyzed, and clear edge features are extracted by edge detection algorithm, and texture features are obtained by texture analysis algorithm. These image features are combined with basic geometric parameters extracted from the geological database and preprocessed by hierarchical analysis algorithm. For example, for fault length data, correction is performed based on its edge continuity and texture changes in the image. Finally, a 50×3 basic parameter matrix is ​​formed.

[0106] Based on the above basic parameter matrix, the size of the grid unit is calculated using the formula according to the target area (measured as 10 square kilometers) and the set number of grids (1000). Considering the complex geological characteristics of the region, the adjustment factor for image data analysis is determined to be 0.8 through the analysis of the texture density and edge complexity of the aerial images. This generates a set of grid units containing local features, each with an area of ​​about 10,000 square meters, and combines local geological features, such as finer grid division in dense texture areas.

[0107] The geometric features of the grid cell set are fitted using the shape parameterization model. Taking one of the grid cells as an example, by analyzing its height change (assumed to be 5 meters) and horizontal change (assumed to be 3 meters) in the aerial image, combined with the adjustment factor related to the height change (assumed to be 0.6) and the adjustment factor related to the direction angle (assumed to be 0.7) obtained from the image data, the local curvature and main direction angle of the grid cell are calculated using the formula. The shape feature description set is obtained, which includes the main direction angle, curvature and boundary attributes of each grid cell.

[0108] Cluster analysis is performed based on the shape feature description set and the target geometric attributes. For example, grid cells with similar curvature and orientation angles are clustered together to generate a geometric description matrix for the target area. The clustering results are optimized through a deep learning-based target detection method, so that the matrix can more accurately reflect the spatial distribution of geological features.

[0109] In the geometric description matrix, a set of grid cells with a local curvature greater than 0.3 and a main direction angle that conforms to the change trend of the target area is selected as the target feature point set. After screening, 50 target feature points are obtained, which can well locate the geometric change information of the key target area, such as the intersection of faults.

[0110] Based on the extracted target feature point set, the aerial photography mode is matched according to its spatial distribution density and directional characteristics. For example, in areas where the feature points are evenly distributed and have consistent directions, the parallel arrangement aerial photography mode is selected; in areas where the feature points are distributed in a curve, the free curve aerial photography mode is selected. At the same time, the deep convolutional neural network (CNN) is used to identify the target of the aerial image and extract key geological features, such as 10 obvious fault lines and 5 mineral veins. Based on these recognition results, the selection of aerial photography mode is further optimized to generate an aerial photography mode matching matrix.

[0111] According to the output of the aerial photography pattern matching matrix, the pattern rules are mapped to the grid unit set. For example, for the grid cells corresponding to the parallel aerial photography pattern, routes are generated according to a certain spacing and direction. Finally, an initial route set is generated, which contains 20 different initial routes.

[0112] Obtain the weather forecast data for the target area for the next week, including wind speed (range 0-10m / s), wind direction (0°-360°) and rainfall probability (0-100%). Combine the target area grid cell set and the geological features in the aerial image to calculate the comprehensive weight of each grid cell affected by weather conditions. When the comprehensive weight is greater than 0.5, mark the grid cell as a high-impact area. After calculation, 200 grid cells are marked as high-impact areas, which are mainly concentrated in places such as mountain tops and valleys with high wind speeds and high rainfall probabilities.

[0113] The initial route set is modified according to the set of high-impact areas. For example, for route nodes close to high-impact areas, the formula is used to adjust the node spacing and path direction according to their comprehensive weather impact weights and distance attenuation effects. The route correction intensity factor is set to 0.3 and the distance smoothing factor is set to 0.1. After modification, the route can effectively avoid high-impact areas and generate a route set that adapts to multiple weather conditions.

[0114] According to each route in the revised route, the route priority is assigned based on the comprehensive weight of weather conditions, the proportion of high-impact areas covered by route nodes, and geological characteristics. For example, for routes that pass through high-risk fault areas and are greatly affected by weather conditions, they are given a higher priority so that they can be focused on during aerial photography operations.

[0115] According to the parameters of the acquisition equipment, the sensor resolution of the drone used this time is 0.1 meters, and the effective ranging range is 500 meters. Combined with the geological properties of the target area, such as complex geological structures and diverse surface features, the sampling points in the target area are optimized. Considering the uniformity of the spatial distribution of the sampling points, 500 sampling points are evenly distributed throughout the 10 square kilometers of the target area. At the same time, the sampling points are encrypted in key areas such as faults and mineral layers to ensure the pertinence and effectiveness of data collection. According to the performance of the equipment, the spacing between the sampling points is controlled between 10-50 meters to form a sampling point layout matrix.

[0116] Based on the generated route set and sampling point layout matrix, the drone is dispatched to perform aerial photography operations. The drone flies according to the predetermined route. During the flight, it collects and records aerial images, aerial photography parameters (such as aerial photography altitude between 100-500 meters, speed between 10-20m / s, inclination between 0°-30°, shooting time interval of 5 seconds) and corresponding meteorological conditions data (temperature between 10-30℃, humidity between 30%-70%, wind speed between 0-10m / s) in real time. These data are processed and stored in real time through airborne sensors to ensure data integrity and accuracy.

[0117] Table 1 Comparison of key parameters of geological aerial photography test area

[0118]

[0119] It can be seen from the fault length, formation boundary dip angle and ore layer width data in Table 1 that the present invention can effectively process different geological features. For example, the fault length of area 3 is 450 meters, the formation boundary dip angle is 50°, and the ore layer width is 20 meters. Through the method of the present invention, these features can be accurately reflected in the geometric description matrix. Compared with the traditional geological aerial photography method, the traditional method may simply record these data, but cannot be closely combined with the aerial photography mode like the present invention.

[0120] The data of the grid unit area reflects the accuracy of the present invention in grid division. The grid unit area of ​​each area is between 9800 and 10200 square meters, which is the result obtained after considering the spatial variation of geological characteristics and the adjustment factor of image data analysis. The traditional grid division method is usually based on fixed rules and cannot be dynamically adjusted according to the actual situation of geological characteristics. The present invention can flexibly divide the grid according to the complexity of geology and the characteristics of aerial photography images, ensuring the accuracy of subsequent aerial photography pattern matching.

[0121] The data of local curvature and main direction angle reflect the advantages of the present invention in geometric feature fitting. The local curvature and main direction angle of different regions are different. These data are calculated by combining the shape parameterization model with the texture and edge features in the aerial image. In the prior art, the geometric feature fitting of geological targets is often not accurate enough, but the present invention can accurately obtain these key geometric features through precise calculation and image feature assistance, providing a reliable basis for the selection of aerial photography mode.

[0122] The data of weather comprehensive weight and route priority demonstrate the innovation of the present invention in route planning under multiple weather conditions. For example, the weather comprehensive weight of region 3 is 0.5 and the route priority is 1, which indicates that the route in this region is given priority adjustment when the weather has a greater impact. Traditional route planning methods are often difficult to dynamically adjust according to actual weather conditions and geological characteristics when faced with complex meteorological conditions. However, the present invention can reasonably allocate route priorities and effectively avoid high-risk areas by comprehensively considering weather data and geological characteristics, thus ensuring the safety and efficiency of aerial photography operations.

[0123] In general, the present invention forms a complete geological aerial photography scheme by organically combining the geometric characteristics of geological targets with aerial photography modes, meteorological conditions and sampling point layout. Compared with the prior art, the present invention has significant advantages in terms of accuracy of geological feature extraction, accuracy of aerial photography mode matching, flexibility of route planning under multiple meteorological conditions and rationality of sampling point layout, and can provide more comprehensive and accurate data for geological analysis.

[0124] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for planning multi-scale aerial photography acquisition of geological targets, characterized by: include: S1: Using the target geometric parameters in the geological database, the spatial distribution characteristics of the geological target are calculated through the shape parameterization model, and image data is collected to extract and fit the geometric features to generate a geometric description matrix of the target area; S2: Based on the geometric description matrix and multi-aerial photography pattern matching rules, an optimal aerial photography model is generated for different target features, and through image analysis, a pattern matching matrix is ​​output and an initial route set is generated; S3: combining weather forecast data with the initial route set, correcting the route in real time, generating a route set adapted to multiple meteorological conditions, and marking the route priority of high-risk areas by analyzing geological features in the aerial images; S4: Based on the generated route set, the sampling points in the target area are optimized and arranged in combination with the acquisition equipment parameters and the geological properties of the target area; S5: Based on the route set and sampling point layout, drive the UAV to complete the aerial photography operation and record the collected data matrix in real time, including aerial images, aerial photography parameters and corresponding meteorological conditions data.

2. The method for planning geological target multi-scale aerial photography acquisition according to claim 1, characterized in that: The S1 comprises the following steps: The basic geometric parameters of the target area are extracted from the geological database, and the edge, contour and texture features extracted from the image data are combined. The extracted parameters are preprocessed based on the hierarchical analysis algorithm to form the basic parameter matrix P b ; Based on the basic parameter matrix P b , combined with the local geological features extracted from the aerial images, the target area is gridded by regional segmentation to generate a grid unit set G containing local features c ; Combined with the geometric features extracted from the aerial images, the geometric features of the grid unit set are fitted to obtain the shape feature description set of the target area, and the local curvature and direction angle are calculated through the shape parameterization model; Based on the shape feature description set, cluster analysis is performed on the description set in combination with the target geometric attributes to generate a geometric description matrix M of the target area. g .

3. The method for planning geological target multi-scale aerial photography acquisition according to claim 2, characterized in that: The grid unit set G c Generated by the following formula: Among them, g i is the area of ​​a single grid unit, A t is the total area of ​​the target area, N g is the number of grid divisions, f im is the adjustment factor based on image data analysis.

4. The method for planning geological target multi-scale aerial photography acquisition according to claim 3, characterized in that: The geometric feature fitting of the grid unit set comprises: Geometric features are fitted to the grid cell set. The texture and edge features in the image are used as input to determine the curvature and orientation angle of the grid cell: The local curvature k of the grid cell is i The main direction angle α of the grid cell is calculated by the change of cell height. i Determined by the level change of the unit, F s is the shape feature description set, Δz is the height change, Δx is the horizontal change, Δy is the vertical change, is the adjustment factor related to the height change in the image data, is the adjustment factor related to the direction angle in the image data.

5. The method for planning the multi-scale aerial photography acquisition of geological targets according to claim 4, characterized in that: The S2 comprises the following steps: According to the local curvature and angular characteristics of the geometric description matrix, combined with the grid unit set of the target area, the target feature point set T is extracted p ; Based on the extracted target feature point set T p , match the coverage rules of different aerial photography modes, and optimize the aerial photography mode selection through image recognition algorithm to generate the aerial photography mode matching matrix M p ; According to the aerial photography mode matching matrix M p The output result generates the initial route set L s .

6. The method for planning geological target multi-scale aerial photography acquisition according to claim 5, characterized in that: The aerial photography pattern matching matrix M p The generation includes: Using the target feature point set T p According to the spatial distribution density and directional characteristics, select the corresponding aerial photography model; Combined with the target features extracted from aerial images, deep learning algorithms are used to identify targets in aerial images and extract key geological features; Based on the image recognition results, the extracted geological features are combined with the geometric description matrix to optimize the selection of aerial photography mode.

7. The method for planning geological target multi-scale aerial photography acquisition according to claim 6, characterized in that: The S3 comprises the following steps: Analyze weather forecast data and combine it with the target area grid unit set G c and the geological features in the aerial images, and calculate the comprehensive weight W of the grid cells affected by weather conditions c (i) and mark the set of grid cells in high-impact areas; Modify the initial route set L according to the high impact area set s , adjust the position and interval of route nodes to generate a route set R1 that adapts to multiple weather conditions; For each route in the revised route, route priority is assigned according to the comprehensive weight of weather conditions, the proportion of high-impact areas covered by route nodes and geological characteristics, and the route priority in high-risk areas is marked.

8. The method for planning geological target multi-scale aerial photography acquisition according to claim 6, characterized in that: The comprehensive weight W of the grid unit affected by weather conditions c The calculation formula of (i) is as follows: W c (i)=ω v V i +oh d cos(φ i -f tar )+ω r P i +oh f W im Among them, V i is the wind speed value of the i-th grid unit, φ i is the wind direction angle of the i-th grid unit, φ tar is the target area direction of the i-th grid unit, P i is the rainfall probability of the i-th grid cell, ω v ,ω d ,ω f and ω r is the corresponding weight coefficient, W im is the image feature influence weight of the i-th grid unit; The calculation of the route set R1 adapted to multiple weather conditions is as follows: Among them, R1(k) is the node coordinate of the corrected route k, L s (k) is the node coordinate of route k in the initial route set, G c (k) is the set of grid cells covered by route k, which is used to filter the interaction between the route and the grid. c (i) is the comprehensive weather impact weight of the i-th grid cell, d ik is the distance from the ith grid unit to route k, ε is the distance smoothing factor to avoid division by zero, γ is the route correction intensity factor, is the unit direction vector from the ith grid cell to the route k, used to describe the direction of the route correction; The node coordinates R1(k) of the corrected route k are aggregated to form the route set R1.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for multi-scale aerial photography acquisition planning of geological targets described in any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for multi-scale aerial photography acquisition planning of geological targets described in any one of claims 1 to 8 are implemented.

Images