Unmanned aerial vehicle multi-data fusion geological disaster detection method and system

Abstract

The invention relates to the field of geological disaster detection, and discloses an unmanned aerial vehicle multi-data fusion geological disaster detection method and system, and the method comprises the steps: firstly obtaining an unmanned aerial vehicle initial image through the collection of an unmanned aerial vehicle, carrying out the marking, obtaining a three-dimensional image marking element, and building an environment training set large model; inputting the three-dimensional image marking elements into an environment training set large model to obtain a first geological disaster detection result; performing geological disaster prediction according to the environment training set large model to obtain an unmanned aerial vehicle geological disaster prediction image, converting the unmanned aerial vehicle geological disaster prediction image into an unmanned aerial vehicle index feature image, and performing big data analysis to obtain a second geological disaster detection result; comparing and verifying the first detection result and the second detection result to obtain an unmanned aerial vehicle image geological disaster detection result; the problems that in the prior art, integration analysis of various data is not considered, the geological disaster detection effect is poor, and geological disaster data analysis is simple are solved.

Description

Technical Field

[0001] This invention relates to the field of geological disaster detection, and in particular to a geological disaster detection method and system based on multi-data fusion from unmanned aerial vehicles (UAVs). Background Technology

[0002] With the development of drone technology, drones have gradually been integrated into various industries, greatly improving people's production speed. Especially in the field of environmental monitoring, drones play an irreplaceable role. However, in the field of geological disaster detection, drones play a very small role, and there is no technology to use drones to detect geological disasters. Currently, existing geological hazard detection technologies mainly rely on ground-mounted equipment to detect changes and vibrations in the geological magnetic field. However, the results are generally poor, and the technology is relatively outdated. It does not involve simulating various scenarios or combining multiple data for analysis and simulation. Summary of the Invention

[0003] The present invention aims to provide a geological disaster detection method and system based on multi-data fusion from unmanned aerial vehicles (UAVs), which solves the problems of existing technologies that do not consider the integration and analysis of multiple data, resulting in poor geological disaster detection effects and relatively simple geological disaster data analysis.

[0004] To achieve the above objectives, the present invention provides the following method: This invention provides a method for geological disaster detection using UAV multi-data fusion: S1: Obtain the actual situation of the target area through drones and collect initial drone images; S2: Establish a three-dimensional image coordinate system based on the initial image of the UAV, and mark the elements in the initial image of the UAV to obtain the three-dimensional image marked elements; S3: Perform cluster analysis on historical images of the target area acquired by the UAV to obtain a training set of the target area and establish a large model of the environmental training set; S4: Input the three-dimensional image marker elements into the environmental training set large model to perform the first geological disaster detection process and obtain the first geological disaster detection result; S5: Based on the large model of the environmental training set, geological disaster prediction is performed. The geological disaster prediction results are combined with historical images of the target area for preprocessing to obtain UAV geological disaster prediction images. S6: Convert the UAV geological disaster prediction image into a UAV index feature image, and perform big data analysis on the UAV index feature image and the UAV initial image to obtain the second geological disaster detection result; S7: Compare and verify the first geological disaster detection result and the second geological disaster detection result to obtain the UAV image geological disaster detection result.

[0005] Preferably, the step of establishing a three-dimensional image coordinate system based on the initial image of the UAV and marking the elements in the initial image of the UAV to obtain three-dimensional image marked elements includes: establishing a three-dimensional image coordinate system with the lowest center position of the target area in the initial image of the UAV as the origin of the three-dimensional image coordinate system; traversing the pixel values ​​of all pixels in the initial image of the UAV and determining the area to be marked based on the pixel values; obtaining the first minimum and first maximum values ​​of the pixels in the area to be marked on the X-axis, and obtaining the second minimum and second maximum values ​​of the pixels in the area to be marked on the Y-axis; obtaining four vertex coordinate information based on the first minimum, first maximum, second minimum, and second maximum values, and obtaining a rectangular marking box corresponding to the area to be marked based on the four vertex coordinate information; marking fixed elements in the rectangular marking box, recording the position and size of the fixed elements, and mapping the fixed elements to the Z-axis of the three-dimensional image coordinate system to obtain the three-dimensional image marked elements.

[0006] Preferably, the step of performing cluster analysis on historical images of the target area acquired by the UAV to obtain a training set for the target area and establish a large-scale environmental training set model includes: performing grayscale stretching on the historical images of the target area; constructing an HS color histogram and a gradient direction histogram on the grayscale stretched reference transmission line images; obtaining the corresponding probability distributions based on the HS color histogram and gradient direction histogram; selecting initial cluster centers after grayscale stretching of the historical images of the target area; determining the number of initial cluster centers; performing JSD calculation on each of the reference transmission line images; obtaining the target area training set based on the probability distribution; and obtaining a large-scale environmental training set model.

[0007] Preferably, the step of inputting the three-dimensional image marker elements into the environmental training set large model for the first geological disaster detection process to obtain the first geological disaster detection result includes: aligning the three-dimensional image marker elements with the historical images of the target area according to the origin of the three-dimensional image coordinate system; determining the number of elements at the historical corresponding positions of the three-dimensional image marker elements in the historical images of the target area; if the number of deviations between the three-dimensional image marker elements and their historical corresponding positions is >1 / 2, the severity of the geological disaster is moderate; if the number of deviations between the three-dimensional image marker elements and their historical corresponding positions is <=1 / 2, the severity of the geological disaster is mild.

[0008] Preferably, after determining the number of elements in the historical corresponding positions of the three-dimensional image marker elements and the historical images of the target area, the method further includes: connecting each three-dimensional image marker element in the three-dimensional image coordinate system with the origin of the three-dimensional image coordinate system to obtain a connecting line of the three-dimensional image marker elements; calculating the angle between each connecting line of the three-dimensional image marker elements and the X-axis of the three-dimensional image coordinate system to obtain an offset angle, and calculating the average offset angle of the connecting lines of the three-dimensional image marker elements; based on the average offset angle of the connecting lines of the three-dimensional image marker elements, drawing a three-dimensional image marker element offset line from the origin of the three-dimensional image coordinate system; calculating the historical three-dimensional image marker element offset line of the target area in the same way as the three-dimensional image marker element offset line; determining the angle difference between the three-dimensional image marker element offset line and the historical image three-dimensional image marker element offset line; if the angle difference between the three-dimensional image marker element offset line and the historical image three-dimensional image marker element offset line of the target area is <30°... o If the angle difference between the offset line of the three-dimensional image marker element and the offset line of the three-dimensional image marker element in the historical image of the target area is >= 30°, then it is considered a moderate disaster situation. o If so, it is considered a severe disaster situation.

[0009] Preferably, the step of performing geological disaster prediction based on the large model of the environmental training set, and preprocessing the geological disaster prediction results with historical images of the target area to obtain UAV geological disaster prediction images includes: parsing the historical images of the target area into continuous image frames in chronological order, and converting each image frame into a grayscale image; calculating the motion speed and direction of pixels between every two image frames based on the optical flow velocity method to obtain a geological disaster prediction motion map of the target area; simulating the geological disaster target area change process by combining the geological disaster prediction motion map with the historical images of the target area; and performing image conversion on the geological disaster target area change process to obtain UAV geological disaster prediction images.

[0010] Preferably, the step of converting the UAV geological disaster prediction image into a UAV exponential feature image includes: performing frame-by-frame analysis on the UAV geological disaster prediction image, marking different elements in each frame according to different labeling signals to obtain an element-labeled prediction image; performing element motion simulation on the element-labeled prediction image according to the UAV geological disaster prediction image, and recording the motion trajectory of the same element in the element-labeled prediction image; recording the motion trajectory of the same element in the element-labeled prediction image, and converting it into exponential feature data to obtain a UAV exponential feature image.

[0011] Preferably, the step of performing big data analysis on the UAV index feature image and the UAV initial image to obtain the second geological disaster detection result includes: marking different elements in the UAV initial image and simulating geological disasters according to a big data model; recording the movement trajectory of different elements during each geological disaster simulation to obtain the element simulation trajectory; determining the overlap between the element simulation trajectory and the UAV index feature image; if the overlap between the element simulation trajectory and the UAV index feature image is greater than 70%, it is a severe disaster situation; if the overlap between the element simulation trajectory and the UAV index feature image is greater than 40% and less than or equal to 70%, it is a moderate disaster situation; if the overlap between the element simulation trajectory and the UAV index feature image is less than or equal to 40%, it is a minor disaster situation.

[0012] Preferably, the step of comparing and verifying the first geological disaster detection result and the second geological disaster detection result to obtain the UAV image geological disaster detection result includes: comparing the first geological disaster detection result and the second geological disaster detection result; if the difference between the first geological disaster detection result and the second geological disaster detection result is less than 20%, then the first geological disaster detection result and the second geological disaster detection result are compared and optimized according to historical geological disaster results to obtain the UAV image geological disaster detection result; if the difference between the first geological disaster detection result and the second geological disaster detection result is greater than 20%, then the first geological disaster detection result and the second geological disaster detection result are input into a big data platform for geological disaster data simulation optimization to obtain the UAV image geological disaster detection result.

[0013] This invention provides a geological disaster detection system based on UAV multi-data fusion, characterized in that the system comprises: The image acquisition module is used to obtain the actual situation of the target area through the drone and acquire the initial image of the drone. The element marking module is used to establish a three-dimensional image coordinate system based on the initial image of the UAV, mark the elements in the initial image of the UAV, and obtain three-dimensional image marked elements. The large model training module is used to perform cluster analysis on historical images of the target area acquired by the UAV, obtain the target area training set, and build a large model of the environment training set. The first detection module is used to input the three-dimensional image marker elements into the environmental training set large model to perform the first geological disaster detection process and obtain the first geological disaster detection result. The disaster prediction module is used to predict geological disasters based on the large model of the environmental training set, and to preprocess the geological disaster prediction results by combining them with historical images of the target area to obtain UAV geological disaster prediction images. The second detection module is used to convert the UAV geological disaster prediction image into a UAV index feature image, and to perform big data analysis on the UAV index feature image and the UAV initial image to obtain the second geological disaster detection result. The data integration module is used to compare and verify the first geological disaster detection result and the second geological disaster detection result to obtain the UAV image geological disaster detection result.

[0014] The beneficial effects of this invention are as follows: This invention establishes a three-dimensional image coordinate system based on initial UAV images and marks relevant elements. The movement trajectory of elements is recorded through element marking. A large environmental training set model is established based on historical images. The movement traces of different elements are monitored based on the coordinate system, which can improve the detection effect of geological disasters. The geological disaster level is detected by judging the number of elements and the element offset before and after the geological disaster. This is better than the existing method of judging the vibration intensity and visual disaster conditions of geological disasters. It is more systematic and can help rescuers carry out corresponding rescue operations under different disaster conditions. Then, geological disaster detection is carried out by multi-data fusion based on big data simulation and repeated verification of disaster conditions, making the geological disaster detection results more accurate. Attached Figure Description

[0015] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0016] Figure 1 A flowchart illustrating a method for geological disaster detection using multi-data fusion from unmanned aerial vehicles (UAVs) provided in an embodiment of the present invention; Figure 2 This is a flowchart illustrating a geological disaster detection system based on multi-data fusion from an unmanned aerial vehicle (UAV) according to an embodiment of the present invention. Detailed Implementation

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

[0018] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this invention are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, apparatus, product, or end that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or ends.

[0019] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0020] Currently, existing geological hazard detection technologies mainly rely on ground-mounted equipment to detect changes and vibrations in the geological magnetic field. However, the results are generally poor, and the technology is relatively outdated. It does not involve simulating various scenarios or combining multiple data for analysis and simulation.

[0021] The present invention aims to provide a geological disaster detection method and system based on multi-data fusion from unmanned aerial vehicles (UAVs), which solves the problems of existing technologies that do not consider the integration and analysis of multiple data, resulting in poor geological disaster detection effects and relatively simple geological disaster data analysis.

[0022] like Figure 1 As shown in the figure, a specific embodiment of the present invention provides a geological disaster detection method based on UAV multi-data fusion, comprising the following steps: S1: Obtain the actual situation of the target area through drones and collect initial drone images.

[0023] S2: Establish a three-dimensional image coordinate system based on the initial UAV image, mark the elements in the initial UAV image, and obtain the three-dimensional image marked elements.

[0024] In this embodiment of the invention, a three-dimensional image coordinate system is established with the lowest center position of the target area in the initial image of the UAV as the origin of the three-dimensional image coordinate system; the pixel values ​​of all pixels in the initial image of the UAV are traversed, and the area to be marked is determined based on the pixel values; the first minimum value and the first maximum value of the pixels in the area to be marked on the X-axis are obtained, and the second minimum value and the second maximum value of the pixels in the area to be marked on the Y-axis are obtained; the coordinate information of four vertices is obtained based on the first minimum value, the first maximum value, the second minimum value, and the second maximum value, and the rectangular marking box corresponding to the area to be marked is obtained based on the coordinate information of the four vertices; fixed elements are marked in the rectangular marking box, the position and size of the fixed elements are recorded, and the fixed elements are mapped to the Z-axis of the three-dimensional image coordinate system to obtain the three-dimensional image marking elements.

[0025] S3: Perform cluster analysis on historical images of the target area acquired by the UAV to obtain a training set of the target area and establish a large model of the environmental training set.

[0026] In this embodiment of the invention, the historical images of the target area are subjected to grayscale stretching; an HS color histogram and a gradient direction histogram are constructed on the grayscale stretched reference transmission line images; the corresponding probability distributions are obtained according to the HS color histogram and the gradient direction histogram; the initial cluster center points of the grayscale stretched historical images of the target area are selected; the number of initial cluster center points is determined; JSD calculation is performed on each reference transmission line image, and the target area training set is obtained according to the probability distribution to obtain a large model of the environment training set.

[0027] S4: Input the 3D image marker elements into the large environmental training set model to perform the first geological disaster detection process and obtain the first geological disaster detection result.

[0028] In this embodiment of the invention, the three-dimensional image marker elements are aligned with the historical images of the target area based on the origin of the three-dimensional image coordinate system; the number of elements is determined by comparing the three-dimensional image marker elements with the historical corresponding positions of the three-dimensional image marker elements in the historical images of the target area; if the deviation between the three-dimensional image marker elements and their historical corresponding positions is greater than 1 / 2, the severity of the geological disaster is classified as moderate; if the deviation between the three-dimensional image marker elements and their historical corresponding positions is less than or equal to 1 / 2, the severity of the geological disaster is classified as minor; after determining the number of elements by comparing the three-dimensional image marker elements with the historical corresponding positions of the three-dimensional image marker elements in the historical images of the target area, the method further includes: connecting each three-dimensional image marker element in the three-dimensional image coordinate system with the origin of the three-dimensional image coordinate system. The process involves: obtaining the connecting lines of 3D image marker elements; calculating the angle between each connecting line and the X-axis of the 3D image coordinate system to obtain the offset angle, and calculating the average offset angle of the connecting lines; based on the average offset angle of the connecting lines, depicting the offset lines of the 3D image marker elements from the origin of the 3D image coordinate system; calculating the historical 3D image marker element offset lines of the target area using the same method; determining the angle difference between the 3D image marker element offset lines and the historical 3D image marker element offset lines; and determining if the angle difference between the 3D image marker element offset lines and the historical 3D image marker element offset lines of the target area is <30°. o If the angle difference between the offset line of the 3D image marker element and the offset line of the 3D image marker element in the historical image of the target area is >= 30°, then it is considered a moderate disaster situation; o If so, it is considered a severe disaster situation.

[0029] S5: Geological disaster prediction is performed based on the large model of the environmental training set. The geological disaster prediction results are combined with historical images of the target area for preprocessing to obtain UAV geological disaster prediction images.

[0030] In this embodiment of the invention, historical images of the target area are parsed into continuous image frames in chronological order, and each image frame is converted into a grayscale image; the motion speed and direction of pixels between every two image frames are calculated based on the optical flow velocity method to obtain a geological disaster prediction motion map of the target area; the geological disaster prediction motion map is combined with historical images of the target area to simulate the change process of the geological disaster target area; the change process of the geological disaster target area is converted into an image to obtain a UAV geological disaster prediction image.

[0031] S6: Convert UAV geological disaster prediction images into UAV index feature images, perform big data analysis on the UAV index feature images and the initial UAV images to obtain the second detection result of geological disasters.

[0032] In this embodiment of the invention, the UAV geological disaster prediction image is analyzed frame by frame. Different elements in each frame are marked according to different labeling signals to obtain element-labeled prediction images. The element-labeled prediction images are used to simulate element motion according to the UAV geological disaster prediction images, and the motion trajectories of the same element in the element-labeled prediction images are recorded. The motion trajectories of the same element in the element-labeled prediction images are recorded and converted into exponential feature data to obtain UAV exponential feature images. Different elements in the initial UAV image are labeled, and geological disaster simulation is performed according to a big data model. The motion trajectories of different elements are recorded during each geological disaster simulation to obtain element simulation trajectories. The overlap between the element simulation trajectory and the UAV exponential feature images is determined. If the overlap between the element simulation trajectory and the UAV exponential feature images is greater than 70%, it indicates a severe disaster situation. If the overlap between the element simulation trajectory and the UAV exponential feature images is greater than 40% and less than or equal to 70%, it indicates a moderate disaster situation. If the overlap between the element simulation trajectory and the UAV exponential feature images is less than or equal to 40%, it indicates a minor disaster situation.

[0033] S7: Compare and verify the first and second geological disaster detection results to obtain the UAV image geological disaster detection results.

[0034] In this embodiment of the invention, the first geological disaster detection result and the second geological disaster detection result are compared. If the difference between the first geological disaster detection result and the second geological disaster detection result is less than 20%, the first geological disaster detection result and the second geological disaster detection result are compared and optimized according to historical geological disaster results to obtain the UAV image geological disaster detection result. If the difference between the first geological disaster detection result and the second geological disaster detection result is greater than 20%, the first geological disaster detection result and the second geological disaster detection result are input into a big data platform for geological disaster data simulation and optimization to obtain the UAV image geological disaster detection result.

[0035] like Figure 2 As shown, the present invention provides a geological disaster detection system based on UAV multi-data fusion, characterized in that the system includes: The image acquisition module is used to obtain the actual situation of the target area through the drone and acquire the initial image of the drone. The element marking module is used to establish a three-dimensional image coordinate system based on the initial image of the UAV, mark the elements in the initial image of the UAV, and obtain three-dimensional image marked elements. The large model training module is used to perform cluster analysis on historical images of the target area acquired by the UAV, obtain the target area training set, and build a large model of the environment training set. The first detection module is used to input the three-dimensional image marker elements into the environmental training set large model to perform the first geological disaster detection process and obtain the first geological disaster detection result. The disaster prediction module is used to predict geological disasters based on the large model of the environmental training set, and to preprocess the geological disaster prediction results by combining them with historical images of the target area to obtain UAV geological disaster prediction images. The second detection module is used to convert the UAV geological disaster prediction image into a UAV index feature image, and to perform big data analysis on the UAV index feature image and the UAV initial image to obtain the second geological disaster detection result. The data integration module is used to compare and verify the first geological disaster detection result and the second geological disaster detection result to obtain the UAV image geological disaster detection result.

[0036] The beneficial effects of this invention are as follows: This invention establishes a three-dimensional image coordinate system based on initial UAV images and marks relevant elements. The movement trajectory of elements is recorded through element marking. A large environmental training set model is established based on historical images. The movement traces of different elements are monitored based on the coordinate system, which can improve the detection effect of geological disasters. The geological disaster level is detected by judging the number of elements and the element offset before and after the geological disaster. This is better than the existing method of judging the vibration intensity and visual disaster conditions of geological disasters. It is more systematic and can help rescuers carry out corresponding rescue operations under different disaster conditions. Then, geological disaster detection is carried out by multi-data fusion based on big data simulation and repeated verification of disaster conditions, making the geological disaster detection results more accurate.

[0037] The above descriptions are merely embodiments of the present invention. Commonly known technical solutions or characteristics are not described in detail here. It should be noted that those skilled in the art can make various modifications and improvements without departing from the present invention, and these should also be considered within the scope of protection of the present invention. These modifications and improvements will not affect the effectiveness of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A method for detecting geological disasters by unmanned aerial vehicle (UAV) multi-data fusion, characterized in that, The method comprises: S1: obtaining the actual situation of the target area by the unmanned aerial vehicle, and collecting the initial image of the unmanned aerial vehicle; S2: establishing a three-dimensional image coordinate system according to the initial image of the unmanned aerial vehicle, marking the elements in the initial image of the unmanned aerial vehicle, and obtaining a three-dimensional image marked element; S3: performing clustering analysis on the historical image of the target area obtained by the unmanned aerial vehicle, obtaining a training set of the target area, and establishing an environment training set large model; S4: inputting the three-dimensional image marked element into the environment training set large model to perform a first geological disaster detection process, and obtaining a first geological disaster detection result; S5: performing geological disaster prediction according to the environment training set large model, preprocessing the geological disaster prediction result in combination with the historical image of the target area, and obtaining an unmanned aerial vehicle geological disaster prediction image; S6: converting the unmanned aerial vehicle geological disaster prediction image into an unmanned aerial vehicle index feature image, performing big data analysis on the unmanned aerial vehicle index feature image and the initial image of the unmanned aerial vehicle, and obtaining a second geological disaster detection result; S7: comparing and verifying the first geological disaster detection result and the second geological disaster detection result, and obtaining an unmanned aerial vehicle image geological disaster detection result. 2.The unmanned aerial vehicle multi-data fusion geological disaster detection method of claim 1, wherein, The step of establishing a three-dimensional image coordinate system according to the initial image of the unmanned aerial vehicle, marking the elements in the initial image of the unmanned aerial vehicle, and obtaining a three-dimensional image marked element comprises: establishing a three-dimensional image coordinate system with the lowest central position of the target area in the initial image of the unmanned aerial vehicle as the origin; traversing the pixel values of all pixel points in the initial image of the unmanned aerial vehicle, and determining a to-be-marked region according to the pixel values; obtaining a first minimum value and a first maximum value of the pixel points in the to-be-marked region on the X-axis, and obtaining a second minimum value and a second maximum value of the pixel points in the to-be-marked region on the Y-axis; obtaining four vertex coordinate information according to the first minimum value, the first maximum value, the second minimum value, and the second maximum value, and obtaining a rectangular marking box corresponding to the to-be-marked region according to the four vertex coordinate information; marking fixed elements in the rectangular marking box, recording the position and size of the fixed elements, and mapping the fixed elements to the Z-axis of the three-dimensional image coordinate system to obtain the three-dimensional image marked element.

3. The unmanned aerial vehicle multi-data fusion geological disaster detection method according to claim 1, characterized in that, The step of performing clustering analysis on the historical image of the target area obtained by the unmanned aerial vehicle, obtaining a training set of the target area, and establishing an environment training set large model comprises: performing gray scale stretching on the historical image of the target area; constructing an H-S color histogram and a gradient direction histogram for the reference power transmission line image after gray scale stretching; obtaining corresponding probability distributions according to the H-S color histogram and the gradient direction histogram; selecting an initial cluster center point of the historical image of the target area after gray scale stretching; determining the number of initial cluster center points; performing JSD calculation on each reference power transmission line image, obtaining a training set of the target area according to the probability distributions, and obtaining an environment training set large model. 4.The unmanned aerial vehicle multi-data fusion geological disaster detection method of claim 1, wherein, The step of inputting the three-dimensional image mark element into the environment training set large model for the first geological disaster detection process to obtain a first geological disaster detection result comprises: The three-dimensional image mark element is overlapped with the historical image of the target area according to the three-dimensional image coordinate system origin. The three-dimensional image mark element and the historical corresponding position of the three-dimensional image mark element in the historical image of the target area are element quantity judged. If the deviation number of the three-dimensional image mark element and the historical corresponding position is > 1 / 2, the geological disaster severity is moderate disaster condition. If the deviation number of the three-dimensional image mark element and the historical corresponding position is <= 1 / 2, the geological disaster severity is mild disaster condition.

5. The unmanned aerial vehicle multi-data fusion geological disaster detection method according to claim 4, characterized in that: After the three-dimensional image mark element and the historical corresponding position of the three-dimensional image mark element in the historical image of the target area are element quantity judged, the following steps are further included: The three-dimensional image mark element connection line is connected with each three-dimensional image mark element in the three-dimensional image coordinate system with the three-dimensional image coordinate system origin to obtain a three-dimensional image mark element connection line. The average deviation angle of the three-dimensional image mark element connection line is calculated, and the average deviation angle of the three-dimensional image mark element connection line is calculated. According to the average deviation angle of the three-dimensional image mark element connection line, the three-dimensional image mark element deviation line is drawn from the three-dimensional image coordinate system origin. The historical three-dimensional image mark element deviation line is calculated in the same way as the three-dimensional image mark element deviation line. The angle difference between the three-dimensional image mark element deviation line and the historical three-dimensional image mark element deviation line is judged. If an angle difference between the three-dimensional image marker element offset line and the three-dimensional image marker element offset line in the historical image of the target region is < 30 o then it is a moderate disaster condition; If the angle difference between the three-dimensional image marker element offset line and the three-dimensional image marker element offset line in the historical image of the target region is >= 30 o , it is a severe disaster condition.

6. The unmanned aerial vehicle multi-data fusion geological disaster detection method according to claim 1, characterized in that, The step of combining the geological disaster prediction result with the historical image of the target area for preprocessing to obtain the unmanned aerial vehicle geological disaster prediction image comprises: The historical image of the target area is parsed into continuous image frames in time sequence, and each image frame is converted into a gray image. Based on the optical flow velocity method, the motion speed and direction of the pixel points between each two image frames are calculated to obtain a geological disaster prediction motion map of the target area. The geological disaster prediction motion map is combined with the historical image of the target area to simulate the geological disaster target area change process. The geological disaster target area change process is converted into an image to obtain the unmanned aerial vehicle geological disaster prediction image.

7. The unmanned aerial vehicle multi-data fusion geological disaster detection method according to claim 1, characterized in that, The step of converting the unmanned aerial vehicle geological disaster prediction image into an unmanned aerial vehicle index feature image comprises: The unmanned aerial vehicle geological disaster prediction image is analyzed frame by frame, and different elements in each frame are marked according to different mark signals to obtain an element mark prediction image. The element mark prediction image is subjected to element motion simulation according to the unmanned aerial vehicle geological disaster prediction image, and the motion trajectory of the same element in the element mark prediction image is recorded. The motion trajectory of the same element in the element mark prediction image is recorded and converted into index feature data to obtain an unmanned aerial vehicle index feature image. 8.The unmanned aerial vehicle multi-data fusion geological disaster detection method of claim 1, wherein, The step of performing big data analysis on the unmanned aerial vehicle index characteristic image and the unmanned aerial vehicle initial image to obtain a second geological disaster detection result comprises: Marking different elements in the unmanned aerial vehicle initial image, and performing geological disaster simulation according to a big data model; Recording the motion trajectory of different elements in each geological disaster simulation process to obtain an element simulation trajectory; Judging the coincidence degree of the element simulation trajectory and the unmanned aerial vehicle index characteristic image; If the coincidence degree of the element simulation trajectory and the unmanned aerial vehicle index characteristic image is greater than 70%, it is a severe disaster condition; If the coincidence degree of the element simulation trajectory and the unmanned aerial vehicle index characteristic image is greater than 40% and less than or equal to 70%, it is a moderate disaster condition; If the coincidence degree of the element simulation trajectory and the unmanned aerial vehicle index characteristic image is less than or equal to 40%, it is a mild disaster condition. 9.The unmanned aerial vehicle multi-data fusion geological disaster detection method of claim 1, wherein, The step of comparing and verifying the first geological disaster detection result and the second geological disaster detection result to obtain an unmanned aerial vehicle image geological disaster detection result comprises: Comparing the first geological disaster detection result and the second geological disaster detection result, if the comparison difference between the first geological disaster detection result and the second geological disaster detection result is less than 20%, comparing and optimizing the first geological disaster detection result and the second geological disaster detection result according to historical geological disaster results to obtain an unmanned aerial vehicle image geological disaster detection result; If the comparison difference between the first geological disaster detection result and the second geological disaster detection result is greater than 20%, inputting the first geological disaster detection result and the second geological disaster detection result into a big data platform to perform geological disaster data simulation optimization to obtain an unmanned aerial vehicle image geological disaster detection result.

10. A geological disaster detection system using unmanned aerial vehicle multi-data fusion, characterized in that, The system comprises: An image acquisition module for acquiring the actual condition of a target area by an unmanned aerial vehicle to obtain an unmanned aerial vehicle initial image; An element marking module for establishing a three-dimensional image coordinate system according to the unmanned aerial vehicle initial image, marking elements in the unmanned aerial vehicle initial image to obtain three-dimensional image marked elements; A big model training module for performing cluster analysis on historical images of the target area acquired by the unmanned aerial vehicle to obtain a target area training set, and establishing an environment training set big model; A first detection module for inputting the three-dimensional image marked elements into the environment training set big model to perform a first geological disaster detection process to obtain a first geological disaster detection result; A disaster prediction module for performing geological disaster prediction according to the environment training set big model, and preprocessing the geological disaster prediction result combined with the historical images of the target area to obtain an unmanned aerial vehicle geological disaster prediction image; A second detection module for converting the unmanned aerial vehicle geological disaster prediction image into an unmanned aerial vehicle index characteristic image, and performing big data analysis on the unmanned aerial vehicle index characteristic image and the unmanned aerial vehicle initial image to obtain a second geological disaster detection result; A data integration module for comparing and verifying the first geological disaster detection result and the second geological disaster detection result to obtain an unmanned aerial vehicle image geological disaster detection result.

Images