Data security analysis system and method applied to low-altitude flight

Abstract

The invention discloses a data safety analysis system and method applied to low-altitude flight, and relates to the technical field of data safety analysis, and the method comprises the steps: obtaining a depth image of a low-altitude aircraft in a flight process, and carrying out the judgment and extraction of feature points in the depth image; obtaining a positioning path of the low-altitude aircraft in the flight process, and obtaining a normal positioning point and an abnormal positioning point in the positioning path; obtaining a feature distance between each positioning point and each feature point, and obtaining an incidence relation expression between the feature distance and a pixel value according to the pixel value of the feature point in the depth image; and correcting each abnormal positioning point in the positioning path according to the incidence relation expression and the geographic position coordinate of each feature point. Visual positioning assistance is carried out by using the depth image to determine the position of the unmanned aerial vehicle, correction of the positioning point is realized, errors generated during positioning of the unmanned aerial vehicle are reduced, the positioning precision of the unmanned aerial vehicle is improved, and the task execution efficiency is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of data security analysis, and in particular to a data security analysis system and method applied to low-altitude flight. Background Art

[0002] At present, with the rapid development of low-altitude aircraft such as drones, they are playing an increasingly important role in many fields. With the advantages of flexibility, convenience and efficient operation, drones have been widely used in many fields such as aerial photography, agricultural plant protection and power inspection. However, they also bring a series of challenges: in satellite navigation and positioning systems such as GPS, the satellite signals received by the receiver are reflected by reflectors such as buildings, water surfaces and the ground, and then superimposed with the signals directly from the satellite, so that the receiver receives signals from multiple different paths, resulting in positioning deviation; in the field of power inspection, positioning deviation may prevent drones from accurately reaching the location of power equipment that needs to be inspected, miss some potential fault points, and increase the safety risks of the power system; in the field of agricultural plant protection, positioning deviation may cause uneven pesticide spraying, over-spraying in some areas and under-spraying in other areas, affecting the growth and prevention and control effects of crops. The positioning error caused by inaccurate data will cause a large deviation between the actual movement path of the drone and the positioning path, affecting subsequent analysis and decision-making. Therefore, solving the above problems, improving the positioning accuracy of UAVs, increasing flight stability, and ensuring mission execution efficiency are of great significance to the development of UAVs and other low-altitude aircraft. Summary of the invention

[0003] The purpose of the present invention is to provide a data safety analysis system and method for low-altitude flight to solve the problems raised in the prior art.

[0004] To achieve the above object, the present invention provides the following technical solutions: A data safety analysis method applied to low-altitude flight comprises the following steps: Step S100: obtaining a depth image of the low-altitude aircraft during flight, establishing a three-dimensional model of the flight area, and extracting all target points in the three-dimensional model; judging and extracting feature points therein according to the coordinate information of each target point; Step S200: obtaining the positioning path of the low-altitude aircraft during flight, obtaining normal positioning points and positioning points to be detected, and classifying each positioning point to be detected as a normal positioning point or an abnormal positioning point according to the coordinate information of each positioning point to be detected; Step S300: according to each feature point in the depth image and the geographical location coordinates of each positioning point in the positioning path, obtain the feature distance between each positioning point and each corresponding feature point, and according to the pixel value of each feature point in the corresponding depth image, obtain the correlation relationship expression between the feature distance and the pixel value; Step S400: Correcting each abnormal positioning point in the positioning path according to the association relationship expression and the geographical location coordinates of each feature point.

[0005] Furthermore, step S100 includes: Step S110: Arrange a laser radar camera on a low-altitude aircraft to capture a number of consecutive frames of three-dimensional point cloud data and generate a corresponding depth image; establish a three-dimensional model of the flight area, construct all components in the three-dimensional model, and use the components corresponding to the facilities or buildings that are fixed and difficult to change as feature components. Starting from the first frame of the depth image DG1, use the area in the depth image DG1 with a pixel value greater than a preset pixel threshold as the target area in the depth image DG1, and then obtain all the target areas in the depth image DG1; This embodiment uses a TOF laser radar camera to collect high-quality continuous multi-frame three-dimensional point cloud data; a depth image is an image that reflects the distance information from an object in a scene to the camera. In a depth image, the value of each pixel represents the distance from the corresponding point in the scene to the camera. The pixel value of a depth image is generally a grayscale value, and its value is related to the distance. When the object is closer to the camera, the pixel value is larger, and when the object is farther from the camera, the pixel value is smaller. However, a depth image cannot be directly taken like an ordinary color photo, but can only be obtained through a laser radar camera. The three-dimensional model of the flight area can be obtained using BIM modeling; Step S120: Obtain a certain target area in the depth image DG1, and the component areas of all components in the three-dimensional model in the depth image DG1, and use the component area with the largest intersection area with the certain target area as the target component area; if the intersection area between the target component area and the certain target area is greater than the preset area threshold, and the target component corresponding to the target component area is a characteristic component, then obtain the actual height h corresponding to the highest point of the target component and the maximum height H of the low-altitude aircraft during the flight; if Hh>h0, h0 is the height threshold, then use the geographical location coordinates of the target component at the actual height h as the target point of the target component, and the geographical location coordinates include latitude, longitude and altitude; and each characteristic component of the flight area corresponds to at most one target point, and so on, and then mark all target points in the three-dimensional model; The actual height h is the actual height corresponding to the highest point of the target component in the actual environment, and the maximum height H is the maximum height of the low-altitude aircraft during flight in the actual environment; in this embodiment, the height threshold h0 is determined according to the actual situation. When the actual height h is large, the height threshold h0 should be adjusted accordingly. When the actual height h is small, the height threshold h0 should be adjusted accordingly. Step S130: Obtain all depth images where a certain target point appears, and convert each depth image into a two-dimensional coordinate system, wherein the maximum values ​​of the horizontal and vertical coordinates of a two-dimensional coordinate system corresponding to a certain depth image are W and W respectively. ma and H mi , the horizontal and vertical coordinates of a target point in a two-dimensional coordinate system are W t and H t , if A1*W is satisfied ma ≤W t ≤A2*W ma And B1*H mi ≤H t ≤B2*H mi , then a target point in a depth image is marked, wherein A1 and A2 are respectively the first horizontal coordinate coefficient and the second horizontal coordinate coefficient, A1<A2, B1 and B2 are respectively the first vertical coordinate coefficient and the second vertical coordinate coefficient, B1<B2, if the total number of marked target points in the depth image is not less than the set total number threshold, then the target point is taken as a feature point; and then all feature points in the three-dimensional model are obtained.

[0006] Feature points are points that are not at the edge of the image, which means that these points are relatively stable in the image. Points that are not at the edge of the image usually have better anti-interference performance. Points at the edge of the image are easily affected by factors such as noise and illumination changes, resulting in unstable features. Only by analyzing such satisfactory feature points can we obtain more accurate and effective information for the next step of calculation. Further, step S200 includes: Step S210: Obtain all positioning points in the positioning path, as well as the time and positioning coordinates corresponding to each positioning point, wherein the positioning coordinates only include the longitude and latitude of the positioning point; obtain the positioning distances between all two temporally adjacent positioning points according to the positioning coordinates, and if a certain positioning distance is less than a preset distance threshold, mark the certain positioning distance; and use the temporally adjacent and marked positioning distances as a reference distance set, and then obtain all reference distance sets in the positioning path; if the number of elements in a certain reference distance set is greater than a preset number threshold, then use all positioning points corresponding to the reference distance set as normal positioning points, and then obtain all normal positioning points in the positioning path, and use the remaining positioning points as positioning points to be detected, and sort the positioning points to be detected in chronological order; Step S220: The first positioning point to be detected is taken as positioning point P1, the time corresponding to positioning point P1 is T1, the previous time of time T1 is T0, the positioning point corresponding to time T0 is P0, the previous time of time T0 is T -1 , time T -1 The corresponding positioning point is P -1 , and P1=(x1,y1), P0=(x0,y0), P -1 =(x -1 ,y -1 ), where x1 and y1 are the latitude and longitude of the positioning point P1, x0 and y0 are the latitude and longitude of the positioning point P0, and x -1 ,y -1 They are the positioning points P -1 Corresponding longitude and latitude; based on the positioning point P -1 The position of the positioning point P0 is the starting point, the direction of the positioning point P0 is the end point as V1, and the opposite direction of the direction V1 is V2; the positioning point deviation of the positioning point P0 is AD0=[(x0-x -1 ) / m0] 2 +[(y0-y -1 ) / n0] 2 , where m0 is the distance m meters forward along the direction V1 from the positioning point P0 as the starting point, n0 is the distance n meters forward along the direction V2 from the positioning point P0 as the starting point, m>n>0; if the positioning point P1 is within the range corresponding to the positioning point deviation AD0, the positioning point P1 is regarded as a normal positioning point, if the positioning point P1 is not within the positioning point deviation, the positioning point P1 is regarded as an abnormal positioning point; It should be noted that when positioning deviation occurs, the longitude and latitude are the first to be affected, and the height is not affected much. Here, the positioning point is calculated only using longitude and latitude. First, it is easy to calculate, and second, the height does not change much. It is not meaningful to analyze the height. Only the longitude and latitude that are mainly affected can be analyzed. If the low-altitude aircraft is in normal flight, then the range of its position at the next moment may be an ellipse. Because if the low-altitude aircraft suddenly accelerates, it will appear at a farther position in front at the next moment. If it suddenly stops or turns around, it will not immediately appear at a farther position in the back at the next moment, but can only appear at a closer position in the back. The calculation formula for the positioning point deviation is AD0=[(x0-x -1 ) / m0] 2 +[(y0-y -1 ) / n0] 2 is the calculation formula of the ellipse, so m>n>0; Step S230: Take the next positioning point of positioning point P1 as P2. If positioning point P2 is a positioning point to be detected and positioning point P1 is a normal positioning point, determine whether positioning point P2 is a normal or abnormal positioning point according to positioning point P0 and positioning point P1 according to steps S210 to S220; if positioning point P2 is a positioning point to be detected and positioning point P1 is an abnormal positioning point, obtain positioning point P2. -1 The distance L-1 0 between the positioning point P0 and the positioning point P0 is taken as the starting point, the direction is V1, and the length is the end point of the distance L-1 0 as the predetermined position P^1. According to the positioning point P0 and the predetermined position P^1, the positioning point P2 is judged to be a normal or abnormal positioning point; and so on, until all normal positioning points and abnormal positioning points in the positioning path are obtained.

[0007] The positioning device continuously receives satellite signals and updates the location information in real time based on the new signal data. Over a period of time, the device will perform positioning calculations at a certain time interval, such as once per second, to obtain a series of continuous positioning points. These positioning points reflect the position changes of objects at different times. All positioning points are collected and the positioning points that are adjacent in time are connected to obtain the positioning path within a period of time, and the path with little deviation is regarded as the normal path.

[0008] Furthermore, step S300 includes: Step S310: Obtain the time corresponding to each depth image. The time corresponding to a depth image is M, and the positioning point corresponding to the low-altitude aircraft at time M is P. M , if the positioning point P M is the normal positioning point, and the positioning point P is obtained. M The geographic location coordinates include latitude, longitude and altitude; obtain a feature point P in a depth image t The geographic location coordinates of the positioning point P M and a feature point P t The characteristic distance FDM t is obtained, and then the characteristic distance between each positioning point and each corresponding characteristic point is obtained; Step S320: According to the feature distance between each positioning point and each corresponding feature point, and the pixel value of each feature point in the corresponding depth image, the least squares method is used to obtain the correlation expression between the feature distance and the pixel value FD=k*GV+b, where FD is the feature distance, k is the slope, GV is the pixel value, and b is the intercept.

[0009] Since the value of each pixel in the depth image represents the distance from the corresponding point in the scene to the camera, the pixel value of the depth image is generally a grayscale value, and its value is related to the distance. When the object is closer to the camera, the pixel value is larger, and when the object is farther from the camera, the pixel value is smaller. In this solution, a linear change is used to obtain the correlation relationship. Obtaining the correlation relationship expression by the least squares method is a prior art and will not be repeated here.

[0010] Furthermore, step S400 includes: Step S410: Obtain the time corresponding to each abnormal positioning point in the positioning path, where the time corresponding to a certain abnormal positioning point is M a , get the time M a The corresponding depth image DG a , get the depth image DG a The pixel values ​​of all feature points in , and according to the association relationship expression, the pixel values ​​of each feature point at time M are obtained. a The corresponding characteristic distance when Step S420: Take the location of each feature point as the sphere center and the feature distance as the radius to obtain the spherical surface corresponding to each feature point; obtain the intersection points corresponding to any three of the spherical surfaces, and obtain two intersection points in total. Take the intersection point closest to a certain abnormal positioning point as the feature positioning point, and then obtain several feature positioning points; and calculate the average value of the geographical location coordinates of the feature positioning points to obtain a virtual positioning point as the corrected result of a certain abnormal positioning point.

[0011] According to the pixel value of each feature point corresponding to each moment, the feature distance of the feature value from the positioning point is obtained, and multiple feature distances are obtained, and each feature distance corresponds to a feature point. For example, if there is a triangular cone, the three vertices (feature points) at the bottom of the triangular cone are known to be determined, and the distances from the three vertices at the bottom to the vertex (positioning point) are known, then the position of this vertex (positioning point) can be determined; because in three-dimensional space, determining the position of a point requires three non-collinear distance information. When the three vertices at the bottom are determined, they form a plane triangle, and the distances from these three vertices to the vertices are like three spheres with the three known points as the center of the sphere and the corresponding distances as the radius. The intersection of these three spheres is the vertex of the triangular cone. This intersection is unique, so the vertex position can be determined.

[0012] A data safety analysis system applied to low-altitude flight, comprising a feature point judgment and extraction module, a positioning point judgment module, an association relationship expression calculation module and an abnormal positioning point correction module; Feature point judgment and extraction module: used to obtain the depth image of the low-altitude aircraft during flight, establish a three-dimensional model of the flight area, and extract all target points in the three-dimensional model; according to the coordinate information of each target point, the feature points are judged and extracted; Positioning point judgment module: used to obtain the positioning path of the low-altitude aircraft during flight, obtain the normal positioning points and the positioning points to be detected, and classify each positioning point to be detected as a normal positioning point or an abnormal positioning point according to the coordinate information of each positioning point to be detected; The correlation expression calculation module is used to obtain the characteristic distance between each positioning point and each corresponding characteristic point according to each characteristic point in the depth image and the geographical location coordinates of each positioning point in the positioning path, and obtain the correlation expression between the characteristic distance and the pixel value according to the pixel value of each characteristic point in the corresponding depth image; Abnormal positioning point correction module: used to correct the abnormal positioning points in the positioning path according to the association relationship expression and the geographical location coordinates of each feature point.

[0013] Further, the positioning point judgment module includes a positioning point judgment unit to be detected, a positioning point deviation calculation unit and an abnormal positioning point judgment unit; The unit for determining the positioning points to be detected is used to obtain all the positioning points in the positioning path, as well as the time and positioning coordinates corresponding to each positioning point, wherein the positioning coordinates only include the longitude and latitude of the positioning point; obtain the positioning distances between all two positioning points adjacent to each other at all times, and obtain all the reference distance sets in the positioning path; obtain all the normal positioning points in the positioning path, and use the remaining positioning points as the positioning points to be detected; Positioning point deviation calculation unit: used to obtain the positioning point deviation of the positioning point according to the coordinate information of each positioning point to be detected; Abnormal positioning point judgment unit: used to judge whether the positioning point is an abnormal positioning point, and whether the predetermined position is obtained, and then obtain all normal positioning points and abnormal positioning points in the positioning path.

[0014] Further, the abnormal positioning point correction module includes a feature distance calculation unit and an abnormal positioning point correction unit; Feature distance calculation unit: used to obtain the time corresponding to each abnormal positioning point in the positioning path, and obtain the feature distance corresponding to each feature point at each time according to the association relationship expression; Abnormal positioning point correction unit: used to obtain a number of characteristic positioning points, and calculate the average value of the geographical location coordinates of the characteristic positioning points to obtain a virtual positioning point as the result of correction of a certain abnormal positioning point.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention provides a data security analysis system and method for low-altitude flight, including: obtaining a depth image of a low-altitude aircraft during flight, and judging and extracting feature points therein; obtaining a positioning path of the low-altitude aircraft during flight, and obtaining normal positioning points and abnormal positioning points therein; obtaining a characteristic distance between each positioning point and each feature point, and obtaining an expression of the correlation relationship between the characteristic distance and the pixel value according to the pixel value of the feature point in the depth image; and correcting each abnormal positioning point in the positioning path according to the expression of the correlation relationship and the geographical location coordinates of each feature point. The present invention uses a depth image for visual positioning assistance, obtains an expression of the correlation relationship between the characteristic distance and the pixel value, determines the position of the drone, corrects the positioning point, reduces the error generated by the drone during positioning, improves the positioning accuracy of the drone, and ensures the efficiency of task execution, which is of great significance to the development of low-altitude aircraft such as drones. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic diagram of a flow chart of a data safety analysis method applied to low-altitude flight according to the present invention; Figure 2 The present invention is a structural diagram of a data safety analysis system applied to low-altitude flight. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] Example: Figure 1 As shown, the present invention provides a technical solution for a data safety analysis method applied to low-altitude flight, comprising the following steps: Step S100: Acquire a depth image of the low-altitude aircraft during flight, establish a three-dimensional model of the flight area, and extract all target points in the three-dimensional model; based on the coordinate information of each target point, determine and extract the feature points therein.

[0019] Step S110: Arrange a laser radar camera on a low-altitude aircraft, capture a number of consecutive frames of three-dimensional point cloud data, and generate a corresponding depth image; establish a three-dimensional model of the flight area, and construct all components in the three-dimensional model, and use the components corresponding to the facilities or buildings that are fixed and difficult to change as feature components, starting from the first frame of the depth image DG1, and use the area in the depth image DG1 with a pixel value greater than a preset pixel threshold as the target area in the depth image DG1, and then obtain all the target areas in the depth image DG1; This embodiment uses a TOF laser radar camera to collect high-quality continuous multi-frame three-dimensional point cloud data; a depth image is an image that reflects the distance information from an object in a scene to a camera. In a depth image, the value of each pixel represents the distance from the corresponding point in the scene to the camera. The pixel value of a depth image is generally a grayscale value, and its value is related to the distance. When the object is closer to the camera, the pixel value is larger, and when the object is farther from the camera, the pixel value is smaller. However, a depth image cannot be directly taken like an ordinary color photo, but can only be obtained through a laser radar camera. The three-dimensional model of the flight area can be obtained by BIM modeling.

[0020] Step S120: Obtain a certain target area in the depth image DG1, and the component areas of all components in the three-dimensional model in the depth image DG1, and use the component area with the largest intersection area with the certain target area as the target component area; if the intersection area between the target component area and the certain target area is greater than the preset area threshold, and the target component corresponding to the target component area is a characteristic component, then obtain the actual height h corresponding to the highest point of the target component and the maximum height H of the low-altitude aircraft during the flight; if Hh>h0, h0 is the height threshold, then use the geographical location coordinates of the target component at the actual height h as the target point of the target component, and the geographical location coordinates include latitude, longitude and altitude; and each characteristic component of the flight area corresponds to at most one target point, and so on, and then mark all target points in the three-dimensional model; The actual height h is the actual height corresponding to the highest point of the target component in the actual environment, and the maximum height H is the maximum height of the low-altitude aircraft during flight in the actual environment; in this embodiment, the height threshold h0 is determined according to the actual situation. When the actual height h is large, the height threshold h0 should be adjusted accordingly. When the actual height h is small, the height threshold h0 should be adjusted accordingly.

[0021] Step S130: Obtain all depth images where a certain target point appears, and convert each depth image into a two-dimensional coordinate system, wherein the maximum values ​​of the horizontal and vertical coordinates of a two-dimensional coordinate system corresponding to a certain depth image are W and W respectively. ma and H mi , the horizontal and vertical coordinates of a target point in a two-dimensional coordinate system are W tand H t , if A1*W is satisfied ma ≤W t ≤A2*W ma And B1*H mi ≤H t ≤B2*H mi , then a target point in a depth image is marked, wherein A1 and A2 are respectively the first horizontal coordinate coefficient and the second horizontal coordinate coefficient, A1<A2, B1 and B2 are respectively the first vertical coordinate coefficient and the second vertical coordinate coefficient, B1<B2, if the total number of marked target points in the depth image is not less than the set total number threshold, then the target point is taken as a feature point; and then all feature points in the three-dimensional model are obtained.

[0022] In this embodiment, the values ​​of the first abscissa coefficient and the second abscissa coefficient A1 and A2 are 0.1 and 0.9 respectively, and the values ​​of the first ordinate coefficient and the second ordinate coefficient B1 and B2 are 0.1 and 0.9 respectively. The total number threshold should be determined according to the number of target points appearing in the depth image. When the target point appears in 10 depth images, the total number threshold is 8, that is, the total number of target points marked in the depth image is not less than 8, and the target point can be used as a feature point. Therefore, in this embodiment, the total number threshold is the number of points appearing in the depth image multiplied by 0.8.

[0023] Feature points are points that are not at the edge of the image, which means that these points are relatively stable in the image. Points that are not at the edge of the image usually have better anti-interference properties. Points at the edge of the image are easily affected by factors such as noise and lighting changes, resulting in unstable features. Only by analyzing such satisfactory feature points can we obtain more accurate and effective information for the next step of calculation.

[0024] Step S200: Acquire the positioning path of the low-altitude aircraft during flight, obtain normal positioning points and positioning points to be detected, and classify each positioning point to be detected as a normal positioning point or an abnormal positioning point according to the coordinate information of each positioning point to be detected.

[0025] Step S210: Obtain all positioning points in the positioning path, as well as the time and positioning coordinates corresponding to each positioning point, wherein the positioning coordinates only include the longitude and latitude of the positioning point; obtain the positioning distances between all two temporally adjacent positioning points according to the positioning coordinates, and if a certain positioning distance is less than a preset distance threshold, mark the certain positioning distance; and use the temporally adjacent and marked positioning distances as a reference distance set, and then obtain all reference distance sets in the positioning path; if the number of elements in a certain reference distance set is greater than a preset number threshold, then use all positioning points corresponding to the reference distance set as normal positioning points, and then obtain all normal positioning points in the positioning path, and use the remaining positioning points as positioning points to be detected, and sort the positioning points to be detected in chronological order; Step S220: The first positioning point to be detected is taken as positioning point P1, the time corresponding to positioning point P1 is T1, the previous time of time T1 is T0, the positioning point corresponding to time T0 is P0, the previous time of time T0 is T -1 , time T -1 The corresponding positioning point is P -1 , and P1=(x1,y1), P0=(x0,y0), P -1 =(x -1 ,y -1 ), where x1 and y1 are the latitude and longitude of the positioning point P1, x0 and y0 are the latitude and longitude of the positioning point P0, and x -1 ,y -1 They are the positioning points P -1 Corresponding longitude and latitude; based on the positioning point P -1 The position of the positioning point P0 is the starting point, the direction of the positioning point P0 is the end point as V1, and the opposite direction of the direction V1 is V2; the positioning point deviation of the positioning point P0 is AD0=[(x0-x -1 ) / m0] 2 +[(y0-y -1 ) / n0] 2 , where m0 is the distance m meters forward along the direction V1 from the positioning point P0 as the starting point, n0 is the distance n meters forward along the direction V2 from the positioning point P0 as the starting point, m>n>0; if the positioning point P1 is within the range corresponding to the positioning point deviation AD0, the positioning point P1 is regarded as a normal positioning point, if the positioning point P1 is not within the positioning point deviation, the positioning point P1 is regarded as an abnormal positioning point; It should be noted that when positioning deviation occurs, the longitude and latitude are the first to be affected, and the height is not affected much. Here, the positioning point is calculated only using longitude and latitude. First, it is easy to calculate, and second, the height does not change much. It is not meaningful to analyze the height. Only the longitude and latitude that are mainly affected can be analyzed. If the low-altitude aircraft is in normal flight, then the range of its position at the next moment may be an ellipse. Because if the low-altitude aircraft suddenly accelerates, it will appear at a farther position in front at the next moment. If it suddenly stops or turns around, it will not immediately appear at a farther position in the back at the next moment, but can only appear at a closer position in the back. The calculation formula for the positioning point deviation is AD0=[(x0-x -1 ) / m0] 2 +[(y0-y -1 ) / n0] 2 is the calculation formula of the ellipse, so m>n>0. In this embodiment, m=3, n=2, and the unit is meter.

[0026] Step S230: Take the next positioning point of positioning point P1 as P2. If positioning point P2 is a positioning point to be detected and positioning point P1 is a normal positioning point, determine whether positioning point P2 is a normal or abnormal positioning point according to positioning point P0 and positioning point P1 according to steps S210 to S220; if positioning point P2 is a positioning point to be detected and positioning point P1 is an abnormal positioning point, obtain positioning point P2. -1 The distance L-1 0 between the positioning point P0 and the positioning point P0 is taken as the starting point, the direction is V1, and the length is the end point of the distance L-1 0 as the predetermined position P^1. According to the positioning point P0 and the predetermined position P^1, the positioning point P2 is judged to be a normal or abnormal positioning point; and so on, until all normal positioning points and abnormal positioning points in the positioning path are obtained.

[0027] The positioning device continuously receives satellite signals and updates the location information in real time based on the new signal data. Over a period of time, the device will perform positioning calculations at a certain time interval, such as once per second, to obtain a series of continuous positioning points. These positioning points reflect the position changes of objects at different times. All positioning points are collected and the positioning points that are adjacent in time are connected to obtain the positioning path within a period of time, and the path with little deviation is regarded as the normal path.

[0028] Step S300: According to each feature point in the depth image and the geographical location coordinates of each positioning point in the positioning path, the feature distance between each positioning point and each corresponding feature point is obtained, and according to the pixel value of each feature point in the corresponding depth image, the correlation expression between the feature distance and the pixel value is obtained.

[0029] Step S310: Obtain the time corresponding to each depth image. The time corresponding to a depth image is M, and the positioning point corresponding to the low-altitude aircraft at time M is P. M , if the positioning point P M is the normal positioning point, and the positioning point P is obtained. M The geographic location coordinates include latitude, longitude and altitude; obtain a feature point P in a depth image t The geographic location coordinates of the positioning point P M and a feature point P t The characteristic distance FDM t is obtained, and then the characteristic distance between each positioning point and each corresponding characteristic point is obtained; Step S320: According to the feature distance between each positioning point and each corresponding feature point, and the pixel value of each feature point in the corresponding depth image, the least squares method is used to obtain the correlation expression between the feature distance and the pixel value FD=k*GV+b, where FD is the feature distance, k is the slope, GV is the pixel value, and b is the intercept.

[0030] Since the value of each pixel in the depth image represents the distance from the corresponding point in the scene to the camera, the pixel value of the depth image is generally a grayscale value, and its value is related to the distance. When the object is closer to the camera, the pixel value is larger, and when the object is farther from the camera, the pixel value is smaller. In this solution, a linear change is used to obtain the correlation relationship. Obtaining the correlation relationship expression by the least squares method is a prior art and will not be repeated here.

[0031] Step S400: Correcting each abnormal positioning point in the positioning path according to the association relationship expression and the geographical location coordinates of each feature point.

[0032] Step S410: Obtain the time corresponding to each abnormal positioning point in the positioning path, where the time corresponding to a certain abnormal positioning point is M a , get the time M a The corresponding depth image DG a , get the depth image DG a The pixel values ​​of all feature points in , and according to the association relationship expression, the pixel values ​​of each feature point at time M are obtained. a The corresponding characteristic distance when Step S420: Take the location of each feature point as the sphere center and the feature distance as the radius to obtain the spherical surface corresponding to each feature point; obtain the intersection points corresponding to any three of the spherical surfaces, and obtain two intersection points in total. Take the intersection point closest to a certain abnormal positioning point as the feature positioning point, and then obtain several feature positioning points; and calculate the average value of the geographical location coordinates of the feature positioning points to obtain a virtual positioning point as the corrected result of a certain abnormal positioning point.

[0033] According to the pixel value of each feature point corresponding to each moment, the feature distance of the feature value from the positioning point is obtained, and multiple feature distances are obtained, and each feature distance corresponds to a feature point. For example, if there is a triangular cone, the three vertices (feature points) at the bottom of the triangular cone are known to be determined, and the distances from the three vertices at the bottom to the vertex (positioning point) are known, then the position of this vertex (positioning point) can be determined; because in three-dimensional space, determining the position of a point requires three non-collinear distance information. When the three vertices at the bottom are determined, they form a plane triangle, and the distances from these three vertices to the vertices are like three spheres with the three known points as the center of the sphere and the corresponding distances as the radius. The intersection of these three spheres is the vertex of the triangular cone. This intersection is unique, so the vertex position can be determined.

[0034] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A data safety analysis method applied to low-altitude flight, characterized in that: The following steps are involved: Step S100: obtaining a depth image of the low-altitude aircraft during flight, establishing a three-dimensional model of the flight area, and extracting all target points in the three-dimensional model; judging and extracting feature points therein according to the coordinate information of each target point; Step S200: obtaining the positioning path of the low-altitude aircraft during flight, obtaining normal positioning points and positioning points to be detected, and classifying each positioning point to be detected as a normal positioning point or an abnormal positioning point according to the coordinate information of each positioning point to be detected; Step S300: according to each feature point in the depth image and the geographical location coordinates of each positioning point in the positioning path, obtain the feature distance between each positioning point and each corresponding feature point, and according to the pixel value of each feature point in the corresponding depth image, obtain the correlation relationship expression between the feature distance and the pixel value; Step S400: Correcting each abnormal positioning point in the positioning path according to the association relationship expression and the geographical location coordinates of each feature point.

2. The data safety analysis method for low-altitude flight according to claim 1 is characterized in that: Step S100 includes: Step S110: Arrange a laser radar camera on a low-altitude aircraft to capture a number of consecutive frames of three-dimensional point cloud data and generate a corresponding depth image; establish a three-dimensional model of the flight area, construct all components in the three-dimensional model, and use the components corresponding to the facilities or buildings that are fixed and difficult to change as feature components. Starting from the first frame of the depth image DG1, use the area in the depth image DG1 with a pixel value greater than a preset pixel threshold as the target area in the depth image DG1, and then obtain all the target areas in the depth image DG1; Step S120: obtaining a certain target area in the depth image DG1 and the component areas of all components in the three-dimensional model in the depth image DG1, and taking the component area with the largest intersection area with the certain target area as the target component area; if the intersection area between the target component area and the certain target area is greater than the preset area threshold, and the target component corresponding to the target component area is a characteristic component, then obtaining the actual height h corresponding to the highest point of the target component and the maximum height H of the low-altitude aircraft during the flight, if Hh>h0, h0 is the height threshold, then taking the geographical location coordinates of the target component at the actual height h as the target point of the target component, the geographical location coordinates include latitude, longitude and altitude; and each characteristic component of the flight area corresponds to at most one target point, and so on, and then marking all target points in the three-dimensional model; Step S130: Obtain all depth images where a certain target point appears, and convert each depth image into a two-dimensional coordinate system, wherein the maximum values ​​of the horizontal and vertical coordinates of a two-dimensional coordinate system corresponding to a certain depth image are W and W respectively. ma and H mi , the horizontal and vertical coordinates of the target point in a two-dimensional coordinate system are W t and H t , if A1*W is satisfied ma ≤W t ≤A2*W ma And B1*H mi ≤H t ≤B2*H mi , then a certain target point in the depth image is marked, wherein A1 and A2 are respectively the first horizontal coordinate coefficient and the second horizontal coordinate coefficient, A1<A2, B1 and B2 are respectively the first vertical coordinate coefficient and the second vertical coordinate coefficient, B1<B2, if the total number of marked target points in the depth image is not less than the set total number threshold, then the certain target point is taken as a feature point; and then all feature points in the three-dimensional model are obtained.

3. The data safety analysis method for low-altitude flight according to claim 2 is characterized in that: Step S200 includes: Step S210: Obtain all positioning points in the positioning path, as well as the time and positioning coordinates corresponding to each positioning point, wherein the positioning coordinates only include the longitude and latitude of the positioning point; obtain the positioning distances between all two temporally adjacent positioning points according to the positioning coordinates, and if a certain positioning distance is less than a preset distance threshold, mark the certain positioning distance; and use the temporally adjacent and marked positioning distances as a reference distance set, and then obtain all reference distance sets in the positioning path; if the number of elements in a certain reference distance set is greater than a preset number threshold, then use all positioning points corresponding to the reference distance set as normal positioning points, and then obtain all normal positioning points in the positioning path, and use the remaining positioning points as positioning points to be detected, and sort the positioning points to be detected in chronological order; Step S220: The first positioning point to be detected is taken as positioning point P1, the time corresponding to positioning point P1 is T1, the previous time of time T1 is T0, the positioning point corresponding to time T0 is P0, the previous time of time T0 is T -1 , time T -1 The corresponding positioning point is P -1 , and P1=(x1,y1), P0=(x0,y0), P -1 =(x -1 ,y -1 ), where x1 and y1 are the latitude and longitude of the positioning point P1, x0 and y0 are the latitude and longitude of the positioning point P0, and x -1 ,y -1 They are the positioning points P -1 Corresponding longitude and latitude; based on the positioning point P -1 The position of the positioning point P0 is the starting point, the direction of the positioning point P0 is the end point as V1, and the opposite direction of the direction V1 is V2; the positioning point deviation of the positioning point P0 is AD0=[(x0-x -1 ) / m0] 2 +[(y0-y -1 ) / n0] 2 , where m0 is the distance m meters forward along the direction V1 from the positioning point P0 as the starting point, n0 is the distance n meters forward along the direction V2 from the positioning point P0 as the starting point, m>n>0; if the positioning point P1 is within the range corresponding to the positioning point deviation AD0, the positioning point P1 is regarded as a normal positioning point, if the positioning point P1 is not within the positioning point deviation, the positioning point P1 is regarded as an abnormal positioning point; Step S230: The next positioning point of positioning point P1 is taken as P2. If positioning point P2 is a positioning point to be detected and the positioning point P1 is a normal positioning point, according to positioning point P0 and positioning point P1, according to steps S210 to S220, it is determined whether positioning point P2 is a normal or abnormal positioning point; if positioning point P2 is a positioning point to be detected and the positioning point P1 is an abnormal positioning point, the positioning point P2 is obtained. -1 The distance L-1 0 between the positioning point P0 and the positioning point P0 is taken as the starting point, the direction is V1, and the length is the end point of the distance L-1 0 as the predetermined position P^1. According to the positioning point P0 and the predetermined position P^1, the positioning point P2 is judged to be a normal or abnormal positioning point; and so on, until all normal positioning points and abnormal positioning points in the positioning path are obtained.

4. The data safety analysis method for low-altitude flight according to claim 3 is characterized in that: Step S300 includes: Step S310: Obtain the time corresponding to each depth image. The time corresponding to a depth image is M, and the positioning point corresponding to the low-altitude aircraft at time M is P. M , if the positioning point P M is the normal positioning point, and the positioning point P is obtained. M The geographic location coordinates include latitude, longitude and altitude; obtaining a feature point P in a depth image t The geographic location coordinates of the positioning point P M and a feature point P t The characteristic distance FDM t is obtained, and then the characteristic distance between each positioning point and each corresponding characteristic point is obtained; Step S320: According to the feature distance between each positioning point and each corresponding feature point, and the pixel value of each feature point in the corresponding depth image, the least squares method is used to obtain the correlation expression between the feature distance and the pixel value FD=k*GV+b, where FD is the feature distance, k is the slope, GV is the pixel value, and b is the intercept.

5. The data safety analysis method for low-altitude flight according to claim 4 is characterized in that: Step S400 includes: Step S410: Obtain the time corresponding to each abnormal positioning point in the positioning path, where the time corresponding to a certain abnormal positioning point is M a , get the time M a The corresponding depth image DG a , get the depth image DG a The pixel values ​​of all feature points in , and according to the association relationship expression, the pixel values ​​of each feature point at time M are obtained. a The corresponding characteristic distance when Step S420: taking the location of each feature point as the sphere center and the feature distance as the radius, obtaining the spherical surface corresponding to each feature point; obtaining the intersection points corresponding to any three of the spherical surfaces, obtaining a total of two intersection points, taking the intersection point closest to the abnormal positioning point as the feature positioning point, and then obtaining several feature positioning points; and calculating the average value of the geographical location coordinates of the feature positioning points to obtain a virtual positioning point as the corrected result of the abnormal positioning point.

6. A data safety analysis system, used to execute a data safety analysis method for low-altitude flight as claimed in any one of claims 1 to 5, characterized in that: The system includes a feature point judgment and extraction module, a positioning point judgment module, an association relationship expression calculation module and an abnormal positioning point correction module; Feature point judgment and extraction module: used to obtain the depth image of the low-altitude aircraft during flight, establish a three-dimensional model of the flight area, and extract all target points in the three-dimensional model; according to the coordinate information of each target point, the feature points are judged and extracted; Positioning point judgment module: used to obtain the positioning path of the low-altitude aircraft during flight, obtain the normal positioning points and the positioning points to be detected, and classify each positioning point to be detected as a normal positioning point or an abnormal positioning point according to the coordinate information of each positioning point to be detected; The correlation expression calculation module is used to obtain the characteristic distance between each positioning point and each corresponding characteristic point according to each characteristic point in the depth image and the geographical location coordinates of each positioning point in the positioning path, and obtain the correlation expression between the characteristic distance and the pixel value according to the pixel value of each characteristic point in the corresponding depth image; Abnormal positioning point correction module: used to correct each abnormal positioning point in the positioning path according to the association relationship expression and the geographical location coordinates of each feature point.

7. A data security analysis system according to claim 6, characterized in that: The positioning point judgment module includes a positioning point judgment unit to be detected, a positioning point deviation calculation unit and an abnormal positioning point judgment unit; The unit for determining the positioning points to be detected is used to obtain all the positioning points in the positioning path, as well as the time and positioning coordinates corresponding to each positioning point, wherein the positioning coordinates only include the longitude and latitude of the positioning point; obtain the positioning distances between all two positioning points adjacent to each other at all times, and obtain all the reference distance sets in the positioning path; obtain all the normal positioning points in the positioning path, and use the remaining positioning points as the positioning points to be detected; Positioning point deviation calculation unit: used to obtain the positioning point deviation of the positioning point according to the coordinate information of each positioning point to be detected; Abnormal positioning point judgment unit: used to judge whether the positioning point is an abnormal positioning point, and whether the predetermined position is obtained, and then obtain all normal positioning points and abnormal positioning points in the positioning path.

8. A data security analysis system according to claim 7, characterized in that: The abnormal positioning point correction module includes a characteristic distance calculation unit and an abnormal positioning point correction unit; Feature distance calculation unit: used to obtain the time corresponding to each abnormal positioning point in the positioning path, and obtain the feature distance corresponding to each feature point at each time according to the association relationship expression; Abnormal positioning point correction unit: used to obtain a number of characteristic positioning points, and calculate the average value of the geographical location coordinates of the characteristic positioning points to obtain a virtual positioning point as a result of correction of the abnormal positioning point.

Images