A surveying and mapping method based on low-altitude UAV remote sensing technology

Abstract

The present invention discloses a surveying and mapping method based on low-altitude UAV remote sensing technology, which belongs to the field of surveying and mapping, and includes the following steps: S1, flight path planning; S2, remote sensing image acquisition; S3, remote sensing image processing; S4, remote sensing image feature extraction; S41, building a three-dimensional scanning model, extracting remote sensing image feature points; S42, matching and screening the extracted remote sensing image feature points; S43, obtaining the projection relationship from three-dimensional space to two-dimensional remote sensing image plane based on the camera's projection matrix; S44, using the projection matrix to perform coordinate transformation; S45, calculating the specific position and posture of the target object in three-dimensional space, and completing the positioning of the target scene; S5, drawing a map based on the positioning of the target scene. The surveying and mapping method based on low-altitude UAV remote sensing technology is adopted, and by combining augmented reality with remote sensing mapping, it has the advantages of real-time visualization, precise positioning, high work efficiency, etc., and thus has broad application prospects.

Description

Technical Field

[0001] The present invention relates to the field of surveying and mapping technology, and in particular to a surveying and mapping method based on low-altitude unmanned aerial vehicle remote sensing technology. Background Art

[0002] Drone remote sensing mapping is a technology that uses drones to collect remote sensing data and draw maps. Through the cameras and sensors carried on the drones, high-resolution remote sensing images and detailed data of the surface can be obtained. At this time, if combined with advanced positioning and navigation technologies, drone remote sensing mapping can achieve high-precision surveying and mapping, and is widely used in land surveys, surveying and mapping, urban planning, environmental monitoring and other fields.

[0003] Compared with traditional surveying and mapping methods, UAV remote sensing surveying and mapping has the following significant advantages: 1. Efficient and fast: UAVs can quickly cover large areas, greatly shortening the time for data collection. 2. Low cost: UAVs have low operating and maintenance costs, reducing the cost of overall surveying and mapping projects. 3. High data accuracy: High-resolution sensors and advanced data processing technology ensure high data accuracy and reliability. 4. Strong flexibility: UAVs can easily adapt to various terrains and environmental conditions, improving the flexibility and adaptability of surveying and mapping work.

[0004] At the same time, augmented reality (AR) technology uses optoelectronic display technology, interactive technology, multiple sensor technologies, computer graphics and multimedia technology to integrate the virtual environment generated by the computer with the real environment around the user, so that the user can be convinced from the sensory effect that the virtual environment is a part of the real environment around him. Augmented reality has the characteristics of combining virtual and real, real-time interaction, and three-dimensional registration.

[0005] However, there is no precedent in existing technology for combining augmented reality with remote sensing mapping. Summary of the invention

[0006] The purpose of the present invention is to provide a surveying and mapping method based on low-altitude UAV remote sensing technology to solve the above technical problems.

[0007] To achieve the above object, the present invention provides a surveying and mapping method based on low-altitude UAV remote sensing technology, comprising the following steps:

[0008] S1. Flight path planning: Plan the flight path of the UAV based on graph theory, potential field method or particle swarm optimization algorithm;

[0009] S2, remote sensing image acquisition: after taking off, the UAV flies in the area to be surveyed according to the flight path described in step S1, and collects remote sensing image information during the flight;

[0010] S3, remote sensing image processing: grayscale processing is performed on the collected remote sensing images, and the noise in the remote sensing images is removed by the median filtering method, and then the Canny operator is used for edge detection to obtain a binary remote sensing image containing only the target object;

[0011] S4, remote sensing image feature extraction;

[0012] S41, constructing a three-dimensional scanning model, and extracting feature points of the remote sensing image using the three-dimensional scanning model;

[0013] S42, using the target manager to match and filter the extracted remote sensing image feature points;

[0014] S43, obtaining a projection relationship from the three-dimensional space to the two-dimensional remote sensing image plane based on the projection matrix of the camera;

[0015] S44, using a projection matrix to perform coordinate transformation, mapping feature points in the remote sensing image to actual positions in the three-dimensional space;

[0016] S45, calculating the specific position and posture of the target object in the three-dimensional space according to the converted coordinates, and completing the positioning of the target scene;

[0017] S5. Draw a map based on the positioning of the target scene.

[0018] Preferably, step S3 specifically includes the following steps:

[0019] S31, Grayscale processing:

[0020] S311, calculating the error after pixel displacement:

[0021] (1);

[0022] In the formula, Represents the total error after pixel displacement; represents the weight function; Represents the gray value of the original remote sensing image; Represents the gray value of the remote sensing image after displacement;

[0023] S312. Obtaining the autocorrelation matrix of remote sensing images using second-order derivatives :

[0024] (2);

[0025] In the formula, , Respectively represent the remote sensing images in Direction and Directional gradient;

[0026] S32, Canny operator edge detection;

[0027] S321. Use a Gaussian filter formula to smooth the grayscale remote sensing image, and the Gaussian filter formula is as follows:

[0028] (3);

[0029] In the formula, represents the response of the Gaussian filter; represents the standard deviation of the Gaussian filter;

[0030] S322, Grayscale Remote Sensing Image Perform a first-order difference to obtain the gradient of each pixel:

[0031] (4);

[0032] (5);

[0033] In the formula, , Represents pixel values exist Direction and Directional gradient; , , , Represents grayscale remote sensing images In pixel value , , , The gray value at ;

[0034] S323, based on and Calculate the gradient magnitude and gradient direction:

[0035] (6);

[0036] (7);

[0037] In the formula, represents the gradient magnitude; Indicates the gradient direction;

[0038] S324, perform non-maximum suppression: pixel values ​​in the remote sensing image The neighborhood along To interpolate, When the pixel value is greater than the interpolation point, is regarded as a candidate edge point, and then the candidate edge points are deleted by the double threshold method to obtain the edge of the remote sensing image; at this time, it is assumed that The number of optimized feature points in a frame is ,and , the number of characteristic lines is ,and , then the calculation formula for the number of iterations is as follows:

[0039] (8);

[0040] (9);

[0041] In the formula, Indicates the temporary number of iterations; , Respectively represent the weights of feature points and feature lines; Indicates the maximum number of iterations; represents the denominator; Indicates the number of iterations;

[0042] S325, using a keyframe-based method to perform hybrid feature matching and calculate camera motion parameters, and calculate the number and length of feature lines:

[0043] (10);

[0044] (11);

[0045] In the formula, The number of temporary points representing the characteristic line; Indicates the number of optimized feature points; Indicates The length of the characteristic line; Indicates the length of the characteristic line; and Represent the mass coefficients of characteristic points and characteristic lines respectively; Indicates the maximum number of points of the feature line; Indicates the number of points after the characteristic line is discretized;

[0046] S326, Remote Sensing Image Forward Matching: Assuming the features of the previous frame of remote sensing image The search window is ,use As the matching template, calculate the Frame remote sensing image features and Matching degree between frame remote sensing image features :

[0047] (12);

[0048] In the formula, Indicates Frame remote sensing image in coordinates The eigenvalue at ; Indicates Frame remote sensing image in coordinates The eigenvalue at ; , All represent displacement;

[0049] S327, reverse calculation: Calculate the matching template Frame remote sensing image features and Matching degree between frame remote sensing image features :

[0050] (13);

[0051] In the formula, Indicates Frame remote sensing image in coordinates The eigenvalue at ;

[0052] S328, set the current Frame and The scene change rate of the frame remote sensing image is set to , search window width The upper and lower thresholds are set to and , then the search window width The rate of scene change The relationship between them is:

[0053] (14);

[0054] in,

[0055] (15);

[0056] In the formula, and are the maximum and minimum values ​​of the scene change rate, respectively;

[0057] S329. Based on the scene change rate of the remote sensing image, an adaptive remote sensing image feature matching search window is obtained to perform a matching search.

[0058] Preferably, step S41 specifically includes the following steps:

[0059] S411, constructing a three-dimensional scanning model;

[0060] S4111, converting the collected remote sensing image into point cloud data to obtain a point set in three-dimensional space;

[0061] S4112, Point cloud registration: Align point cloud data from multiple perspectives to generate a complete 3D model;

[0062] S4113, Model Reconstruction: Use surface reconstruction algorithms to convert point cloud data into a continuous 3D scan model;

[0063] S412, extracting remote sensing image feature points;

[0064] S4121, Feature point detection: Use the BRISK algorithm to detect feature points in remote sensing images;

[0065] S4122, feature description: using FREAK descriptors to represent detected feature points;

[0066] S4123, feature matching: determining the correspondence between remote sensing images by comparing feature points and descriptors between different remote sensing images;

[0067] S413, Remote Sensing Image Classification:

[0068] S4131, Extracting texture features of remote sensing images using Log-Gabor transform;

[0069] S4132, using Krawtchouk moment to extract edge shape features of remote sensing images;

[0070] S4133, fusing the texture features and edge shape features of the remote sensing image to obtain a feature vector;

[0071] S4134, input the feature vector into a support vector machine for classification,

[0072] Preferably, step S42 specifically includes the following steps:

[0073] S421. Feature point management: The extracted feature points are stored in the target manager;

[0074] S422, feature point matching: using a feature point matching algorithm to perform feature point matching between different remote sensing images;

[0075] S423, screening matching points: randomly selecting matching points through the RANSAC algorithm, fitting the three-dimensional scanning model, and then calculating the errors between all matching points and the three-dimensional scanning model to screen out the internal points.

[0076] Preferably, step S43 specifically includes the following steps:

[0077] S431, obtain projection matrix: calculate the projection matrix through camera calibration ;

[0078] S432, coordinate transformation: Use the projection matrix to transform points in three-dimensional space into points on the two-dimensional remote sensing image plane.

[0079] Preferably, in step S5, when prompt information is received, the canvas is first used to draw the text portion of the required prompt information onto the label material remote sensing image, and stored as a picture in the form of a bitmap; then OpenGL ES is used to draw the prompt information of the model or remote sensing image type on the currently captured remote sensing image.

[0080] Preferably, in step S5, the prompt information includes text information and image information.

[0081] Therefore, the present invention adopts the above-mentioned surveying and mapping method based on low-altitude UAV remote sensing technology, which has the following beneficial effects:

[0082] 1. Real-time visualization: AR technology can display remote sensing mapping data in real time in a real environment, allowing users to intuitively see terrain, buildings and other geographic information;

[0083] 2. Wide application: In urban planning, AR can display the three-dimensional effect of the planning scheme, helping decision makers and the public to better understand the planning content; Disaster emergency: In disaster emergency response, AR can display remote sensing images and rescue information of the disaster area in real time to improve rescue efficiency; Agricultural monitoring: In the agricultural field, AR can display remote sensing data such as crop growth and soil moisture to help farmers conduct precise management.

[0084] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] Figure 1 This is a flow chart of a surveying and mapping method based on low-altitude UAV remote sensing technology described in the present invention. DETAILED DESCRIPTION

[0086] In the description of the present invention, it should be noted that the terms "upper", "lower", "inside", "outside", etc. indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, or the positions or positional relationships in which the invented product is usually placed when in use. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In the description of the present invention, it should also be noted that, unless otherwise clearly specified and limited, the terms "setting", "installation", and "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be a connection between the two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0087] The embodiments of the present invention are described in detail below in conjunction with the accompanying drawings.

[0088] like Figure 1 As shown, a surveying and mapping method based on low-altitude UAV remote sensing technology includes the following steps:

[0089] S1. Flight path planning: Plan the flight path of the UAV based on graph theory, potential field method or particle swarm optimization algorithm;

[0090] S2, remote sensing image acquisition: after taking off, the UAV flies in the area to be surveyed according to the flight path described in step S1, and collects remote sensing image information during the flight;

[0091] S3, remote sensing image processing: grayscale processing is performed on the collected remote sensing images, and the noise in the remote sensing images is removed by the median filtering method, and then the Canny operator is used for edge detection to obtain a binary remote sensing image containing only the target object;

[0092] Step S3 specifically includes the following steps:

[0093] S31, Grayscale processing:

[0094] S311, calculating the error after pixel displacement:

[0095] (1);

[0096] In the formula, Represents the total error after pixel displacement; represents the weight function; Represents the gray value of the original remote sensing image; Represents the gray value of the remote sensing image after displacement;

[0097] S312. Obtaining the autocorrelation matrix of remote sensing images using second-order derivatives :

[0098] (2);

[0099] In the formula, , Respectively represent the remote sensing images in Direction and Directional gradient;

[0100] S32, Canny operator edge detection;

[0101] S321. Use a Gaussian filter formula to smooth the grayscale remote sensing image, and the Gaussian filter formula is as follows:

[0102] (3);

[0103] In the formula, represents the response of the Gaussian filter; represents the standard deviation of the Gaussian filter;

[0104] S322, Grayscale Remote Sensing Image Perform a first-order difference to obtain the gradient of each pixel:

[0105] (4);

[0106] (5);

[0107] In the formula, , Represents pixel values exist Direction and Directional gradient; , , , Represents grayscale remote sensing images In pixel value , , , The gray value at ;

[0108] S323, based on and Calculate the gradient magnitude and gradient direction:

[0109] (6);

[0110] (7);

[0111] In the formula, represents the gradient magnitude; Indicates the gradient direction;

[0112] S324, perform non-maximum suppression: pixel values ​​in the remote sensing image The neighborhood along To interpolate, When the pixel value is greater than the interpolation point, is regarded as a candidate edge point, and then the candidate edge points are deleted by the double threshold method to obtain the edge of the remote sensing image; at this time, it is assumed that The number of optimized feature points in a frame is ,and , the number of characteristic lines is ,and , then the calculation formula for the number of iterations is as follows:

[0113] (8);

[0114] (9);

[0115] In the formula, Indicates the temporary number of iterations; , Respectively represent the weights of feature points and feature lines; Indicates the maximum number of iterations; represents the denominator; Indicates the number of iterations;

[0116] S325, using a keyframe-based method to perform hybrid feature matching and calculate camera motion parameters, and calculate the number and length of feature lines:

[0117] (10);

[0118] (11);

[0119] In the formula, The number of temporary points representing the characteristic line; Indicates the number of optimized feature points; Indicates The length of the characteristic line; Indicates the length of the characteristic line; and Represent the mass coefficients of characteristic points and characteristic lines respectively; Indicates the maximum number of points of the feature line; Indicates the number of points after the characteristic line is discretized;

[0120] S326, Remote Sensing Image Forward Matching: Assuming the features of the previous frame of remote sensing image The search window is ,use As the matching template, calculate the Frame remote sensing image features and Matching degree between frame remote sensing image features :

[0121] (12);

[0122] In the formula, Indicates Frame remote sensing image in coordinates The eigenvalue at ; Indicates Frame remote sensing image in coordinates The eigenvalue at ; , All represent displacement;

[0123] S327, reverse calculation: Calculate the matching template Frame remote sensing image features and Matching degree between frame remote sensing image features :

[0124] (13);

[0125] In the formula, Indicates Frame remote sensing image in coordinates The eigenvalue at ;

[0126] S328, set the current Frame and The scene change rate of the frame remote sensing image is set to , search window width The upper and lower thresholds are set to and , then the search window width The rate of scene change The relationship between them is:

[0127] (14);

[0128] in,

[0129] (15);

[0130] In the formula, and are the maximum and minimum values ​​of the scene change rate, respectively;

[0131] S329. Based on the scene change rate of the remote sensing image, an adaptive remote sensing image feature matching search window is obtained to perform a matching search.

[0132] S4, remote sensing image feature extraction;

[0133] S41, constructing a three-dimensional scanning model, and extracting feature points of the remote sensing image using the three-dimensional scanning model;

[0134] Step S41 specifically includes the following steps:

[0135] S411, constructing a three-dimensional scanning model;

[0136] S4111, converting the collected remote sensing image into point cloud data to obtain a point set in three-dimensional space;

[0137] S4112, Point cloud registration: Align point cloud data from multiple perspectives to generate a complete 3D model;

[0138] S4113, Model Reconstruction: Use surface reconstruction algorithms to convert point cloud data into a continuous 3D scan model;

[0139] S412, extracting remote sensing image feature points;

[0140] S4121, Feature point detection: Use the BRISK algorithm to detect feature points in remote sensing images;

[0141] S4122, feature description: using FREAK descriptors to represent detected feature points;

[0142] S4123, feature matching: determining the correspondence between remote sensing images by comparing feature points and descriptors between different remote sensing images;

[0143] S413, Remote Sensing Image Classification:

[0144] S4131, Extracting texture features of remote sensing images using Log-Gabor transform;

[0145] S4132, using Krawtchouk moment to extract edge shape features of remote sensing images;

[0146] S4133, fusing the texture features and edge shape features of the remote sensing image to obtain a feature vector;

[0147] S4134, input the feature vector into a support vector machine for classification,

[0148] S42, using the target manager to match and filter the extracted remote sensing image feature points;

[0149] Step S42 specifically includes the following steps:

[0150] S421. Feature point management: The extracted feature points are stored in the target manager;

[0151] S422, feature point matching: using a feature point matching algorithm to perform feature point matching between different remote sensing images;

[0152] S423, screening matching points: randomly selecting matching points through the RANSAC algorithm, fitting the three-dimensional scanning model, and then calculating the errors between all matching points and the three-dimensional scanning model to screen out the internal points.

[0153] S43, obtaining a projection relationship from the three-dimensional space to the two-dimensional remote sensing image plane based on the projection matrix of the camera;

[0154] Step S43 specifically includes the following steps:

[0155] S431, obtain projection matrix: calculate the projection matrix through camera calibration ;

[0156] S432, coordinate transformation: Use the projection matrix to transform points in three-dimensional space into points on the two-dimensional remote sensing image plane.

[0157] S44, using a projection matrix to perform coordinate transformation, mapping feature points in the remote sensing image to actual positions in the three-dimensional space;

[0158] S45, calculating the specific position and posture of the target object in the three-dimensional space according to the converted coordinates, and completing the positioning of the target scene;

[0159] S5. Draw a map based on the positioning of the target scene.

[0160] In step S5, when prompt information is received, the canvas is first used to draw the text portion of the required prompt information to the label material remote sensing image, and stored as a picture in the form of a bitmap; then OpenGL ES is used to draw the prompt information of the model or remote sensing image type on the currently captured remote sensing image.

[0161] In step S5, the prompt information includes text information and image information.

[0162] Table 1 Evaluation of different fusion algorithms for image processing

[0163] ;

[0164] Table 1 demonstrates the effectiveness of the present invention.

[0165] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A surveying and mapping method based on low-altitude UAV remote sensing technology, characterized in that: The following steps are involved: S1. Flight path planning: Plan the flight path of the UAV based on graph theory, potential field method or particle swarm optimization algorithm; S2, remote sensing image acquisition: after taking off, the UAV flies in the area to be surveyed according to the flight path described in step S1, and collects remote sensing image information during the flight; S3, remote sensing image processing: grayscale processing is performed on the collected remote sensing images, and the noise in the remote sensing images is removed by the median filtering method, and then the Canny operator is used for edge detection to obtain a binary remote sensing image containing only the target object; S4, remote sensing image feature extraction; S41, constructing a three-dimensional scanning model, and extracting feature points of the remote sensing image using the three-dimensional scanning model; S42, using the target manager to match and filter the extracted remote sensing image feature points; S43, obtaining a projection relationship from the three-dimensional space to the two-dimensional remote sensing image plane based on the projection matrix of the camera; S44, using a projection matrix to perform coordinate transformation, mapping feature points in the remote sensing image to actual positions in the three-dimensional space; S45, calculating the specific position and posture of the target object in the three-dimensional space according to the converted coordinates, and completing the positioning of the target scene; S5, mapping based on the positioning of the target scene; In step S5, when prompt information is received, the text portion of the required prompt information is first drawn to the label material remote sensing image using the canvas, and stored as a picture in the form of a bitmap; then, the prompt information of the model or remote sensing image type is drawn on the currently captured remote sensing image using OpenGL ES; In step S5, the prompt information includes text information and image information.

2. The surveying and mapping method based on low-altitude UAV remote sensing technology according to claim 1 is characterized in that: Step S3 specifically includes the following steps: S31, Grayscale processing: S311, calculating the error after pixel displacement: （1）； In the formula, Represents the total error after pixel displacement; represents the weight function; Represents the gray value of the original remote sensing image; Represents the gray value of the remote sensing image after displacement; S312. Obtaining the autocorrelation matrix of remote sensing images using second-order derivatives : （2）； In the formula, , Respectively represent the remote sensing images in Direction and Directional gradient; S32, Canny operator edge detection; S321. Use a Gaussian filter formula to smooth the grayscale remote sensing image, and the Gaussian filter formula is as follows: （3）； In the formula, represents the response of the Gaussian filter; represents the standard deviation of the Gaussian filter; S322, Grayscale Remote Sensing Image Perform a first-order difference to obtain the gradient of each pixel: （4）； （5）； In the formula, , Represents pixel values exist Direction and Directional gradient; , , , Represents grayscale remote sensing images In pixel value , , , The gray value at ; S323, based on and Calculate the gradient magnitude and gradient direction: （6）； （7）； In the formula, represents the gradient magnitude; Indicates the gradient direction; S324, perform non-maximum suppression: pixel values ​​in the remote sensing image The neighborhood along To interpolate, When the pixel value is greater than the interpolation point, is regarded as a candidate edge point, and then the candidate edge points are deleted by the double threshold method to obtain the edge of the remote sensing image; at this time, it is assumed that The number of optimized feature points in a frame is ,and , the number of characteristic lines is ,and , then the calculation formula for the number of iterations is as follows: （8）； （9）； In the formula, Indicates the temporary number of iterations; , Respectively represent the weights of feature points and feature lines; Indicates the maximum number of iterations; represents the denominator; Indicates the number of iterations; S325, using a keyframe-based method to perform hybrid feature matching and calculate camera motion parameters, and calculate the number and length of feature lines: （10）； （11）； In the formula, The number of temporary points representing the characteristic line; Indicates the number of optimized feature points; Indicates The length of the characteristic line; Indicates the length of the characteristic line; and Represent the mass coefficients of characteristic points and characteristic lines respectively; Indicates the maximum number of points of the feature line; Indicates the number of points after the characteristic line is discretized; S326, Remote Sensing Image Forward Matching: Assuming the features of the previous frame of remote sensing image The search window is ,use As the matching template, calculate the Frame remote sensing image features and Matching degree between frame remote sensing image features : （12）； In the formula, Indicates Frame remote sensing image in coordinates The eigenvalue at ; Indicates Frame remote sensing image in coordinates The eigenvalue at ; , All represent displacement; S327, reverse calculation: Calculate the matching template Frame remote sensing image features and Matching degree between frame remote sensing image features : （13）； In the formula, Indicates Frame remote sensing image in coordinates The eigenvalue at ; S328, set the current Frame and The scene change rate of the frame remote sensing image is set to , search window width The upper and lower thresholds are set to and , then the search window width The rate of scene change The relationship between them is: （14）； in, （15）； In the formula, and are the maximum and minimum values ​​of the scene change rate, respectively; S329. Based on the scene change rate of the remote sensing image, an adaptive remote sensing image feature matching search window is obtained to perform a matching search.

3. The surveying and mapping method based on low-altitude UAV remote sensing technology according to claim 2 is characterized in that: Step S41 specifically includes the following steps: S411, constructing a three-dimensional scanning model; S4111, converting the collected remote sensing image into point cloud data to obtain a point set in three-dimensional space; S4112, Point cloud registration: Align point cloud data from multiple perspectives to generate a complete 3D model; S4113, Model Reconstruction: Use surface reconstruction algorithms to convert point cloud data into a continuous 3D scan model; S412, extracting remote sensing image feature points; S4121, Feature point detection: Use the BRISK algorithm to detect feature points in remote sensing images; S4122, feature description: using FREAK descriptors to represent detected feature points; S4123, feature matching: determining the correspondence between remote sensing images by comparing feature points and descriptors between different remote sensing images; S413, Remote Sensing Image Classification: S4131, Extracting texture features of remote sensing images using Log-Gabor transform; S4132, using Krawtchouk moment to extract edge shape features of remote sensing images; S4133, fusing the texture features and edge shape features of the remote sensing image to obtain a feature vector; S4134. Input the feature vector into a support vector machine for classification.

4. The surveying and mapping method based on low-altitude UAV remote sensing technology according to claim 3 is characterized by: Step S42 specifically includes the following steps: S421. Feature point management: The extracted feature points are stored in the target manager; S422, feature point matching: using a feature point matching algorithm to perform feature point matching between different remote sensing images; S423, screening matching points: randomly selecting matching points through the RANSAC algorithm, fitting the three-dimensional scanning model, and then calculating the errors between all matching points and the three-dimensional scanning model to screen out the internal points.

5. The surveying and mapping method based on low-altitude UAV remote sensing technology according to claim 4 is characterized in that: Step S43 specifically includes the following steps: S431, obtain projection matrix: calculate the projection matrix through camera calibration ; S432, coordinate transformation: Use the projection matrix to transform points in three-dimensional space into points on the two-dimensional remote sensing image plane.

Images