Target relative pose resolving method based on monocular vision depth estimation

Abstract

The invention discloses a target relative pose resolving method based on monocular vision depth estimation, which comprises the following steps: firstly, calibrating an internal reference matrix of a monocular camera, shooting a target image for preprocessing and target detection, and extracting pixel coordinates of at least two feature points; the method comprises the following steps: firstly, acquiring the depth of a feature point through a monocular distance measurement algorithm based on a pinhole imaging model and a similarity principle, acquiring attitude data of a flight control system, carrying out air communication, and finally calculating the relative pose between aircrafts by combining a projection transformation relationship between 2D-3D point pairs established based on a camera mathematical model. According to the method, the relative pose between the aircrafts can be solved only by adopting the monocular camera to process the single-frame image and the at least two feature points, the problems that a traditional monocular vision positioning algorithm is high in complexity and long in consumed time are solved, the detection requirement for the feature points is reduced, meanwhile, the depth of the feature points is directly measured through a monocular distance measurement algorithm, and the accuracy of the relative pose between the aircrafts is improved. And the calculation complexity is further reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of monocular vision positioning, and in particular to a method for solving target relative posture based on monocular vision depth estimation. Background Art

[0002] Aerial refueling technology is a technical means to effectively improve combat distance and mobility in modern warfare. For drones, which are emerging air combat forces in modern warfare, one of their most prominent weaknesses is their short flight time. They need to return to the base regularly for resupply, which greatly weakens the ability of drones to perform long-duration and complex tasks, limiting their use. To resolve this contradiction, organizations led by the National Aeronautics and Space Administration (NASA) have launched a number of automated aerial refueling (AAR) verification programs. Applying autonomous aerial refueling technology to drones is both a trend and a necessity.

[0003] Autonomous aerial refueling technology is essentially to complete the aerial docking of drones. The solution of relative posture is an important part of drone aerial docking. At present, the sensors used to obtain relative posture mainly include inertial navigation system (INS), differential global positioning system (DGPS), visual sensors, etc. The inertial navigation system will produce cumulative errors over time and cannot meet the accuracy requirements during the docking process. DGPS can meet the accuracy requirements, but the GPS signal is easily blocked during the docking process, resulting in a decrease in accuracy. Visual sensors have been widely used in autonomous aerial refueling due to their high accuracy and low price.

[0004] Liu Aichao et al. proposed a relative pose estimation method based on ellipse detection and contour detection in the UAV docking process in their paper "Visual Navigation Method in UAV Aerial Docking". This method first uses the mobile docking icon and the GPS / INS information of the UAV to perform rough navigation of the UAV and complete the rendezvous process; then uses the color and shape information of the mobile docking icon to select possible target areas through color segmentation method, and performs fast ellipse detection and contour detection in these areas to obtain accurate icon edge description information; finally, the improved orthogonal iteration (Orthogonal Iteration, OI) algorithm is used to estimate the relative pose and achieve accurate navigation at close range. The disadvantage of this method is that it uses an iterative solution method when solving the relative pose. This method has high computational complexity, occupies more resources, and has a long running time, which will affect the real-time performance of the relative pose estimation of the aircraft to a certain extent.

[0005] Xidian University has disclosed a method for estimating target pose based on the fusion of monocular vision and inertial navigation in its patent application "Target pose estimation method based on the fusion of monocular vision and inertial navigation" (patent application number CN202410265306.3, patent publication number CN117968640A). This method obtains the relative pose between the aircraft and the identifiable target on the ground through the inertial navigation data of the aircraft; by measuring the pixel coordinates of two feature points, the distance formula is combined with the projection transformation relationship between the 2D-3D point pairs established based on the camera mathematical model to solve the relative position between the aircraft. The shortcomings of this method are: first, since this method relies on inertial navigation to obtain attitude data and lacks communication, it is only suitable for measuring the relative attitude between an aircraft and a fixed target on the ground, and cannot measure the relative attitude between an aircraft and another movable target; second, this method applies the projection transformation relationship between 2D-3D point pairs established based on the camera mathematical model. In this projection transformation relationship, since the number of unknowns is greater than the number of equations, the relative position cannot be solved when only the pixel coordinates of the feature points are obtained. To solve this problem, this method introduces a distance formula as a supplement to the equation. Although this solves the problem, the introduced distance formula is a nonlinear equation, which undoubtedly increases the computational complexity. Summary of the invention

[0006] The technical problem to be solved by the present invention is to provide a method for solving the relative pose of a target based on monocular vision depth estimation in view of the defects involved in the background technology.

[0007] The present invention adopts the following technical solutions to solve the above technical problems:

[0008] The target relative pose solution method based on monocular vision depth estimation includes the following steps:

[0009] Step 1), perform monocular camera calibration;

[0010] Step 2), using a calibrated monocular camera to capture a target image and preprocess it, and then performing target detection in the preprocessed image to obtain a target image containing at least two feature points;

[0011] Step 3), extracting the pixel coordinates of each feature point in the target image;

[0012] Step 4), obtaining the depth information of each feature point through a monocular ranging algorithm based on a pinhole imaging model and similar principles;

[0013] Step 5), obtaining attitude data of the flight control system and performing air communication to obtain the relative attitude between the aircraft;

[0014] Step 6), using the pixel coordinates and depth information of each feature point, combined with the corresponding projection transformation relationship between 2D-3D point pairs established based on the camera mathematical model, the relative positions between the aircraft are solved;

[0015] Step 7), the relative posture and relative position between the aircraft are sorted into a homogeneous transformation matrix to obtain the target relative posture.

[0016] As a further optimization scheme of the target relative pose solution method based on monocular vision depth estimation of the present invention, the method for calibrating the monocular camera in step 1) adopts any one of the following: Tsai two-step method, Zhang Zhengyou calibration method, active vision camera calibration method, camera self-calibration method, calibration based on genetic algorithm, calibration based on neural network, calibration based on error back propagation.

[0017] As a further optimization scheme of the target relative pose solution method based on monocular visual depth estimation of the present invention, the monocular camera intrinsic parameter matrix parameters in step 1) include: the focal length of the monocular camera, the horizontal pixel offset, the vertical pixel offset, and the pixel coordinates of the center of the monocular camera image.

[0018] As a further optimization scheme of the target relative pose solution method based on monocular vision depth estimation of the present invention, the method for target image preprocessing described in step 2) adopts any one of the following: binarization, grayscale, filtering, histogram equalization, image enhancement, geometric change, image interpolation, image segmentation, and image morphological operation.

[0019] As a further optimization scheme of the target relative pose solution method based on monocular vision depth estimation of the present invention, the target detection method for the preprocessed image described in step 2) adopts any one of the following: contour detection, edge detection, key point detection, interest point detection, texture analysis, feature matching.

[0020] As a further optimization scheme of the target relative pose solution method based on monocular vision depth estimation of the present invention, the method for extracting pixel coordinates of feature points in the target image described in step 3) adopts any one of the following: image moment calculation, fast corner detection, Harris corner detection, and Hough transform method.

[0021] As a further optimization scheme of the target relative pose solution method based on monocular vision depth estimation of the present invention, the formula of the monocular ranging algorithm based on the pinhole imaging model and similar principle for obtaining the depth information of the feature points described in step 4) is as follows:

[0022]

[0023] In the formula, z c is the depth of the feature point, and also the z-axis coordinate of the feature point in the camera coordinate system; f is the focal length of the monocular camera; w is the actual length of the line segment formed by the two feature points as endpoints; P is the Euclidean distance between the two feature points in the pixel coordinate system.

[0024] As a further optimization scheme of the target relative pose solution method based on monocular vision depth estimation of the present invention, the detailed steps of step 5) are as follows:

[0025] Step 5.1), during the flight, the pitch angle, yaw angle and roll angle of the aircraft at any time are obtained through the flight control system, and the aircraft attitude matrix is ​​obtained by combining the coordinate transformation law. The aircraft attitude matrix expression is as follows:

[0026]

[0027] Where R is the aircraft attitude matrix; θ is the aircraft pitch angle; ψ is the aircraft yaw angle; φ is the aircraft roll angle;

[0028] Step 5.2), transmit the aircraft attitude matrix through air communication, and use the three-dimensional coordinate system transformation relationship to solve the relative attitude between the aircraft. The three-dimensional coordinate system transformation relationship expression is as follows:

[0029]

[0030] In the formula, R BA is the attitude matrix of aircraft B relative to aircraft A; is the transpose of the attitude matrix of aircraft A; R B is the attitude matrix of aircraft B.

[0031] As a further optimization scheme of the target relative pose solution method based on monocular vision depth estimation of the present invention, the projection transformation relationship between the 2D-3D point pairs established based on the camera mathematical model described in step 6) is as follows:

[0032]

[0033] In the formula, x c ,y c 、z c are the x-axis, y-axis, and z-axis coordinates of the feature point in the camera coordinate system; u and v are the x-axis and y-axis coordinates of the feature point in the pixel coordinate system; f is the focal length of the monocular camera; dx is the horizontal pixel offset; dy is the vertical pixel offset; u0 and v0 are the x-axis and y-axis coordinates of the center of the monocular camera image in the pixel coordinate system.

[0034] Compared with the prior art, the present invention adopts the above technical solution and has the following technical effects:

[0035] 1. The relative position and posture between aircraft can be obtained by processing a single frame of image with a monocular camera, which overcomes the problems of high complexity and long time consumption of traditional monocular vision positioning algorithms, and ensures the real-time and accuracy of relative position and posture estimation during aircraft docking.

[0036] 2. Only at least two feature points are needed to estimate the relative pose between aircrafts, which solves the problem that traditional monocular vision positioning algorithms require at least three feature points to estimate the relative pose between aircrafts, and reduces the detection requirements for feature points in relative pose estimation.

[0037] 3. The depth information of the feature points is directly measured by using a monocular ranging algorithm, which reduces the number of unknown quantities in the projection transformation relationship between 2D-3D point pairs established based on the camera mathematical model, so that the unique relative position can be directly solved. Compared with solving by supplementary equations, the present invention has lower computational complexity and shorter running time. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 is a flow chart of the present invention;

[0039] Figure 2 (a) Figure 2 (b) are respectively a scene simulation diagram established in the simulation experiment of the present invention and a captured visual sensor image containing feature points;

[0040] Figure 3 (a) Figure 3 (b) Figure 3 (c) are the original target image captured by the calibrated monocular camera and the Figure 3 (a) The result after image segmentation and image morphological operation based on HSV color space. Figure 3 (b) The result after binarization processing;

[0041] Figure 4 It is a schematic diagram of the principle of the monocular ranging algorithm based on the pinhole imaging model and similar principles in the present invention;

[0042] Figure 5 It is a test result diagram of the simulation experiment of the present invention. DETAILED DESCRIPTION

[0043] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings:

[0044] The present invention can be implemented in many different forms and should not be considered to be limited to the embodiments described herein. On the contrary, these embodiments are provided to make this disclosure thorough and complete, and will fully express the scope of the present invention to those skilled in the art. In the accompanying drawings, components are enlarged for clarity.

[0045] The idea of ​​achieving the purpose of the present invention is that the present invention uses the real-time communication between the flight control system and the aircraft to obtain the relative posture between the aircraft, without using a multi-eye camera or processing multiple frames of images to obtain the relative posture, thereby solving the problem of long running time and high resource consumption of conventional visual positioning algorithms. After preprocessing and detecting the target image, the present invention first obtains the depth information of the feature points in the target image through a monocular ranging algorithm, and then uses the pixel coordinates of at least two feature points, combined with the projection transformation relationship between 2D-3D point pairs established based on the camera mathematical model, to solve the relative position between the aircraft, thereby solving the problem that the traditional monocular visual positioning algorithm requires at least three feature points to estimate the relative posture between aircraft, and reduces the detection requirements for feature points in relative posture estimation. In terms of posture solution, the present invention uses a flight control system to obtain the current pitch angle, yaw angle and roll angle of the aircraft, and obtains the relative posture between the aircraft through real-time communication; uses the depth information and pixel coordinates of the feature points in the target image, and directly calculates the relative position between the aircraft through the projection transformation relationship between 2D-3D point pairs established based on the camera mathematical model, without the need to use complex solution methods such as neural networks and iterations, thereby solving the problems of high computational complexity and long algorithm running time in solving the posture of traditional monocular vision positioning algorithms.

[0046] The implementation steps of the present invention are as follows:

[0047] Step 1), perform monocular camera calibration;

[0048] Step 2), using a calibrated monocular camera to capture a target image and preprocess it, and then performing target detection in the preprocessed image to obtain a target image containing at least two feature points;

[0049] Step 3), extracting the pixel coordinates of each feature point in the target image;

[0050] Step 4), obtaining the depth information of the feature points through a monocular ranging algorithm based on a pinhole imaging model and similar principles;

[0051] Step 5), obtaining attitude data of the flight control system and performing air communication to obtain the relative attitude between the aircraft;

[0052] Step 6), using the pixel coordinates and depth information between each feature point, combined with the corresponding projection transformation relationship between 2D-3D point pairs established based on the camera mathematical model, the relative positions between the aircraft are solved;

[0053] Step 7) Arrange the relative posture and relative position between the aircraft into a homogeneous transformation matrix to obtain the target relative posture. Figure 1 , the implementation steps of the present invention are further described.

[0054] Step 1: calibrate the monocular camera.

[0055] For the calibration of monocular cameras, the methods that can be selected include but are not limited to: Tsai two-step method, Zhang Zhengyou calibration method, active vision camera calibration method, camera self-calibration method, calibration based on genetic algorithm, calibration based on neural network, and calibration based on error back propagation.

[0056] In an embodiment of the present invention, the Zhang Zhengyou calibration method is used to calibrate the monocular camera. The specific operation is: place the black and white chessboard used for calibration in a suitable position, take 23 pictures with a complete chessboard with different camera postures, import the taken pictures into the Camera Calibrator APP program of the MATLAB (Matrix Laboratory) software, and calibrate after entering the specifications of the black and white chessboard. During the calibration process, if the reprojection error of some pictures is too large, they can be deleted and recalibrated. It is only necessary to ensure that the number of pictures used for calibration is 15-20. If the calibration reprojection error is not greater than 0.3, the calibration is considered successful, and the internal parameter matrix of the monocular camera after calibration is obtained from the results output by the MATLAB software. The standard format of the internal parameter matrix is ​​as follows:

[0057]

[0058] Where I is the intrinsic parameter matrix of the monocular camera; f is the focal length of the monocular camera; dx is the horizontal pixel offset; dy is the vertical pixel offset; u0 and v0 are the x-axis and y-axis coordinates of the center of the monocular camera image in the pixel coordinate system, respectively.

[0059] Step 2: Use a calibrated monocular camera to capture the target image and preprocess it, and then perform target detection in the preprocessed image.

[0060] Figure 2 (a) is a simulation diagram of a drone aerial docking scenario created in CoppeliaSim Edu 4.6 simulation software. Figure 2 (b) is the image containing feature points captured by the visual sensor in the simulation software.

[0061] For image preprocessing, the methods that can be selected include but are not limited to: binarization, grayscale, filtering, histogram equalization, image enhancement, geometric transformation, image interpolation, image segmentation, and image morphological operations.

[0062] The embodiment of the present invention sequentially uses image segmentation, image morphological operation, and binarization to preprocess the image of the monocular camera. Figure 3 (a) is the original target image captured by the monocular camera after calibration. Figure 3 (b) Yes Figure 3 (a) The result after image segmentation and image morphological operation based on HSV color space. Figure 3 (c) Yes Figure 3 (b) The result after binarization processing.

[0063] For object detection in preprocessed images, the methods that can be selected include but are not limited to: contour detection, edge detection, key point detection, interest point detection, texture analysis, and feature matching.

[0064] In the embodiment of the present invention, a contour detection algorithm is used to find the contours of all feature points in the preprocessed image.

[0065] Step 3: Extract the pixel coordinates of the feature points in the target image.

[0066] Extract pixel coordinates of feature points in the target image. Selectable methods include but are not limited to: image moment calculation, fast corner detection, Harris corner detection, and Hough transform method.

[0067] In the embodiment of the present invention, the pixel coordinates of the centroid of all feature points in the target image are obtained by calculating the image moments of all feature point contours found and combining the relationship between the image moments and the pixel coordinates of the image centroid. The relationship between the image moments and the pixel coordinates of the image centroid is as follows:

[0068]

[0069] In the formula, C x 、M y Respectively represent the x-axis and y-axis coordinates of the image centroid in the pixel coordinate system; M 10 and M 01 is the first-order moment of the image, indicating the quality distribution of the image; M 00is the zeroth-order moment of the image, which represents the overall quality of the image.

[0070] Step 4: Obtain the depth information of the feature points through a monocular ranging algorithm based on the pinhole imaging model and similar principles.

[0071] The monocular ranging algorithm used in the present invention is based on the pinhole imaging model and similar principles, and its principle diagram is shown in FIG. Figure 4 As shown, the feature point depth z can be derived from the similar principle c As shown below:

[0072]

[0073] In the formula, z c is the depth of the feature point, and also the z-axis coordinate of the feature point in the camera coordinate system; f is the focal length of the monocular camera; w is the actual length of the line segment formed by the two feature points as endpoints; P is the Euclidean distance between the two feature points in the pixel coordinate system.

[0074] Step 5, obtain the attitude data of the flight control system and perform air communication to calculate the relative attitude between the aircraft.

[0075] In an embodiment of the present invention, the pitch angle, yaw angle and roll angle of the aircraft at any time can be obtained by utilizing the flight control system, and the aircraft attitude matrix can be obtained by combining the coordinate transformation rule. The aircraft attitude matrix is ​​then transmitted via air communication, and finally the relative attitude between aircraft can be calculated using the three-dimensional coordinate system transformation relationship.

[0076] The aircraft attitude matrix is ​​expressed as follows:

[0077]

[0078] Where R is the aircraft attitude matrix; θ is the aircraft pitch angle; ψ is the aircraft yaw angle; φ is the aircraft roll angle.

[0079] Specifically, suppose there are two aircraft A and B in the air, and the attitude matrices obtained by their respective flight control systems are R A and R B , if you want to find the attitude matrix R of aircraft B relative to aircraft A BA , then aircraft B will change its attitude matrix R B By transmitting it to aircraft A through communication and combining it with the three-dimensional coordinate system conversion relationship, we can get:

[0080]

[0081] In the formula, R BA is the attitude matrix of aircraft B relative to aircraft A; is the transpose of the attitude matrix of aircraft A; R B is the attitude matrix of aircraft B.

[0082] Step 6, using the pixel coordinates and depth information of the two feature points, combined with the corresponding projection transformation relationship between 2D-3D point pairs established based on the camera mathematical model, the relative positions between the aircraft are solved.

[0083] The projection transformation relationship between the 2D-3D point pairs established based on the camera mathematical model is shown in the following formula:

[0084]

[0085] In the formula, x c ,y c 、z c are the x-axis, y-axis, and z-axis coordinates of the feature point in the camera coordinate system; u and v are the x-axis and y-axis coordinates of the feature point in the pixel coordinate system; f is the focal length of the monocular camera; dx is the horizontal pixel offset; dy is the vertical pixel offset; u0 and v0 are the x-axis and y-axis coordinates of the center of the monocular camera image in the pixel coordinate system.

[0086] In the embodiment of the present invention, since the pixel coordinates (u, v) of the feature point have been obtained by calculating the image moment, the depth z of the feature point has been obtained by the monocular ranging algorithm. c , at this time directly pixel coordinates (u, v) and feature point depth z c Substituting into the above formula, we can solve x by solving the linear equation c ,y c , that is, the relative positions between the aircraft are solved.

[0087] The relative positions between the aircrafts obtained by the final solution are compared with the relative positions between the aircrafts given by the simulation software. The comparison results are as follows: Figure 5 As shown, when the lateral distance between the aircraft is not too large, the algorithm result and the simulation result are almost consistent, and the error is at the millimeter level; when the lateral distance is large, although there is a certain error between the algorithm result and the simulation result, the error still remains at the centimeter level, meeting the accuracy requirements of the aircraft air docking. In addition, since the present invention only needs to solve linear equations, the algorithm running time is very short and occupies very few computing resources. Therefore, it is shown that the present invention can use at least two feature points to quickly and accurately solve the relative posture between aircraft by performing monocular visual depth estimation and air communication.

[0088] It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as those generally understood by those skilled in the art in the art to which the present invention belongs. It should also be understood that terms such as those defined in common dictionaries should be understood to have meanings consistent with the meanings in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless defined as herein.

[0089] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for solving target relative pose based on monocular vision depth estimation, characterized in that: The following steps are included: Step 1), perform monocular camera calibration; Step 2), using a calibrated monocular camera to capture a target image and preprocess it, and then performing target detection in the preprocessed image to obtain a target image containing at least two feature points; Step 3), extracting the pixel coordinates of each feature point in the target image; Step 4), obtaining the depth information of each feature point through a monocular ranging algorithm based on a pinhole imaging model and similar principles; Step 5), obtaining attitude data of the flight control system and performing air communication to obtain the relative attitude between the aircraft; Step 6), using the pixel coordinates and depth information of each feature point, combined with the corresponding projection transformation relationship between 2D-3D point pairs established based on the camera mathematical model, the relative positions between the aircraft are solved; Step 7), the relative posture and relative position between the aircraft are sorted into a homogeneous transformation matrix to obtain the target relative posture.

2. The method for solving target relative pose based on monocular vision depth estimation according to claim 1, characterized in that: In step 1), the method for calibrating the monocular camera is any one of the following: Tsai two-step method, Zhang Zhengyou calibration method, active vision camera calibration method, camera self-calibration method, calibration based on genetic algorithm, calibration based on neural network, and calibration based on error back propagation.

3. The method for solving target relative pose based on monocular vision depth estimation according to claim 1, characterized in that: The monocular camera intrinsic parameter matrix parameters in step 1) include: the focal length of the monocular camera, the horizontal pixel offset, the vertical pixel offset, and the pixel coordinates of the center of the monocular camera image.

4. The target relative pose solution method based on monocular vision depth estimation and air communication according to claim 1 is characterized in that: The method for preprocessing the target image described in step 2) is any one of the following: binarization, grayscale, filtering, histogram equalization, image enhancement, geometric change, image interpolation, image segmentation, and image morphological operation.

5. The target relative pose solution method based on monocular vision depth estimation and air communication according to claim 1 is characterized in that: The target detection method for the preprocessed image described in step 2) adopts any one of the following: contour detection, edge detection, key point detection, interest point detection, texture analysis, feature matching.

6. The target relative pose solution method based on monocular vision depth estimation and air communication according to claim 1 is characterized in that: The method for extracting pixel coordinates of feature points in the target image described in step 3) adopts any one of the following methods: image moment calculation, fast corner point detection, Harris corner point detection, and Hough transform method.

7. The target relative pose solution method based on monocular vision depth estimation and air communication according to claim 1 is characterized in that: The formula of the monocular ranging algorithm based on the pinhole imaging model and similar principle for obtaining the depth information of the feature points described in step 4) is as follows: In the formula, z c is the depth of the feature point, and also the z-axis coordinate of the feature point in the camera coordinate system; f is the focal length of the monocular camera; w is the actual length of the line segment formed by the two feature points as endpoints; P is the Euclidean distance between the two feature points in the pixel coordinate system.

8. The target relative pose solution method based on monocular vision depth estimation and air communication according to claim 1 is characterized in that: The detailed steps of step 5) are as follows: Step 5.1), during the flight, the pitch angle, yaw angle and roll angle of the aircraft at any time are obtained through the flight control system, and the aircraft attitude matrix is ​​obtained by combining the coordinate transformation law. The aircraft attitude matrix expression is as follows: Where R is the aircraft attitude matrix; θ is the aircraft pitch angle; ψ is the aircraft yaw angle; φ is the aircraft roll angle; Step 5.2), transmit the aircraft attitude matrix through air communication, and use the three-dimensional coordinate system transformation relationship to solve the relative attitude between the aircraft. The three-dimensional coordinate system transformation relationship expression is as follows: In the formula, R BA is the attitude matrix of aircraft B relative to aircraft A; is the transpose of the attitude matrix of aircraft A; R B is the attitude matrix of aircraft B.

9. The method for solving target relative pose based on monocular vision depth estimation and air communication according to claim 1, characterized in that: The projection transformation relationship between the 2D-3D point pairs established based on the camera mathematical model described in step 6) is as follows: In the formula, x c ,y c 、z c are the x-axis, y-axis, and z-axis coordinates of the feature point in the camera coordinate system; u and v are the x-axis and y-axis coordinates of the feature point in the pixel coordinate system; f is the focal length of the monocular camera; dx is the horizontal pixel offset; dy is the vertical pixel offset; u0 and v0 are the x-axis and y-axis coordinates of the center of the monocular camera image in the pixel coordinate system.

Images