CN114812512A - Automatic imaging system of unmanned aerial photography based on AI exempts from image control point - Google Patents

Abstract

The invention provides an AI-free image control point-based unmanned aerial automatic imaging system, which comprises unmanned aerial vehicle equipment, wherein the unmanned aerial vehicle equipment is a fixed-wing unmanned aerial vehicle, and is loaded with a power module, a communication module, a laser scanner, a GNSS high-precision positioning module, an MEMS inertial navigation module, an IMU attitude measurement module, a high-speed data acquisition and storage module and a full-frame aerial camera; the method comprises the following specific steps: the method comprises the steps of accurately measuring camera parameters, drawing a three-dimensional route, erecting a base station, obtaining data information stored by a mobile station and a reference station, calculating three-dimensional geographic coordinate information, automatically flying and shooting, importing a current unmanned aerial vehicle low-altitude photography aerial photograph, reading EXIF information and calibrating the precision of rare image control points. According to the invention, aerial photos can be imported in batches for multiple times, the imported photos are subjected to image quality detection, the coordinate error range in a flight line is analyzed, the writing of PPK data and the drawing of a track map are rapidly carried out, high-precision mapping is realized in dangerous areas, and safety risks are effectively avoided.

Description

An unmanned aerial photography automatic imaging system based on AI-free image control point

technical field

The invention relates to the technical field of aerial photography, in particular to an unmanned aerial photography automatic imaging system based on AI-free image control points.

Background technique

Traditional aerial photogrammetry of large aircraft has shortcomings such as long flight time, poor maneuverability, low real-time performance, high cost, limited accuracy, and unsuitable for dangerous areas. The rapid development of UAV low-altitude digital aerial photogrammetry technology in recent years makes up for it. Overcoming the shortcomings of traditional manned aerial surveys, it is gradually becoming an effective supplement to satellite remote sensing, manned remote sensing and ground remote sensing, and is especially suitable for quickly obtaining large-scale digital maps with high precision in small areas and difficult-to-fly areas. However, UAV aerial survey images have smaller image frames and more image pairs, and the image quality is easily affected by weather and flight quality. UAV low-altitude photogrammetry has more image control points and higher distribution requirements. The field workload of the control measurement is larger.

The GPS single-point positioning accuracy of the UAV flight control is too poor. 3D modeling requires the use of a large number of image control points to correct the distortion of the image. However, it is difficult for field personnel to give up the image control points in some special terrains. In order to reduce the workload , it is necessary to increase the longitude of the POS point of the aircraft to reduce most of the image control points or even without the need for image control points. PPK technology can achieve centimeter-level accuracy. The PPK base station maintains continuous observation, and the initialized rover moves to the next to-be-determined point and maintains continuous tracking of the satellite. Calculate the coordinate position of the rover in response time. The photos taken by the UAV oblique photography are written into the photos by using the coordinates generated by the PPK technology, which is of great significance to the 3D modeling, and there is no need to correct the image distortion in the later stage. Therefore, it is very necessary to design an intelligent writing method of oblique photographic images based on PPK data without image control.

SUMMARY OF THE INVENTION

The present invention proposes an unmanned aerial photography automatic imaging system based on AI-free image control points, which can solve the problems raised in the above background technology.

The technical solution of the present invention is realized as follows: an unmanned aerial photography automatic imaging system based on AI free image control point, including unmanned aerial vehicle equipment, the unmanned aerial vehicle equipment is a fixed-wing unmanned aerial vehicle, and a power supply is mounted on it. modules, communication modules, laser scanners, GNSS high-precision positioning modules, MEMS inertial navigation modules, IMU attitude measurement modules, high-speed data acquisition and storage modules, and full-frame aerial cameras;

Specific steps include:

a. Accurate measurement of camera parameters: Accurately calibrate the interior orientation elements of the aerial camera based on the outdoor 3D calibration field to obtain accurate camera parameters, lens distortion parameters and the eccentricity of the camera GNSS antenna installation;

b. 3D route drawing and base station erection: Calculate route interval, photographic baseline, relative flight height, lowest point resolution, highest point course overlap, highest point lateral overlap; Before the modeling and mapping device takes off, set up a base station consisting of a base station GNSS receiver and a static base station radio assembly;

c. Obtain the data information stored by the mobile station and the base station, solve the three-dimensional geographic coordinate information, and linearly combine the obtained position data through specific computer software to form a virtual carrier phase observation, and determine the relative relationship between the receivers. position, introduce the known coordinates of the base station, and obtain the three-dimensional coordinates of the UAV aerial photography;

d. Automatic flight and shooting, remote control by the erected base station or automatic flight of the fixed-wing UAV controlled by the autopilot without image control point 3D modeling and mapping device according to the flight route designed in step 2;

e. Import the current low-altitude photography aerial photos of the UAV, and read the EXIF information: The UAV takes photos at an isometric or isochronous time according to the specified trajectory route, records the attribute information and shooting data of the digital photos, and reads the EXIF through the software. information to obtain the location information of the corresponding photo;

f. Import the coordinate information collected by the current drone flight after PPK processing, save the processed data as a CSV format file, import the CSV format file through this system and read the PPK processing data, and import the PPK data with one click. aerial photo

g. Additional integrated system error parameters for free image control point air three calculation;

h. Accuracy calibration of rare image control points. Add four image control points in the four corners of the survey area to calibrate the object square of the free image control point in step 7, and eliminate the systematic difference of scale, direction and system offset.

Preferably, the full-frame aerial camera is rigidly and fixedly connected under the IMU attitude measurement module and connected with its electrical signal; the autopilot module is respectively connected with the GNSS high-precision positioning module, the communication module and the IMU attitude measurement module with electrical signals, and the MEMS The inertial navigation module is connected with the full-frame aerial camera through the camera exposure line; the ground is provided with an electrical signal connection control module with the communication module; the GNSS high-precision positioning module, the MEMS inertial navigation module and the communication module are all electrically connected to the power supply module.

Preferably, the GNSS high-precision positioning module includes at least an airborne multi-mode high-frequency GNSS receiver, a GNSS receiving antenna, an epoch data storage, an RTK communication link radio station and an electronic coupling connection accessory.

As a preference, when measuring mountainous terrain, if the mountainous area has small undulations, that is, the height difference inside the mountainous area is small, and the relative flight height is basically the same under the requirement of the same ground resolution, so that all routes are at the same absolute flight height; if the mountain has large undulations , that is, the height difference inside the mountainous area is large, and under the requirements of the same ground resolution, the routes are at different absolute altitudes.

Preferably, in the step a, the camera parameters are accurately measured, and the camera parameters include: the camera principal point position (x0, y0) and the camera principal distance (f); the lens distortion parameters include: radial distortion coefficient k1, diameter Directional distortion coefficient k2, radial distortion coefficient k3, tangential distortion coefficient p1, tangential distortion coefficient p2, area array deformation coefficient α and area array deformation coefficient β; camera GNSS antenna installation eccentricity (ΔX, ΔY, ΔZ).

Preferably, in the step b, using the measured camera parameters, loading the published global DEM data and the range line in the KML format of the shooting area, according to the principle of flight altitude calculation, carry out the calculation of the three-dimensional route drawing parameters, and obtain and calculate the route interval and photography baseline. , relative altitude, resolution of the lowest point, course overlap of the highest point, and side overlap of the highest point.

Preferably, in the step d, when flying, a single-lens camera or a tilting multi-lens camera is mounted to perform automatic aerial photography to obtain real-time dynamic differential RTK data or post-differential PPK data, IMU attitude data and aerial photography images.

Preferably, in the step f, one-key import of PPK data into the aerial photo includes the following steps: judging whether the imported aerial photo has POS point coordinates; according to the current UAV aerial photo and the PPK post-processing data, carry out quantitative matching, and then match and import in turn .

Compared with the prior art, the present invention has the advantages that: the aerial photos can be imported in batches for many times, the image quality of the imported photos can be detected, the coordinate error range within the flight route can be analyzed, and the PPK data writing and the drawing of the trajectory map can be performed quickly. . The fixed-wing UAV is equipped with a non-measurement aerial camera. After the aerial photography is completed, the aerial triangulation can be completed without any ground image control point measurement work. The aerial survey internal product processing can be carried out directly. Under the control point calibration, the surveying and mapping accuracy is better than 1:500 scale. In the operation process, the measurement process of the field ground image control point is removed, and the direct connection of the operation mode from aerial photography to the internal calculation is realized, which reduces the time and cost of the field image control point measurement. Accurate mapping and effectively avoid security risks.

Description of drawings

FIG. 1 is a schematic flow chart of the present invention.

Detailed ways

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present invention.

Embodiment: Referring to FIG. 1, an unmanned aerial photography automatic imaging system based on AI free image control point, including unmanned aerial vehicle equipment, the unmanned aerial vehicle equipment is a fixed-wing unmanned aerial vehicle, and it is equipped with a power supply module, communication modules, laser scanners, GNSS high-precision positioning modules, MEMS inertial navigation modules, IMU attitude measurement modules, high-speed data acquisition and storage modules, and full-frame aerial cameras;

Specific steps include:

a. Accurate measurement of camera parameters: Accurately calibrate the interior orientation elements of the aerial camera based on the outdoor 3D calibration field to obtain accurate camera parameters, lens distortion parameters and the eccentricity of the camera GNSS antenna installation;

b. 3D route drawing and base station erection: Calculate route interval, photographic baseline, relative flight height, lowest point resolution, highest point course overlap, highest point lateral overlap; Before the modeling and mapping device takes off, set up a base station consisting of a base station GNSS receiver and a static base station radio assembly;

c. Obtain the data information stored by the mobile station and the base station, solve the three-dimensional geographic coordinate information, and linearly combine the obtained position data through specific computer software to form a virtual carrier phase observation, and determine the relative relationship between the receivers. position, introduce the known coordinates of the base station, and obtain the three-dimensional coordinates of the UAV aerial photography;

d. Automatic flight and shooting, remote control by the erected base station or automatic flight of the fixed-wing UAV controlled by the autopilot without image control point 3D modeling and mapping device according to the flight route designed in step 2;

e. Import the current low-altitude photography aerial photos of the UAV, and read the EXIF information: The UAV takes photos at an isometric or isochronous time according to the specified trajectory route, records the attribute information and shooting data of the digital photos, and reads the EXIF through the software. information to obtain the location information of the corresponding photo;

f. Import the coordinate information collected by the current drone flight after PPK processing, save the processed data as a CSV format file, import the CSV format file through this system and read the PPK processing data, and import the PPK data with one click. aerial photo

g. Additional integrated system error parameters for free image control point air three calculation;

h. Accuracy calibration of rare image control points. Add four image control points in the four corners of the survey area to calibrate the object square of the free image control point in step 7, and eliminate the systematic difference of scale, direction and system offset.

Preferably, the full-frame aerial camera is rigidly and fixedly connected under the IMU attitude measurement module and connected with its electrical signal; the autopilot module is respectively connected with the GNSS high-precision positioning module, the communication module and the IMU attitude measurement module with electrical signals, and the MEMS The inertial navigation module is connected with the full-frame aerial camera through the camera exposure line; the ground is provided with an electrical signal connection control module with the communication module; the GNSS high-precision positioning module, the MEMS inertial navigation module and the communication module are all electrically connected to the power supply module.

Preferably, the GNSS high-precision positioning module includes at least an airborne multi-mode high-frequency GNSS receiver, a GNSS receiving antenna, an epoch data storage, an RTK communication link radio station and an electronic coupling connection accessory.

As a preference, when measuring mountainous terrain, if the mountainous area has small undulations, that is, the height difference inside the mountainous area is small, and the relative flight height is basically the same under the requirement of the same ground resolution, so that all routes are at the same absolute flight height; if the mountain has large undulations , that is, the height difference inside the mountainous area is large, and under the requirements of the same ground resolution, the routes are at different absolute altitudes.

Preferably, in the step a, the camera parameters are accurately measured, and the camera parameters include: the camera principal point position (x0, y0) and the camera principal distance (f); the lens distortion parameters include: radial distortion coefficient k1, diameter Directional distortion coefficient k2, radial distortion coefficient k3, tangential distortion coefficient p1, tangential distortion coefficient p2, area array deformation coefficient α and area array deformation coefficient β; camera GNSS antenna installation eccentricity (ΔX, ΔY, ΔZ).

Preferably, in the step b, using the measured camera parameters, loading the published global DEM data and the range line in the KML format of the shooting area, according to the principle of flight altitude calculation, carry out the calculation of the three-dimensional route drawing parameters, and obtain and calculate the route interval and photography baseline. , relative altitude, resolution of the lowest point, course overlap of the highest point, and side overlap of the highest point.

Preferably, in the step d, when flying, a single-lens camera or a tilting multi-lens camera is mounted to perform automatic aerial photography to obtain real-time dynamic differential RTK data or post-differential PPK data, IMU attitude data and aerial photography images.

Preferably, in the step f, one-key import of PPK data into the aerial photo includes the following steps: judging whether the imported aerial photo has POS point coordinates; according to the current UAV aerial photo and the PPK post-processing data, carry out quantitative matching, and then match and import in turn .

The intelligent writing method of oblique photography images based on image control-free PPK data of the present invention can import aerial photos in batches for many times, perform image quality detection on the imported photos, analyze the coordinate error range within the flight route, and quickly write PPK data. input and plotting of trajectory graphs. The fixed-wing UAV is equipped with a non-measurement aerial camera. After the aerial photography is completed, the aerial triangulation can be completed without any ground image control point measurement work. The aerial survey internal product processing can be carried out directly. Under the control point calibration, the surveying and mapping accuracy is better than 1:500 scale. In the operation process, the measurement process of the field ground image control point is removed, and the direct connection of the operation mode from aerial photography to the internal calculation is realized, which reduces the time and cost of the field image control point measurement. Accurate mapping and effectively avoid security risks.

In actual use, the communication module is used to receive instructions from the ground control module, and the power module is responsible for supplying power to the fixed-wing flight platform and various electronic modules on it. The base station GNSS receiver of the control module on the ground receives the coordinate information, and the static base station radio assembly radio transmits the coordinate to the airborne communication module and sends it to the autopilot module. The autopilot module is responsible for controlling the flight of the entire fixed-wing flight platform 1 according to the coordinate information and the pre-drawn three-dimensional route, and triggers pulses to the IMU attitude measurement module, full-frame and airborne GNSS high-precision positioning modules at the same time. The IMU attitude measurement module marks the time according to the pulse and records the angle information. Full frame time stamps according to the pulse and takes pictures. The airborne GNSS high-precision positioning module marks the time according to the pulse and records the coordinate information. Complete a signal transmission process.

The epoch acquisition frequency of the airborne multi-mode high-frequency GNSS receiver is not less than 20HZ, the read and write speed of the epoch data memory is not less than 100MB/s, the communication radius of the RTK communication link is not less than 5km when the radio is not blocked, and the electronic coupling connection The marked time difference from the attachment sent from the autopilot pulse signal to the airborne multi-mode high-frequency GNSS receiver and the IMU attitude measurement module is not more than 1ms. In actual use, when the speed of the fixed-wing UAV flight platform is not greater than 20 m/s, the GNSS high-precision positioning module can accurately obtain the spatial coordinates of the exposure point using static PPK or dynamic RTK modes.

Use the software to analyze the UAV route coordinate trajectory, and linearly combine the data received synchronously by the base station and the UAV rover to form a virtual carrier phase observation, determine the relative position between the receivers, and introduce the known data of the base station. Coordinates to obtain the three-dimensional coordinates of the drone aerial photography. The obtained three-dimensional coordinates are intelligently written into the photo EXIF information by means of comparison, search and trapping, etc., to achieve accurate three-dimensional modeling effects; through the PPK technology, the coordinate information can be imported from images taken without image control, and high-precision aerial photography is completed. In the process of drawing and large-area real scene 3D reconstruction, it is easy to avoid the problems of dislocation and delamination when the model blocks are merged, which will easily lead to the failure of the model, and then the model cannot be reconstructed.

Advantages of the free image control solution No need to deploy phase control in the field: shorten the field time, and the mountain, river and valley area can be reduced by 10 times; personnel and equipment are safer, and there is no need to travel frequently. The internal industry improves the efficiency of aerial triangulation: aerial triangulation does not require puncture points, making aerial triangulation intelligent; the aerial triangulation time is shortened, and professional steps are simplified. Ensure the uniform accuracy of the whole map: Stereo mapping, plane 10cm, elevation 5cm; 3D mapping, plane 3cm, elevation 5cm. Kering Huifei’s image control-free solution greatly reduces the user’s field and office time, improves aerial survey accuracy, and fundamentally solves the problem of aerial survey accuracy control. Kering Huifei’s two core technologies—POS refinement and high-precision camera inspection The two core technologies of the school solve the core problems of aerial triangulation, the three-dimensional position (external orientation element) at the time of aerial photo exposure, the distortion parameters of the aerial camera (internal orientation element), and provide accurate internal and external orientation elements to the Air3 software, Guaranteed 100% success of air three. POS Refinement The POS position is a POS refinement of the pole-arm effect for RTK antenna estimation.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the scope of the present invention. within the scope of protection.

Claims (8)

1. a kind of unmanned aerial photography automatic imaging system based on AI free image control point, comprises unmanned aerial vehicle equipment, it is characterized in that: described unmanned aerial vehicle equipment is fixed-wing unmanned aerial vehicle, is equipped with power supply module, communication module on it , laser scanner, GNSS high-precision positioning module, MEMS inertial navigation module, IMU attitude measurement module, high-speed data acquisition and storage module and full-frame aerial camera; Specific steps include: a. Accurate measurement of camera parameters: Accurately calibrate the interior orientation elements of the aerial camera based on the outdoor 3D calibration field to obtain accurate camera parameters, lens distortion parameters and the eccentricity of the camera GNSS antenna installation; b. 3D route drawing and base station erection: Calculate route interval, photographic baseline, relative flight height, lowest point resolution, highest point course overlap, highest point lateral overlap; Before the modeling and mapping device takes off, set up a base station consisting of a base station GNSS receiver and a static base station radio assembly; c. Obtain the data information stored by the mobile station and the base station, solve the three-dimensional geographic coordinate information, and linearly combine the obtained position data through specific computer software to form a virtual carrier phase observation, and determine the relative relationship between the receivers. position, introduce the known coordinates of the base station, and obtain the three-dimensional coordinates of the UAV aerial photography; d. Automatic flight and shooting, remote control by the erected base station or automatic flight of the fixed-wing UAV controlled by the autopilot without image control point 3D modeling and mapping device according to the flight route designed in step 2; e. Import the current low-altitude photography aerial photos of the UAV, and read the EXIF information: The UAV takes photos at an isometric or isochronous time according to the specified trajectory route, records the attribute information and shooting data of the digital photos, and reads the EXIF through the software. information to obtain the location information of the corresponding photo; f. Import the coordinate information collected by the current drone flight after PPK processing, save the processed data as a CSV format file, import the CSV format file through this system and read the PPK processing data, and import the PPK data with one click. aerial photo g. Additional integrated system error parameters for free image control point air three calculation; h. Accuracy calibration of rare image control points. Add four image control points in the four corners of the survey area to calibrate the object square of the three-free image control points in step 7, and eliminate the systematic differences in scale, direction and system offset. 2. A kind of unmanned aerial photography automatic imaging system based on AI free image control point according to claim 1, it is characterized in that: described full-frame aerial photography camera is rigidly and fixedly connected below the IMU attitude measurement module and connected with its electrical signal The autopilot module is respectively connected with the GNSS high-precision positioning module, the communication module and the IMU attitude measurement module with electrical signals, and the MEMS inertial navigation module is connected with the full-frame aerial camera through the camera exposure line; the ground is provided with an electrical signal connection with the communication module A control module; the GNSS high-precision positioning module, the MEMS inertial navigation module and the communication module are all electrically connected to the power supply module. 3. A kind of unmanned aerial photography automatic imaging system based on AI free image control point according to claim 2, is characterized in that: described GNSS high-precision positioning module at least comprises airborne multi-mode high-frequency GNSS receiver, GNSS receiver Antenna, epoch data storage, RTK communication link radio and electronic coupling attachments. 4. A kind of unmanned aerial photography automatic imaging system based on AI image-free control point according to claim 1, it is characterized in that: when measuring mountainous terrain, if the mountain area is small in undulation, that is, the height difference inside the mountain area is small, and it can be resolved on the same ground. Under the requirements of the flight rate, the relative flight heights are basically the same, so that all the routes are at the same absolute flight height; if the mountains are undulating, that is, the height difference within the mountains is large, and under the same ground resolution requirements, the flight routes are in different absolute flight heights. high. 5. a kind of unmanned aerial photography automatic imaging system based on AI free image control point according to claim 1, is characterized in that: in described step a, camera parameter is measured accurately, and described camera parameter comprises: camera principal point position ( x0, y0) and the main distance of the camera (f); the lens distortion parameters include: radial distortion coefficient k1, radial distortion coefficient k2, radial distortion coefficient k3, tangential distortion coefficient p1, tangential distortion coefficient p2, surface Array deformation coefficient α and area array deformation coefficient β; camera GNSS antenna installation eccentricity (ΔX, ΔY, ΔZ). 6. A kind of unmanned aerial photography automatic imaging system based on AI free image control point according to claim 5, it is characterized in that: in described step b, utilize the measured camera parameters, load public global DEM data and photograph area KML format range line, according to the principle of flight height calculation, calculate the parameters of 3D route drawing, and obtain and calculate the route interval, photographic baseline, relative flight height, lowest point resolution, highest point course overlap degree, and highest point sideways overlap degree. 7. A kind of unmanned aerial photography automatic imaging system based on AI free image control point according to claim 6, it is characterized in that: in described step d, when flying, carry down-view single-lens camera or tilt multi-lens camera to carry out automatic navigation. Capture real-time dynamic differential RTK data or post-differential PPK data, IMU attitude data and aerial images. 8. a kind of unmanned aerial photography automatic imaging system based on AI free image control point according to claim 7, is characterized in that: in described step f, one-key importing PPK data to aerial photograph comprises the following steps: judging to import aerial photograph Whether there are POS point coordinates; according to the current UAV aerial photo and PPK post-processing data, the quantity is matched, and the matching is imported.

Images