CN-121973951-A - Unmanned aerial vehicle photoelectric pod installation error online calibration method based on target tracking

Abstract

The invention discloses an unmanned aerial vehicle photoelectric pod installation error online calibration method based on target tracking, which comprises the steps of S1 designing an online calibration system comprising a photoelectric pod, a flight controller, an onboard processor and a calibration reference, S2 selecting a target with known absolute position information as the calibration reference of the online calibration system and training a detection target on the ground, S3 searching the detection target through the pod in the unmanned aerial vehicle taking-off process and carrying out closed-loop tracking on the detection target through a pod control module when the detection target appears in an image view field, and S4 solving an optimal closed-loop solution of the installation error by the onboard processor based on each information parameter acquired in the process of carrying out closed-loop tracking on the target so as to correct the installation error of the photoelectric pod. According to the method, the pod installation error calibration problem is modeled into the optimal rotation matrix estimation problem through combining multiple observations in the time domain, so that an optimal closed solution of the installation error is obtained, and the calibration precision is improved.

Inventors

• XU YONG
• YAO TIAN
• MAO ZHONGJUN
• YAN HONGTAO
• XU HAIHANG
• MA YUE
• WANG AN
• SONG HONGCHUAN
• REN YOUCHENG
• HUANG HAO

Assignees

• 中国空气动力研究与发展中心空天技术研究所

Dates

Publication Date

20260505

Application Date

20260121

Claims (4)

1. 1. The unmanned aerial vehicle photoelectric pod installation error on-line calibration method based on target tracking is characterized by comprising the following steps of: S1, designing an online calibration system comprising a photoelectric pod, a flight controller, an onboard processor and a calibration reference; S2, selecting a target with known absolute position information as a calibration standard of an online calibration system, and training a detection target on the ground; S3, in the take-off process of the unmanned aerial vehicle, the nacelle is controlled by ground control software of the nacelle to search for a detection target, and when the detection target appears in an image view field, closed-loop tracking is carried out on the detection target by a nacelle control module in an onboard processor; S4, in the process of carrying out closed-loop tracking on the detection target, the airborne processor acquires the position and posture information of the unmanned aerial vehicle at each moment through the flight controller, and acquires the absolute position information of the detection target at each moment, the state information of the photoelectric pod and the state information of image tracking through the photoelectric pod; S5, the onboard processor obtains an optimal closed solution of the installation error based on the information parameters obtained in the S4 through the following formula: In the above formula, U and V are orthogonal matrixes obtained by singular value decomposition of the measurement information matrix B, To singular values 、 、 The diagonal matrix obtained by the decomposition is used, The mounting error to be calibrated corresponding to the angle mounting in the nacelle d system from the machine body coordinate b system; S6, correcting the installation error of the electro-nacelle based on the resolving result of the S4 so as to complete on-line calibration.
2. 2. The unmanned aerial vehicle photoelectric pod installation error online calibration method based on target tracking according to claim 1, wherein in S1, the onboard processor adopts a multi-task asynchronous parallel processing architecture, and comprises a communication thread, a detection thread, a tracking thread and a comprehensive thread; The communication thread comprises a nacelle communication module, a flight control communication module and a link communication module; The comprehensive thread comprises a nacelle control module and an installation error resolving module.
3. 3. The unmanned aerial vehicle photoelectric pod installation error online calibration method based on target tracking according to claim 2, wherein the workflow of the online calibration system comprises: the photoelectric pod transmits the image information acquired in real time to a detection thread and a tracking thread through a pod communication module, and the state information of the pod and the frame angle information of the pod are transmitted to a comprehensive thread through the pod communication module; the detection thread identifies whether a detection target appears or not based on the received image information, and sends a detection result to the tracking thread; The tracking thread determines whether to track the detection target based on the detection result and sends the tracking result to the pod control module; The pod control module generates a pod closed-loop control instruction by adopting active disturbance rejection control based on a tracking result and pod state information so as to transmit the control instruction to the photoelectric pod through the pod communication module to complete the control of the photoelectric pod; The flight control communication module acquires unmanned aerial vehicle state information from the flight controller and transmits the unmanned aerial vehicle state information to the installation error resolving module; The link communication module acquires the position information of the reference target in real time through a data link and transmits the position information to the installation error resolving module; The installation error calculation module calculates and obtains the solution of the installation error of the nacelle according to the formula in S5 based on the nacelle frame angle information, the unmanned plane state information, the target position information and the target closed loop tracking result The method for on-line calibration of the installation error of the unmanned aerial vehicle photoelectric pod based on target tracking according to claim 1, wherein in S4, the acquisition flow of the optimal closed solution of the installation error is as follows: Coordinates of the detection target in the camera coordinate system Characterization was performed by the following formula: ; In the above-mentioned method, the step of, A rotation matrix of the camera system to the nacelle system, For a rotation matrix of the north-east coordinate system to the body system, For the position installation error based on the three-dimensional digital-analog or measurement means of the unmanned aerial vehicle, Is a coefficient of proportionality and is used for the control of the power supply, The coordinates of the projected image point in the camera system that is the reference target, Coordinates of vectors pointing to the detection target of the unmanned aerial vehicle in a north-east coordinate system; The observation equation containing nacelle installation errors is: ; In the above-mentioned method, the step of, The coordinates of the unit vector under the nacelle system, which is obtained by the coordinate transformation and points to the target at the origin of the nacelle system, Coordinates in the machine system of a unit vector pointing to the target for the origin of the nacelle system; Combining n observations in the time domain, modeling the calibration problem as an optimization problem with constraint conditions of the following formula: In the above-mentioned method, the step of, Is that Is to be used in the present invention, Is a unit matrix of order 3, To optimize the objective function of the problem, a rotation matrix can be obtained according to the Kabsch algorithm Is a closed-form solution.
4. 4. The unmanned aerial vehicle photoelectric pod installation error online calibration method based on target tracking according to claim 1, wherein the following different calibration schemes are adopted in different application scenarios: According to the scheme I, for a rotor unmanned aerial vehicle system, a cooperative target is arranged at a flying spot of the unmanned aerial vehicle or other known geographic positions, and nacelle closed-loop tracking is carried out on the cooperative target in the process of taking off and forward-out execution of a task of the unmanned aerial vehicle, so that the calibration of nacelle installation errors is realized; in the scheme II, for a fixed wing unmanned aerial vehicle system, a cooperative target is arranged in front of the take-off of the unmanned aerial vehicle, the cooperative target is tracked in the process of the sliding take-off of the unmanned aerial vehicle, and meanwhile, the online calibration of the nacelle installation error is completed; The third scheme can be used for calibrating the air-ground coordination system and the cluster unmanned aerial vehicle system in a mode of setting a coordination target in the first scheme and the second scheme, calibrating the air-ground coordination system based on a ground unit in the air-ground coordination system and an air unit in the cluster unmanned aerial vehicle system, and selecting a machine for calibrating or recalibrating in the process of executing tasks.

Description

Unmanned aerial vehicle photoelectric pod installation error online calibration method based on target tracking Technical Field The invention relates to the field of unmanned aerial vehicle photoelectric load installation error calibration. More particularly, the invention relates to an unmanned aerial vehicle photoelectric pod installation error online calibration method based on target tracking. Background With the rapid development of unmanned aerial vehicle technology and AI technology, unmanned aerial vehicle is continuously expanded to civil fields such as emergency rescue, agriculture and forestry plant protection, electric power inspection and the like from the traditional single military application, and gradually becomes a key node of current low-altitude economy. The photoelectric nacelle is used as the core task load of the unmanned aerial vehicle, the quick and high-precision calibration of the installation error is a basic premise for completing various tasks, and the quality of the calibration result directly influences the task efficiency of the unmanned aerial vehicle. Under ideal conditions, the pod base coordinate system and the carrier coordinate system are completely coincident, but are influenced by factors such as space limitations on the carrier, process errors, assembly process, deformation of the shock absorber and the like, so that a non-negligible installation error exists between the pod base coordinate system and the carrier coordinate system. If the calibration compensation is not performed, the milli-centimeter difference of the installation error of the photoelectric pod can cause the effect of kilo-liter during the remote operation of the unmanned aerial vehicle. Particularly, in the case of target positioning, the mounting error will significantly affect the positioning accuracy of the target. The deviation is particularly prominent in long-distance and high-precision task scenes (such as accurate striking, disaster search and rescue and accurate delivery), and directly restricts the fight efficiency and the task reliability of the unmanned aerial vehicle system. According to different calibration means, the existing calibration methods can be divided into two types, namely ground static calibration and air dynamic calibration. The traditional ground calibration method relies on special instruments such as a high-precision turntable, a laser theodolite and the like, and error compensation is realized through repeated mechanical adjustment and optical measurement. Although this approach can achieve higher accuracy in laboratory environments, it still faces a number of critical issues, firstly, its operation is cumbersome, time consuming, and requires stringent environmental stability, with poor adaptability. Secondly, the ground calibration method is designed based on static or quasi-static conditions, the calibration environment is too different from a high-dynamic scene in the actual flight of the unmanned aerial vehicle, and therefore the calibration result is not optimal for the actual task. More critical is, after unmanned aerial vehicle transition or nacelle dismantles maintenance, must recalibrate, has reduced the attendance of equipment obviously, has increased the maintenance cost, is difficult to satisfy unmanned aerial vehicle quick deployment's demand, becomes the bottleneck that restricts unmanned aerial vehicle emergency response ability. The aerial dynamic calibration method breaks through the environmental constraint of the traditional ground calibration method, and the calibration of the nacelle installation error is realized by utilizing real-time measurement data in the flight process of the unmanned aerial vehicle. Most of the existing aerial calibration methods need to calibrate special calibration tracks for unmanned aerial vehicles, track and position ground cooperative targets, and finally calculate the installation errors of the nacelle by combining state information of a carrier and an optoelectronic nacelle. However, in the existing air dynamic calibration method, the error range of the initial installation error angle is limited or the approximate processing is carried out in the resolving process, so that the resolved result is not the optimal solution. In addition, because the target needs to be positioned, the existing aerial dynamic calibration method is mostly suitable for calibrating the installation error of the photoelectric pod with laser ranging, and cannot cope with the situations of a single-camera pod, a simple cradle head camera and the like. Therefore, the rapid, high-precision, universal and on-line calibration supporting unmanned aerial vehicle photoelectric pod installation error calibration technology is developed, and the following three challenges are mainly met at present: 1. High-precision requirement of unmanned aerial vehicle accurate operation scene on calibration method Along with the deepening application of unmanned a