77results about How to "Improve target positioning accuracy" patented technology

The present invention discloses a simplified unmanned aerial vehicle multi-target location method. The simplified unmanned aerial vehicle multi-target location method can perform location in real time while realizing multiple targets, and is simple in hardware device. An area array CCD camera is employed to perform grounding imaging, and obtaining the pixel coordinates of each target; performing distortion correction, and obtaining the ideal imaging points of each target; constructing the sight vectors of a main target, a sub target and a lower machine point in the camera coordinate systems; employing the coordinate transformation to obtain the sight vectors of the main target, the sub target and the lower machine point in the aerial carrier geographic coordinate system; calculating the coordinates of each target in the aerial carrier geographic coordinate system according to the distance from the main target and the sub target to a photoelectricity platform and the sight vectors of the main target and the sub target in the aerial carrier geographic coordinate system; and finally, obtaining the geodetic coordinates through the coordinate transformation, namely locating the result.

The invention discloses an unmanned aerial vehicle ground target real-time positioning method with automatic extraction and gathering of multiple control points, and belongs to the unmanned aerial vehicle reconnaissance image processing field. The method comprises the following steps: a first step, acquiring N (N>=3) unmanned aerial vehicle target reconnaissance images satisfying a certain condition; a second step, extracting and correcting coordinates of a quasi-ground control point (GCP) of each reconnaissance image, and forming a quasi-GCP group; a third step, establishing a panoramic remote sensing image through an image stitching algorithm, and achieving quasi-GCP gathering; a fourth step, carrying out geometric correction of the panoramic image based on the quasi-GCP; and a fifth step, calculating geographic latitude and longitude coordinates of a reconnaissance target, and achieving target positioning. By introducing the multi-image coordinate extraction mechanism, the algorithm effectively overcomes a contradiction between real-time performance and positioning accuracy in a conventional unmanned aerial vehicle target positioning algorithm, and can effectively improve the positioning accuracy while ensuring the algorithm real-time performance.

The invention discloses a vehicle location positioning information fusion method based on a vehicular ad hoc network. The method comprises the following steps of S1, determining the number of vehicles around a target vehicle in the ad hoc network, if the number of the vehicles is n, running a single vehicle data fusion method for n times, and if four vehicles are located around the target vehicle, running the single vehicle data fusion method for four times; and S2, using a fuzzy nearness fusion method to remove negligent data therein, and working out a weight according to a close degree to obtain the vehicle location positioning information. According to the method, the error influence on ranging caused by network delay is reduced and eliminated through the difference between relative driving distances within the delay, and it assumes that the own vehicle image ranging and radar ranging are real-time. When the data such as distance and the like is used for fusing, the received positioning information of the other vehicles in the system is pre-corrected, the data fusion is performed, and at last fusion is performed again aiming at each auxiliary vehicle fusion result, thus the target positioning accuracy is improved.

The invention discloses an unmanned aerial vehicle visible light and infrared image target positioning method under a large squint angle. The method includes the steps that firstly, target positioning is performed based on visible light and infrared images of an imaging model; secondly, target positioning error characteristic extraction and representation are performed under various-factor influences; thirdly, target positioning error prediction and compensation are performed at variable heights and the large squint angle. Under the condition of the large squint angle, the unmanned aerial vehicle target positioning precision can be effectively improved. The target positioning and error compensation method is small in calculation amount and can meet the real-time requirement of airborne computation. The method is suitable for target positioning application of images, conforming to the central projection imaging model, such as the visible light images and the infrared images.

The invention discloses an external radiation source TDOA/FDOA error correction based positioning method. For the problem of external radiation source positioning under a system deviation in measurement, according to obtained TDOA and FDOA measurement information, an auxiliary variable is introduced to convert a nonlinear equation into a pseudo linear equation; an estimation model for a target position, a speed and the system deviation is built; and estimation is performed by adopting an iterative weighted least square method. A correlation least square estimation model is built by utilizing the correlation between the auxiliary variable, and the target position and the speed; and an estimation result is improved by means of correlation weighted least squares. The target position, the speed and the system deviation are jointly estimated to improve the target positioning precision. The auxiliary variable is introduced to reasonably convert a nonlinear measurement model into a pseudo linear estimation model, so that the complexity of the external radiation source positioning is reduced on the premise that the estimation performance is ensured.

The invention discloses a mine moving target positioning method based on RSS (received signal strength) and TOA (time of arrival) complementation, which combines an RSS method and a TOA method and overcomes the defect of overlarge positioning error caused by a traditional single positioning method. The method comprises adopting a distance measurement interval method to suppress an NLOS error signal, solving a distance correction weight value according to a distance relative error value between every two nearest neighbor positioning base stations in the roadway, and performing distance correction on the distance measurement value; obtaining the weight of each distance measurement value by adopting a speed threshold function method, and optimizing the distance measurement values according to the weights; and performing accurate positioning calculation on the corrected distance estimation value through a weighted positioning method. The method solves the problem that an existing mine personnel positioning system and method is influenced by the environment of a roadway and is insufficient in positioning accuracy, is simple and effective in positioning process, has strong anti-electromagnetic interference capability, and can effecively realize the accurate positioning of underground coal mine personnel, vehicles and other moving targets.

The invention discloses a probing and tracking camera for targets on solar observation blind areas, which can be arranged on a satellite to observe and track targets on the solar observation blind areas. The camera comprises an attenuation filter, an optical system, a probe and a data processing circuit; wherein, a ray of light from an object space passes through the attenuation filter and the optical system and forms an image on the probe. The attenuation filter is used to adjust the intensity of light coming into the camera, and the optical system adopts a double-Gauss near infrared teleobjective system; the probe adopts a CMOS probe, and the dynamic range is designed as follows: the brightness of the sun, sunspot images is designed out of the dynamic range of the camera, and the brightness of the observed targets is designed within the saturation value of the camera. The data processing circuit conducts reverse-phase processing over the images output by the probe, so as to enable the system to acquire higher positioning precision of the targets. The probing and tracking camera in the invention, together with a sun sensor, a two-dimensional tracker and an image processing system, forms a whole probing and tracking system for targets.

The invention discloses an indoor underwater target positioning method based on a genetic algorithm, wherein an underwater target positioning problem is converted to an optimization problem, and furthermore the optimization problem is solved through applying a hybrid genetic algorithm, thereby obtaining a positioning result. An objective function, a fitness function and a constrained condition which are resolved in the genetic algorithm are determined according to a specific indoor underwater object positioning system. Compared with the prior art, the indoor underwater target positioning method has advantages of greatly improving object positioning precision, realizing high convergence reliability and relatively high convergence speed, and obtaining relatively low effect of a measurement error to positioning precision. Furthermore the indoor underwater target positioning method can be expanded for being applied in other systems for positioning the object based on range finding information.

The invention discloses a target positioning method for double-flow convolutional neural network regression learning based on a depth image. In the offline stage, at each reference position, the grayscale image and the depth image corresponding to the grayscale image are collected by a binocular camera. By using image preprocessing techniques, the grayscale image and the depth image are convertedinto three-channel images. Then, the double-flow CNN with the shared weight coefficient is used for offline regression learning. Finally, a regression model based on the distance is obtained. In the online stage, after the obtained grayscale image and depth image are preprocessed, the final distance is estimated through a regression model based on the distance.

The invention relates to a method for estimating the installation error angle of an optical pod of an unmanned aerial vehicle, and the method comprises the following steps that during the flight of the unmanned aerial vehicle, the optical pod is used for locking a fixed target with known accurate coordinates, an optical pod installation error angle resolving mathematical model which takes unmanned aerial vehicle coordinates, an unmanned aerial vehicle attitude angle, an optical pod pointing angle and a laser ranging value as known parameters is established, the installation error angle of theoptical pod is estimated through a multi-Newton method, and the target positioning is carried out with the installation error angle of the optical pod as the known parameter. The method can accurately estimate the installation error angle of the optical pod, improve the target positioning precision of the optical pod of the unmanned aerial vehicle, realize the installation error calibration and the target accurate positioning of the optical pod in the same flight frame, simplify the installation procedure of the optical pod of the unmanned aerial vehicle, and improve the quick response capability of the unmanned aerial vehicle.

The invention discloses a passive positioning method of employing cooperative short wave and satellite system for an over-the-horizon target. The method comprises the steps of establishing pseudo-linear observation equations about a plurality of satellite uplink time difference parameters by using an auxiliary variable; establishing a mathematic relation of various short wave observation station arrival azimuths about horizontal coordinate parameters of a target on various stations on the basis of local horizontal coordinate systems of various short wave observation stations, and establishingpseudo-linear observation equations based on a plurality of short wave observation station azimuth parameters by using conversion of the coordinate parameters of the target in horizontal coordinate systems of various stations and earth-centered earth-fixed coordinate parameters; establishing a global least squares optimization model in a quadratic constraint condition by combining the pseudo-linear observation equations of the plurality of satellite uplink time difference parameters with the pseudo-linear observation equations of the plurality of short wave observation station azimuth parameters and combining earth ellipsoid constraints; and designing an augmented Lagrange multiplier algorithm to achieve accurate positioning of the ground over-the-horizon target. According to the passive positioning method, two positioning mechanisms can be cooperatively utilized effectively, so that the positioning accuracy on the target is significantly improved.

The invention discloses an improved YOLOv3 model side-scan sonar shipwreck target automatic identification method based on transfer learning, and belongs to the technical field of side-scan sonar image target identification and deep learning. The improved YOLOv3 model side-scan sonar image shipwreck target identification method based on transfer learning solves the problems in existing side-scan sonar image manual interpretation and manual feature extraction, and also solves the problems that a Faster R-CNN model is poor in small target recognition effect, high in missing alarm rate and low inrecognition speed. The identification and positioning precision of a shipwreck target is further improved, so that the model achieves a better convergence effect, and finally, the purposes of improving the overall performance of the model and achieving real-time detection are achieved.

The invention relates to a target positioning method and device based on multiple satellites, electronic equipment and a computer readable medium. The method comprises: according to a positioning request of a target object, obtaining a plurality of direction-finding positioning data sets of a plurality of satellites for the target object and a plurality of time-frequency difference data of the plurality of satellites, wherein the direction-finding positioning data sets comprise direction-finding positioning data of a plurality of periods; training a machine learning model according to the plurality of direction-finding positioning data sets and the plurality of time-frequency difference data to generate a multi-satellite positioning model; and inputting the plurality of direction-finding positioning data sets into the multi-satellite positioning model to obtain the accurate position of the target object. According to the target positioning method and device based on the multiple satellites, the electronic equipment and the computer readable medium, the positioning long-period data of the multiple satellites are fused, the incidence relation among the long-period data of the multiple satellites of the same target is fully mined, decision fusion is achieved, and the target positioning precision is improved.

The invention discloses a multi-station and multiple-external-radiation-source radar moving target positioning method under a receiving station position error. The method comprises the steps of, according to an obtained bistatic range and a bistatic range rate, using the distance from a target to the receiving station and a distance change rate as auxiliary variables to pseudo-linearize a strong nonlinear equation; establishing an estimation model of the target position and velocity; integrating statistical characteristics of an observation station position and a velocity measurement error into a positioning algorithm for designing weights; constructing a constraint weighted least squares localization model by using the association between the auxiliary variables and the target position and velocity as a constraint condition; and optimizing and solving the model by a Lagrange multiplier method. The method introduces the auxiliary variables, converts a nonlinear measurement model into pseudo-linear model, reduces the complexity of multi-station and multiple-external-radiation-source positioning under the premise of ensuring the estimation performance, and designs the weighted leastsquares estimation according to the intermediate variable and the association, thereby reducing the influence of the observation station error on the target positioning performance.

The invention discloses a grid pixel attenuation value adaptive determining method for wireless tomography, and the method sequentially comprises the following steps: (1), deploying n sensor nodes around a monitoring region; (2), calibrating a communication link; (3), determining a weight value w<ij> of a j-th grid pixel of an i-th communication link in a formula; (4), carrying out the Tikhonov normalized image reconstruction of an equation (4), and obtaining a grid pixel attenuation value. According to the invention, the method introduces a distance attenuation factor, and enables a grid pixel weight value in an elliptical area to adaptively change with the distance from the grid pixel to a communication link LOS path. The correction of the weight value is more matched with the propagation characteristics of a radio frequency signal. Through experiments, the method, compared with a standard elliptical model, is higher in energy focusing degree at a target position of the reconstructedimage, and is smaller in regional area at a pseudo position, so the method can achieve the higher positioning precision.

The invention discloses an unmanned aerial vehicle high-precision positioning method based on TDOA and FDOA, wherein the method comprises the following steps: receiving a TDOA measurement value and anFDOA measurement value through a receiving station in space, and enabling the TDOA measurement value and the FDOA measurement value to be equivalent to a distance difference measurement value and a distance difference change rate measurement value; performing joint estimation on the position coordinates of the receiving station to construct a to-be-estimated vector; constructing expressions between the corresponding distance difference measurement value, the distance difference change rate measurement value and the receiving station position measurement value and the to-be-estimated vector; solving a nonlinear weighted least square optimization model of a cost error function; solving an optimal closed-form solution of the to-be-estimated vector through Taylor series joint iteration; and obtaining the position and speed of the target when an iteration termination condition is satisfied. According to the method, the position coordinates of the receiving station are subjected to joint iterative estimation, so that the influence of the position error of the receiving station on the positioning precision is reduced; and a more accurate iterative initial value is provided by utilizing positioning error correction, and the target positioning precision is improved while the convergence is ensured and the convergence speed is improved.

The invention provides an azimuth-equivalent delay difference passive positioning method of double circular arrays. The method comprises the steps: firstly building a double-array space positioning model, and obtaining output data x1 (t) of each array element of two circular arrays, wherein l equals to1 and 2; calculating azimuth angles theta1 and theta2 of a target relative to the two array reference points by utilizing a spatial spectrum estimation algorithm; calculating a time delay difference tauij between the ith array element of the circular array 1 and the jth array element of the circular array 2 by using a time delay estimation algorithm, and calculating an equivalent time delay difference taue by using the tauij; and substituting the azimuth angles theta1 and theta2 and the equivalent delay difference taue into an azimuth-equivalent delay difference positioning formula, and obtaining a distance R1R2 from the target to each array reference point and a coordinate position coordinate (xs, ys) of the target. Reference array elements are prevented from being added to a center of the uniform circular array, an array structure is simplified, system cost is saved, meanwhile, array redundant information is utilized, time delay estimation precision is improved, and then target positioning precision is increased.

The invention provides an indoor positioning method based on channel state information. The indoor positioning method mainly comprises the following steps: measuring and preprocessing the channel state information; constructing a CSI amplitude image and a CSI phase image; extracting color features in the CSI amplitude image and the phase image; constructing a training database based on color features of the CSI amplitude image and the phase image; performing offline learning through a convolutional neural network to obtain a position classification model; and preprocessing the CSI obtained online, constructing a CSI amplitude image and a phase image, respectively extracting color features, and substituting the color features into a position classification model to predict/estimate a target position. According to the invention, the color features of the CSI amplitude image and the CSI phase image are used as positioning fingerprints, and through principal component analysis and wavelettransform, the data dimension can be reduced, the noise of the CSI amplitude information and the CSI phase information can be removed, and the offline learning efficiency can be improved; meanwhile,classification learning is carried out through the convolutional neural network, and the target positioning precision can be improved.