CN112489091B - Full strapdown image seeker target tracking method based on direct-aiming template - Google Patents
Full strapdown image seeker target tracking method based on direct-aiming template Download PDFInfo
- Publication number
- CN112489091B CN112489091B CN202011508984.6A CN202011508984A CN112489091B CN 112489091 B CN112489091 B CN 112489091B CN 202011508984 A CN202011508984 A CN 202011508984A CN 112489091 B CN112489091 B CN 112489091B
- Authority
- CN
- China
- Prior art keywords
- target
- template
- image
- matching
- seeker
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
- G06T7/251—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments involving models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/223—Analysis of motion using block-matching
- G06T7/238—Analysis of motion using block-matching using non-full search, e.g. three-step search
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
- G06V10/751—Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Databases & Information Systems (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Aiming, Guidance, Guns With A Light Source, Armor, Camouflage, And Targets (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a full strapdown image seeker target tracking method based on a direct-view template, which corrects a target image by using the direct-view template and inertial navigation information, so that the target image is consistent with the view angle and the scale between the target image acquired by a full strapdown image seeker, and has the characteristic of high tracking precision; the target tracking range can be reduced by a target memory tracking method by fusing inertial navigation information and image information, and the problems of translation invariance, scale invariance and rotation invariance are solved; filtering out the interference of similar targets, and processing target tracking under the conditions of short-time visual field emergence, shielding and the like of target nonlinear maneuvering; the method adopts a hill climbing search strategy and utilizes the improved mutual information entropy as a similarity measurement function, has the advantages of small calculated amount and strong anti-interference capability, and can improve the target tracking robustness.
Description
Technical Field
The invention belongs to the field of image processing and machine vision, relates to a target tracking method of an air-ground full-strapdown image seeker, in particular to a real-time tracking method for attacking a ground static target by a direct-aiming template, is suitable for an air-ground guided weapon with a full-strapdown image seeker and an inertia measuring device and a pod, and is also suitable for all imaging guided weapons according to the principle.
Background
The instantaneous optical view field of the full strapdown image seeker is large, so that the image signal noise and the image processing calculation amount are large; meanwhile, the photoelectric detector is fixedly connected with the reference seat, the disturbance of the carrier is large, and the phenomena of field of view, geometric deformation, shielding and the like of nonlinear maneuvering of the target are easy to occur in the tracking process; the direct aiming template is generated by directly pointing a target by scouting equipment such as a pod distributed aperture photoelectric system, and the like, and has larger difference of visual angle, size, wave band and background with a real-time image acquired by a full strapdown image seeker. The objective factors cause that the traditional target tracking method based on the direct-view template is difficult to adapt to the full strapdown image seeker; the traditional template matching tracking method only has translation invariance and does not have rotation invariance and scale invariance; meanwhile, similarity measurement (absolute difference, product correlation and the like) has the defects of poor geometric distortion resistance and large calculation amount on a search strategy, and the application range of the similarity measurement is influenced.
Disclosure of Invention
Aiming at the problems, the invention aims to provide a full strapdown image seeker target tracking method based on a direct-view template, which can solve the problem of real-time tracking of a ground static target under the conditions of visual angle difference, scale difference, large maneuverability, field of view, shielding and the like between the direct-view template and a target image acquired by a full strapdown image seeker, can realize that the full strapdown image seeker can quickly and accurately give target position information, and provides support for realizing accurate target striking.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
a full strapdown image seeker target tracking method based on a direct-view template is realized by adopting the following steps:
(1) loading a direct-view template 1 and corresponding longitude and latitude height and posture data in a full strapdown image seeker, and carrying out image correction on the full strapdown image seeker by utilizing the loaded direct-view template 1 and the corresponding longitude and latitude height and posture data to obtain a target template 2 for tracking;
(2) the target template 2 is matched with a target real-time image of an adjacent target and an area acquired by the full strapdown image seeker, a hill-climbing search strategy and normalized gray scale mutual information entropy similarity measurement are adopted during matching, the matching is successful, the initial position of the target in the image acquired by the full strapdown image seeker is obtained, and the target real-time image is continuously issued for matching if the matching is unsuccessful;
taking the initial position of the full-strapdown image seeker obtained in the image as a center, selecting a certain area in a target real-time image as a target template 2 for tracking, matching the target template 2 for tracking with the target real-time image by combining inertial navigation information, and adopting a hill-climbing search strategy and normalized gray scale mutual information entropy similarity measurement during matching; if the matching correlation value is larger than the threshold value 1, the matching is considered to be successful, and if the matching correlation value is not successful, the matching is continued;
(3) and if the correlation value is greater than the threshold value 2, the template changing condition is met, and a target template 3 is obtained.
Further, the target template 2 for tracking is obtained after image correction is carried out by using the direct-view template 1 in the step (1), the direct-view template 1 and position and inertial navigation data of corresponding time are obtained through pod equipment, and the target template 2 which is consistent with the view angle and the scale of a target real-time image when the full-strapdown image seeker is used is converted; the transformation process requires that the position (lambda) of the direct targeting template 1 at the moment is known 0 ,L 0 ,h 0 ) And postureNacelle focal length f 0 Focal length f of full strapdown image seeker 1 And target position data (lambda) t ,L t ,h t ) Correcting the position (lambda) of the full strapdown seeker at the moment 1 ,L 1 ,h 1 ) And postureFrom high and low angle-theta ', correcting to high and low angle-theta', the transformation is as follows:
the same principle is that:
in the formula, θ 'and R are the altitude angle and the distance to the target at the moment of acquiring the direct-view template 1, θ ", R' is the altitude angle and the distance to the target at the moment of correcting the target template 2, and u, v, u ', v' are the target image coordinates at the moment of acquiring the direct-view template 1 and the target template 2 image coordinates at the correction moment; y is the length of the target PP ', Delta theta ' is the imaging opening angle of the target PP ' in the seeker image coordinate system at the moment of the direct collimation template 1, and Delta theta ' is the imaging opening angle of the target PP ' in the seeker image coordinate system at the correction moment.
Through the formulas (5) and (6), the target template 2 consistent with the visual angle and the scale of the target real-time image can be obtained through correction by using the direct aiming template 1.
Further, the target template 2 in the step (2) is matched with a target real-time image of an adjacent target and an area acquired by the full strapdown image seeker, the target real-time image is obtained by predicting the target by using inertial navigation information, the position information (λ, L, h) of the full strapdown image seeker at the current moment is obtained, and the coordinates of the current moment position and the coordinates of the target position in the geocentric system are firstly calculated:
in the formula, a e ,b e Is the earth ellipsoid radius; r wt Is the curvature radius of the mortise-unitary ring at the time and the position R N The radius of curvature of the meridian at the current time position; r wt1 The curvature radius of the prime position fourth prime circle is taken as the curvature radius of the prime position fourth prime circle; r N1 The radius of curvature of the meridian at the target position; (lambda t ,L t ,h t ) The longitude and latitude height data of the target position.
Subtracting the two to obtain a relative position vector delta E:
△E=E t -E (13)
and then, carrying out coordinate transformation on the position vector according to a geocentric system → a geographic system → a bomb system → a camera system → an image system, and obtaining the imaging coordinate (uu, vv) of the target in the current full-strapdown image seeker state according to a camera pinhole imaging model and the coordinate transformation relation between the geocentric system and the image coordinate system:
in the formula (I), the compound is shown in the specification,the earth center is a cosine matrix to the geographic system,cosine matrix of geographical to missile system, (u) 0 ,v 0 ) The center of the image of the full strapdown image seeker; f. of 1 Is the seeker focal length; Δ E (1), Δ E (2), and Δ E (3) are projection components of formula (13) in three coordinate axis directions under the geocentric system coordinates.
By selecting the target real-time graph through the method, the target can be quickly matched, the false target can be effectively avoided, and the problem of short-time visual field of the target nonlinear maneuvering can be solved.
Further, matching the target template 2 for tracking with the target real-time image by combining the inertial navigation information in the step (2), namely determining a scale and a rotation parameter in the tracking process; selecting a certain area in a target real-time image as a target template 2 for tracking by taking an initial position in the image acquired by a full strapdown image seeker as a center, and correcting the target template 2 for tracking through inertial navigation information; because the distance between the full strapdown image seeker and the target is far larger than the depth of field of the target, the template correction scale factor k is as follows:
in the formula (di) 1 For matching the distance dis between the seeker and the target in the current frame full strapdown image 0 The distance theta between the seeker and the target in the last full strapdown image frame during matching 1 And theta 2 The opening angles of a previous frame and a next frame of a target in the full strapdown image seeker are regarded as equal;
during tracking, the rotation angle garmr is corrected for the template:
garmar=garmar1-garmar0 (16)
in the formula, garmr 1 is the leader roll angle for matching the full strapdown image of the current frame, and garmr 0 is the leader roll angle for the full strapdown image of the previous frame when matching.
Further, the normalized gray scale mutual information entropy matching method of the hill-climbing search strategy in the step (2) is to adjust the hill-climbing sequence according to the priority and to reduce the calculation of the correlation by using the sample; the method comprises the following steps of realizing a search process from coarse to fine by setting initial hill climbing points at equal intervals under the condition of ensuring the precision, wherein the specific process comprises the following steps:
assuming that the size of the real-time image S is N × N (pixels) and the size of the template image T is M × M (pixels), the metric function of the improved normalized grayscale mutual information entropy matching method is expressed as:
in the formula (I), the compound is shown in the specification,the edge entropy of the real-time subgraph S is obtained, and ps (i, j) is the probability of the real-time subgraph gray level;the edge entropy of the template graph, and pt (i, j) is the probability of the gray level of the template graph;the real-time subgraph and the template graph are combined entropy, pst (i, j) is the combined probability of the real-time subgraph and the template graph, and PM is the weighted value of the gradient amplitude and the main direction of the real-time graph and the template graph; NMI (u, v) as a real-time subgraph S u,v The matching metric value is more than or equal to 1 and less than or equal to 2 in NMI (u, v), 0 and less than or equal to u, and v and less than or equal to N-M, the size of the matching metric value reflects the similarity degree between the template graph and the real-time subgraph, and the larger the value is, the higher the similarity degree is;
if the correlation value of the real-time graph and the template graph is larger than the threshold value 1, the matching is considered to be successful, otherwise, the matching is failed.
Further, in the step (3), when the correlation value is greater than the threshold 2, the template needs to be updated for the next frame matching, and a complete updating mode is adopted during updating, namely, an image with the best matching position of the current frame as the center and a certain area is used as the template 3 to be matched with the image of the next frame;
T new =T(NMI max ) (18)
in the formula, T (NMI) max ) Matching a template picture with the maximum position of the mutual information entropy as a certain area of the center for the current frame, and then repeatedly entering a cycle of template matching, calculating the gray-scale mutual information entropy, obtaining the current best matching position, updating the target template and matching the template to realize the stable tracking of the target.
Compared with the prior art, the invention has the advantages that:
1. the method uses inertial navigation information to correct the direct-view template, so that the direct-view template is consistent with a real-time image acquired by a full strapdown image seeker in view angle and scale, and the method has the advantage of high tracking precision.
2. By integrating the inertial navigation information and the image information, the target memory tracking method can reduce the target tracking search range, has the advantages of translation invariance, scale invariance and rotation invariance, filters out the interference of similar targets, and can process the target tracking under the conditions of short-time out of view field, shielding and the like of the target nonlinear maneuvering.
3. The method adopts a hill climbing search strategy and utilizes the improved mutual information entropy as a similarity measurement function, has the advantages of small calculated amount and strong anti-interference capability, and can improve the target tracking robustness.
Drawings
FIG. 1 is a flow chart of the method of the present invention;
FIG. 2 is a schematic diagram of the template calibration principle of the present invention;
FIG. 3 is a schematic perspective projection diagram of the present invention;
FIG. 4 is a schematic diagram of an improved hill-climbing search strategy according to the present invention;
FIG. 5 is a schematic diagram of the improved normalized gray scale mutual information entropy matching principle of the present invention;
FIG. 6 is a schematic diagram of the principle of the scale correction factor of the present invention.
Detailed Description
The following describes in detail a specific embodiment of the present invention with reference to the drawings.
As shown in fig. 1, a full strapdown image seeker target tracking method based on a direct-view template includes the following specific steps:
(1) the full strapdown image seeker carries out image correction by utilizing the loaded direct-view template 1 and the corresponding longitude and latitude height and posture data to obtain a target template 2 for tracking.
(2) The target template 2 is matched with a target real-time image of an adjacent target and an area acquired by the full strapdown image seeker, a hill-climbing search strategy and normalized gray scale mutual information entropy similarity measurement are adopted during matching, the matching is successful, the initial position of the target in the image acquired by the full strapdown image seeker is obtained, and the target real-time image is continuously issued for matching if the matching is unsuccessful;
taking the initial position of the full-strapdown image seeker obtained in the image as a center, selecting a certain area in a target real-time image as a target template 2 for tracking, matching the target template 2 for tracking with the target real-time image by combining inertial navigation information, and adopting a hill-climbing search strategy and normalized gray scale mutual information entropy similarity measurement during matching; if the matching correlation value is larger than the threshold value 1, the matching is considered to be successful, and if the matching correlation value is not successful, the matching is continued.
(3) And if the correlation value is greater than the threshold value 2, the template changing condition is met, and a target template 3 is obtained.
In the step (1): the direct-aiming template 1 is combined with inertial navigation information to obtain a target template 2 through image correction, the target template 1 and the position and inertial navigation data at the corresponding moment are obtained through pod equipment, and the target template 2 is converted to be consistent with the view angle and the scale of a real-time image when the full-strapdown image seeker is used.
The transformation process requires that the position (lambda) of the direct-view template 1 at the moment is known 0 ,L 0 ,h 0 ) And postureNacelle focal length f 0 All areFocal length f of strapdown image seeker 1 And target position data (lambda) t ,L t ,h t ) Correcting the position (lambda) of the full strapdown seeker at the moment 1 ,L 1 ,h 1 ) And postureFrom the high and low angle-theta', correcting to the high and low angle-theta ", the schematic diagram of the transformation process is shown in fig. 2;
and as can be seen from fig. 2:
the same principle is that:
in the formula, θ 'and R are the altitude angle and the distance to the target at the moment of acquiring the direct-view template 1, θ ", R' is the altitude angle and the distance to the target at the moment of correcting the target template 2, and u, v, u ', v' are the target image coordinates at the moment of acquiring the direct-view template 1 and the target template 2 image coordinates at the correction moment; y is the length of the target PP ', Delta theta ' is the imaging opening angle of the target PP ' in the seeker image coordinate system at the moment of the direct collimation template 1, and Delta theta ' is the imaging opening angle of the target PP ' in the seeker image coordinate system at the correction moment.
By the formulas (5) and (6), the target template 2 consistent with the view angle and the scale of the real-time image can be obtained by correcting the direct-view template map 1.
In the step (2), the target template 2 is matched with a real-time image (hereinafter referred to as a real-time image) of an adjacent target and an area, which are acquired by the full strapdown image seeker, and the target is predicted by using inertial navigation information to obtain the real-time image.
The position information (lambda, L, h) of the whole strapdown image seeker at the current moment is obtained by firstly calculating the coordinates of the current moment position and the target position in the geocentric system:
in the formula, a e ,b e Is the earth ellipsoid radius; r wt Is the curvature radius of the mortise-unitary ring at the time and the position R N The radius of curvature of the meridian at the current time position; r wt1 The curvature radius of the prime position fourth prime circle is taken as the curvature radius of the prime position fourth prime circle;R N1 the radius of curvature of the meridian at the target position; (lambda t ,L t ,h t ) The longitude and latitude height data of the target position.
Subtracting the two to obtain a relative position vector delta E:
△E=E t -E (13)
and then, performing coordinate transformation on the position vector according to a geocentric system → a geographic system → a missile system → a camera system → an image system, and obtaining imaging coordinates (uu, vv) of the target in the current missile state according to a camera pinhole imaging model and a coordinate transformation relation between the geocentric system and the image coordinate system as shown in fig. 3:
in the formula (I), the compound is shown in the specification,the center of the earth is the cosine matrix of the geographic system,cosine matrix of geographical to missile system, (u) 0 ,v 0 ) The center of the image of the full strapdown image seeker; f. of 1 Is the seeker focal length; Δ E (1), Δ E (2), and Δ E (3) are projection components of formula (13) in three coordinate axis directions under the coordinates of the geocentric system.
(13) The target position is predicted through inertial navigation information, and a real-time image is selected, so that the target can be quickly matched, a false target can be effectively avoided, and the problem of short-time visual field of nonlinear maneuvering of the target can be solved.
And (2) obtaining a real-time image from the target template 2 which is obtained in the step (1) and is consistent with the view angle and the scale of the full strapdown image seeker, and then obtaining the position of the target by adopting a normalized gray scale mutual information entropy matching method based on a hill-climbing search strategy for the target template 2 and the real-time image.
In the invention, the determination of scale and rotation parameters in the tracking process is to match the real-time image obtained in the step (2) with the target template 2, correct the target template 2 and the real-time image by using inertial navigation information, and obtain the target position by using a hill-climbing search strategy and normalized mutual information entropy matching.
The target template 2 is corrected through inertial navigation information, and the normalized mutual information entropy matching method has the advantages of scale invariance and rotation invariance. In the tracking process, because the distance between the full strapdown image seeker and the target is far larger than the depth of field of the target, the template correction scale factor k is as follows:
in the formula (di) 1 For matching the distance dis between the seeker and the target in the current frame full strapdown image 0 The distance theta between the seeker and the target in the last full strapdown image frame during matching 1 And theta 2 For the opening angle of the previous frame and the next frame of the target in the full strapdown image seeker, the adjacent frames can be considered equal, and a schematic diagram of the scale invariant correction principle is shown in fig. 6.
During tracking, the rotation angle garmr is corrected for the template:
garmar=garmar1-garmar0 (16)
in the formula, garmr 1 is the rolling angle of the seeker with the full strapdown image of the current frame, and garmr 0 is the rolling angle of the seeker with the full strapdown image of the previous frame during matching.
In order to meet the requirements of real-time performance and precision, an improved hill-climbing search strategy and a simplified correlation degree measurement mode are adopted. The hill-climbing search strategy is improved, as shown in fig. 4, mainly by adjusting the hill-climbing order by priority and reducing the calculation of correlation with samples. The adjustment of the hill climbing sequence according to the priority also solves many unnecessary calculation search processes on the premise of ensuring the matching accuracy, such as: a plurality of climbers search the point with the maximum correlation value (gray scale mutual information entropy) and a plurality of climbers search the local maximum value, and certainly, in the searching process, the calculation efficiency can be further improved by arranging a correlation matrix table and a searching position matrix table, and redundant calculation is avoided. The search matching is shown in fig. 5, and the search process from coarse to fine can be realized.
Assuming that the size of the real-time image S is N × N (pixels) and the size of the template image T is M × M (pixels), the metric function of the improved normalized grayscale mutual information entropy matching method is expressed as:
in the formula (I), the compound is shown in the specification,the edge entropy of the real-time sub-graph S, and ps (i, j) is the probability of the real-time sub-graph gray level occurring.The edge entropy of the template map and pt (i, j) is the probability of the occurrence of the grayscale of the template map.For the joint entropy of the real-time subgraph and the template graph, pst (i, j) is the joint probability of the gray scale of the real-time subgraph and the template graph, and PM is the weighted value of the gradient amplitude and the principal direction of the real-time graph and the template graph. NMI (u, v) as a real-time subgraph S u,v The matching metric value is more than or equal to 1 and less than or equal to NMI (u, v) and less than or equal to 2, the matching metric value is more than or equal to 0 and less than or equal to u, and v is more than or equal to N-M, the size of the matching metric value reflects the similarity degree between the template graph and the real-time subgraph, and the larger the value is, the higher the similarity degree is.
If the correlation value of the real-time graph and the template graph is larger than the threshold value 1, the matching is considered to be successful, otherwise, the matching is failed.
In the step (3), if the correlation value is greater than the threshold 2, the template needs to be updated for the next frame matching, and a complete updating mode is adopted during general updating, namely, an image with the best matching position of the current frame as the center and a certain area is used as the template 3 to be matched with the image of the next frame.
T new =T(NMI max ) (18)
In the formula, T (NMI) max ) Matching mutual information entropy for current framesThe maximum position is the template picture of a certain area at the center. And then repeatedly entering a cycle of 'template matching, calculating gray scale mutual information entropy, obtaining the current best matching position, target template updating and template matching' to realize stable target tracking.
Claims (5)
1. A full strapdown image seeker target tracking method based on a direct-view template is characterized by comprising the following specific implementation steps of:
(1) loading a direct-view template 1 and corresponding longitude and latitude height and posture data in a full strapdown image seeker, and carrying out image correction on the full strapdown image seeker by utilizing the loaded direct-view template 1 and the corresponding longitude and latitude height and posture data to obtain a target template 2 for tracking;
(2) the target template 2 is matched with a target real-time image of an adjacent target and an area acquired by the full strapdown image seeker, a hill-climbing search strategy and normalized gray scale mutual information entropy similarity measurement are adopted during matching, the matching is successful, the initial position of the target in the image acquired by the full strapdown image seeker is obtained, and the target real-time image is continuously issued for matching if the matching is unsuccessful;
taking the initial position of the full-strapdown image seeker obtained in the image as a center, selecting a certain area in a target real-time image as a target template 2 for tracking, matching the target template 2 for tracking with the target real-time image by combining inertial navigation information, and adopting a hill-climbing search strategy and normalized gray scale mutual information entropy similarity measurement during matching; if the matching correlation value is larger than the threshold value 1, the matching is considered to be successful, and if the matching correlation value is not successful, the matching is continued;
the normalized gray scale mutual information entropy matching method of the hill climbing search strategy is to adjust the hill climbing sequence according to the priority and reduce the calculation of the correlation by using samples; the method comprises the following steps of realizing a search process from coarse to fine by setting initial hill climbing points at equal intervals under the condition of ensuring the precision, wherein the specific process comprises the following steps:
assuming that the size of the real-time graph S is nxn pixels and the size of the template graph T is mxm pixels, the metric function of the improved normalized gray scale mutual information entropy matching method is expressed as:
in the formula (I), the compound is shown in the specification,the edge entropy of the real-time subgraph S is shown, and ps (i, j) is the probability of the real-time subgraph gray level;the edge entropy of the template graph, and pt (i, j) is the probability of the gray level of the template graph;the real-time subgraph and the template graph are combined entropy, pst (i, j) is the combined probability of the real-time subgraph and the template graph, and PM is the weighted value of the gradient amplitude and the main direction of the real-time graph and the template graph; NMI (u, v) as a real-time subgraph S u,v The matching metric value is more than or equal to 1 and less than or equal to 2 in NMI (u, v), 0 and less than or equal to u, and v and less than or equal to N-M, the size of the matching metric value reflects the similarity degree between the template graph and the real-time subgraph, and the larger the value is, the higher the similarity degree is;
if the correlation value of the real-time graph and the template graph is greater than the threshold value 1, the matching is considered to be successful, otherwise, the matching is failed;
(3) and if the correlation value is greater than the threshold value 2, the template changing condition is met, and a target template 3 is obtained.
2. The method for tracking the target of the full strapdown image seeker based on the direct-view template as claimed in claim 1, wherein the step (1) of obtaining the target template 2 for tracking after image correction by using the direct-view template 1 is to obtain the direct-view template 1 and position and inertial navigation data of corresponding time through pod equipment, and to convert the position and inertial navigation data into the target template 2 which is consistent with the view angle and the scale of a target real-time image when the full strapdown image seeker is used; the transformation process requires that the position (lambda) of the direct-view template 1 at the moment is known 0 ,L 0 ,h 0 ) And postureNacelle focal length f 0 Focal length f of full strapdown image seeker 1 And target position data (lambda) t ,L t ,h t ) Correcting the position (lambda) of the full strapdown seeker at the moment 1 ,L 1 ,h 1 ) And postureAnd correcting to the high-low angle-theta 'from the high-low angle-theta', and converting as follows:
the same principle is that:
in the formula, θ ', R is a height angle at the moment of acquiring the direct-view template 1 and a distance from a target, θ ", R ' is a height angle at the moment of correcting the target template 2 and a distance from the target, u, v, u ', v ' is a target image coordinate at the moment of acquiring the direct-view template 1 and a target template 2 image coordinate at the moment of correcting, y is a length of a target PP", Δ θ ' is an imaging opening angle of the target PP "at the moment of the direct-view template 1 in a seeker image coordinate system, and Δ θ" is an imaging opening angle of the target PP "at the moment of correcting in the seeker image coordinate system;
by the formulas (5) and (6), the target template 2 consistent with the visual angle and the scale of the target real-time image can be obtained by correcting by using the direct-view template 1.
3. The method for tracking the target of the full strapdown image seeker based on the direct-view template as claimed in claim 1, wherein the step (2) of matching the target template 2 with the real-time target map of the nearby target and the area obtained by the full strapdown image seeker is to predict the target by using inertial navigation information to obtain the real-time target map, and the position information (λ, L, h) of the full strapdown image seeker at the current time, and firstly calculate the coordinates of the current time position and the target position under the geocentric system:
in the formula, a e ,b e Is the earth ellipsoid radius; r wt Is the curvature radius of the mortise-unitary ring at the time and the position R N The radius of curvature of the meridian at the current time position; r wt1 The curvature radius of the prime position fourth prime circle is taken as the curvature radius of the prime position fourth prime circle; r N1 The radius of curvature of the meridian at the target position; (lambda t ,L t ,h t ) Longitude and latitude height data of a target position;
subtracting the two to obtain a relative position vector delta E:
△E=E t -E (13)
and then, carrying out coordinate transformation on the position vector according to a geocentric system → a geographic system → a bomb system → a camera system → an image system, and obtaining the imaging coordinate (uu, vv) of the target in the current full-strapdown image seeker state according to a camera pinhole imaging model and the coordinate transformation relation between the geocentric system and the image coordinate system:
in the formula (I), the compound is shown in the specification,the center of the earth is the cosine matrix of the geographic system,cosine matrix of geographical to missile system, (u) 0 ,v 0 ) The center of the image of the full strapdown image seeker; f. of 1 Is the seeker focal length; Δ E (1), Δ E (2), and Δ E (3) are projection components of formula (13) in three coordinate axis directions under the geocentric system coordinates.
4. The method for tracking the target of the full strapdown image seeker based on the direct-view template as claimed in claim 1, wherein the step (2) of matching the target template 2 for tracking with the target real-time image by combining the inertial navigation information is to determine the scale and rotation parameters in the tracking process; selecting a certain area in a target real-time image as a target template 2 for tracking by taking an initial position in the image acquired by a full strapdown image seeker as a center, and correcting the target template 2 for tracking through inertial navigation information; in the tracking process, because the distance between the full strapdown image seeker and the target is far larger than the depth of field of the target, the template correction scale factor k is as follows:
in the formula (di) 1 For matching the distance dis between the seeker and the target in the current frame full strapdown image 0 The distance theta between the seeker and the target in the last full strapdown image frame during matching 1 And theta 2 The opening angles of a previous frame and a next frame of a target in the full strapdown image seeker are regarded as equal;
during tracking, the rotation angle garmr is corrected for the template:
garmar=garmar1-garmar0 (16)
in the formula, garmr 1 is the leader roll angle for matching the full strapdown image of the current frame, and garmr 0 is the leader roll angle for the full strapdown image of the previous frame when matching.
5. The method for tracking the target of the full strapdown image seeker based on the direct-view template as claimed in claim 1, wherein in the step (3), when the correlation value is greater than the threshold 2, the template needs to be updated for the next frame matching, and a completely updated mode is adopted during updating, namely, an image with a certain area with the best matching position of the current frame as the center is used as the template 3 to be matched with the image of the next frame;
T new =T(NMI max ) (18)
in the formula, T (NMI) max ) Matching a template picture with the maximum position of mutual information entropy as a certain area of the center for the current frame, and then repeatedly entering' template matching-countingAnd (3) calculating the gray level mutual information entropy, obtaining the current optimal matching position, updating the target template and matching the template to realize stable target tracking.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011508984.6A CN112489091B (en) | 2020-12-18 | 2020-12-18 | Full strapdown image seeker target tracking method based on direct-aiming template |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011508984.6A CN112489091B (en) | 2020-12-18 | 2020-12-18 | Full strapdown image seeker target tracking method based on direct-aiming template |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112489091A CN112489091A (en) | 2021-03-12 |
CN112489091B true CN112489091B (en) | 2022-08-12 |
Family
ID=74914838
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011508984.6A Active CN112489091B (en) | 2020-12-18 | 2020-12-18 | Full strapdown image seeker target tracking method based on direct-aiming template |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112489091B (en) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113395448B (en) * | 2021-06-15 | 2023-02-21 | 西安视成航空科技有限公司 | Airborne pod image searching, tracking and processing system |
CN113554131B (en) * | 2021-09-22 | 2021-12-03 | 四川大学华西医院 | Medical image processing and analyzing method, computer device, system and storage medium |
CN114280978B (en) * | 2021-11-29 | 2024-03-15 | 中国航空工业集团公司洛阳电光设备研究所 | Tracking decoupling control method for photoelectric pod |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108717712A (en) * | 2018-05-29 | 2018-10-30 | 东北大学 | A kind of vision inertial navigation SLAM methods assumed based on ground level |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5146228A (en) * | 1990-01-24 | 1992-09-08 | The Johns Hopkins University | Coherent correlation addition for increasing match information in scene matching navigation systems |
EP3248029A4 (en) * | 2015-01-19 | 2018-10-03 | The Regents of the University of Michigan | Visual localization within lidar maps |
CN107945215B (en) * | 2017-12-14 | 2021-07-23 | 湖南华南光电(集团)有限责任公司 | High-precision infrared image tracker and target rapid tracking method |
-
2020
- 2020-12-18 CN CN202011508984.6A patent/CN112489091B/en active Active
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108717712A (en) * | 2018-05-29 | 2018-10-30 | 东北大学 | A kind of vision inertial navigation SLAM methods assumed based on ground level |
Also Published As
Publication number | Publication date |
---|---|
CN112489091A (en) | 2021-03-12 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112489091B (en) | Full strapdown image seeker target tracking method based on direct-aiming template | |
CN108534782B (en) | Binocular vision system-based landmark map vehicle instant positioning method | |
CN101246590B (en) | Star loaded camera spacing deformation image geometric correction method | |
CN105222788B (en) | The automatic correcting method of the matched aircraft Route Offset error of feature based | |
CN103017762B (en) | The extraterrestrial target fast Acquisition localization method of ground photo-electric telescope | |
CN103047985A (en) | Rapid positioning method for space target | |
CN110111274B (en) | Method for calibrating exterior orientation elements of satellite-borne push-broom optical sensor | |
CN109387192B (en) | Indoor and outdoor continuous positioning method and device | |
US9453731B2 (en) | System and method for determining orientation relative to earth | |
CN109540113B (en) | Total station and star map identification method thereof | |
CN111238540A (en) | Lopa gamma first camera-satellite sensitive installation calibration method based on fixed star shooting | |
CN111366148A (en) | Target positioning method suitable for multiple observations of airborne photoelectric observing and sighting system | |
CN108364279A (en) | Determine the method that stationary orbit remote sensing satellite is directed toward deviation | |
CN108362268A (en) | A kind of automatic astronomical surveing method and measuring system based on video measuring | |
CN112857356A (en) | Unmanned aerial vehicle water body environment investigation and air route generation method | |
CN110889353B (en) | Space target identification method based on primary focus large-visual-field photoelectric telescope | |
CN113218577A (en) | Outfield measurement method for star point centroid position precision of star sensor | |
CN109883400B (en) | Automatic target detection and space positioning method for fixed station based on YOLO-SITCOL | |
CN114860196A (en) | Telescope main light path guide star device and calculation method of guide star offset | |
WO2021135161A1 (en) | Real-time celestial positioning and metering method for space debris based on automatic pointing measurement | |
CN110887477B (en) | Autonomous positioning method based on north polarization pole and polarized sun vector | |
CN115950435A (en) | Real-time positioning method for unmanned aerial vehicle inspection image | |
CN115824190A (en) | Vision and GPS-based target ship fusion positioning method | |
US11580690B1 (en) | Horizon-based navigation | |
CN113240749B (en) | Remote binocular calibration and ranging method for recovery of unmanned aerial vehicle facing offshore ship platform |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |