CN113029108A - Automatic relative orientation method and system based on sequence sea surface images - Google Patents
Automatic relative orientation method and system based on sequence sea surface images Download PDFInfo
- Publication number
- CN113029108A CN113029108A CN202110234110.4A CN202110234110A CN113029108A CN 113029108 A CN113029108 A CN 113029108A CN 202110234110 A CN202110234110 A CN 202110234110A CN 113029108 A CN113029108 A CN 113029108A
- Authority
- CN
- China
- Prior art keywords
- image
- relative orientation
- point
- pairs
- sea surface
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C11/00—Photogrammetry or videogrammetry, e.g. stereogrammetry; Photographic surveying
- G01C11/04—Interpretation of pictures
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C13/00—Surveying specially adapted to open water, e.g. sea, lake, river or canal
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Life Sciences & Earth Sciences (AREA)
- Hydrology & Water Resources (AREA)
- Multimedia (AREA)
- Image Processing (AREA)
- Length Measuring Devices By Optical Means (AREA)
- Image Analysis (AREA)
Abstract
The invention belongs to the technical field of photogrammetry, and discloses an automatic relative orientation method and system based on sequence sea surface images, wherein the automatic relative orientation method based on the sequence sea surface images comprises the following steps: reading a group of sea surface image sequences of not less than 50 pairs; carrying out distortion calibration on the image sequence; selecting a first image pair of the image sequence, and determining the maximum parallax in the selected matching area; extracting characteristic point pairs on the left image, and removing the difference to obtain n more than or equal to 5 residual characteristic point pairs; determining a control equation set based on the first image pair; preliminary relative orientation; extracting characteristic point pairs of the sequence image pairs and rejecting gross errors; the precise relative orientation. In order to increase the practicability, the invention provides a semi-automatic gross error point eliminating method of a first image pair, a gross error point and error larger point eliminating method of an image sequence and an initial value obtaining method. Practice shows that the invention can realize stable and rapid relative orientation of the stereo camera and meet the requirements of stereo photography wave measurement.
Description
Technical Field
The invention belongs to the technical field of photogrammetry, and particularly relates to an automatic relative orientation method and system based on sequence sea surface images.
Background
At present, in a strict sense, the spatial synchronous observation of the wave surface height field and the change thereof is the only effective method for accurately acquiring the wave direction-frequency spectrum. As the only technology capable of realizing the space synchronous observation of the wave height field at present, the stereo photography wave observation technology is rapidly developed in the last decade. The stereo image pair relative orientation is to recover the mutual relation between two adjacent image photographing light beams at the photographing time, so that the same-name light ray pairs are intersected, the key step of the stereo photographing sea wave measurement technology is realized, the premise of realizing the three-dimensional reconstruction of sea waves is realized, and the stable and accurate relative orientation technology has typical practical significance.
As a fundamental problem of photogrammetry and a key step of three-dimensional reconstruction, the relative orientation technology has been a field of photogrammetry focus research. Various relative orientation methods, such as a five-point algorithm, an eight-point algorithm and the like, and various specific equation convergence algorithms derived from the methods are proposed, so that the relative orientation problem is solved well for most application scenarios.
Compared with land images, sea surface image texture information is less, specular reflection is contained, a large amount of rough differences exist in image matching, even if the images are correctly matched with the same name points, experiments show that the matching precision of the images is lower than that of the images in land close-range survey by one magnitude, and the realization of the relative orientation of the stereo camera based on the sea surface images is used for solving the problem of the existing stereo photography and ocean wave survey. The present invention has previously proposed a stereo photographic relative orientation method based on a single sea surface image, however practice has shown that the y rotation angle characterizing the baseline direction may have a difference of 2 ° to 3 ° when relative orientation is performed using different images. This error is relatively large and a large measurement error may come. The reason for this is that, due to the specular reflection image, a single sea surface image has fewer characteristic points with rich texture information, and when the number of homonymous point pairs increases to a certain number, if more characteristic point pairs are artificially extracted, the relative orientation accuracy may be reduced. Therefore, a new method for relative orientation of sea surface images is needed.
Through the above analysis, the problems and defects of the prior art are as follows: according to the existing stereo photography relative orientation method based on a single sea surface image, when different image pairs are used for relative orientation, due to the fact that images are reflected by a mirror surface, texture information of the single sea surface image is rich in a few characteristic points, the y rotation angle representing the baseline direction can have a difference of 2-3 degrees, the error is large, and a large measurement error can be brought; when the number of the same-name point pairs is increased to a certain number, if more feature point pairs are artificially extracted, the relative orientation accuracy may be reduced.
The difficulty in solving the above problems and defects is: increasing the number of effective pairs of feature points by increasing the number of pairs of sea images is one way to solve the above-mentioned problem; however, it is currently difficult to quickly remove the coarse difference points and the points with large matching errors from a large number of pairs of feature points and retain the valid pairs of feature points to obtain accurate relative orientation parameters.
The significance of solving the problems and the defects is as follows: aiming at the problems, the invention provides a semi-automatic gross error point removing method for a primary image pair, a gross error point and error larger point removing method for an image sequence, and an initial value obtaining method; the relative orientation of the stereo camera can be stably and quickly realized, the time and the cost can be saved, and the popularization of the photography sea wave measurement technology is facilitated.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides an automatic relative orientation method and system based on sequence sea surface images, and particularly relates to a novel stereo photography relative orientation calibration method and system aiming at sea wave measurement.
The invention is realized in such a way that an automatic relative orientation method based on sequence sea surface images comprises the following steps:
reading a group of sea surface image sequences of not less than 50 pairs;
secondly, distortion calibration is carried out on the image sequence;
selecting a first image pair of the image sequence, and determining the maximum parallax in the selected matching area;
extracting characteristic point pairs on the left image, and removing gross errors to obtain n more than or equal to 5 residual characteristic point pairs;
step five, determining a control equation set 1 based on the first image pair;
step six, preliminary relative orientation;
extracting characteristic point pairs of the sequence image pairs and removing gross errors;
and step eight, accurate relative orientation.
Further, in the step one, reading a group of not less than 50 pairs of sea surface image sequences includes:
the method comprises the steps of installing a double camera on an observation platform, increasing the area of an image pair overlapping area, collecting a row of sea surface image sequences of not less than 50 pairs, adopting a tiff picture format and reading pictures by using an imread command.
Further, in the second step, the performing distortion calibration on the image sequence includes:
and calibrating each image in the image sequence according to the orientation parameters and the distortion coefficients in the camera, wherein the calibrated images conform to the pinhole imaging model.
Further, in the third step, the selecting a first image pair of the image sequence, and determining the maximum parallax in the selected matching area includes:
selecting a first image pair of an image sequence, displaying the first image pair through 'image', artificially and approximately selecting a homonymous point pair near each vertex in a quadrilateral overlapping region, and determining a maximum matching region, a maximum row parallax and a maximum column parallax of a rectangle through the four homonymous point pairs.
Further, in the fourth step, the extracting the feature point pairs on the left image and eliminating the discrepancy to obtain n more than or equal to 5 remaining feature point pairs includes:
extracting characteristic point pairs of the first image pair and rejecting gross errors:
extracting 10 multiplied by 10 characteristic points from the left image, matching to obtain homologous points of the characteristic points, wherein the characteristic points and the homologous points are called characteristic point pairs, and rejecting the characteristic point pairs with gross errors, and the method comprises the following steps:
and uniformly dividing the left image overlapping area into a plurality of areas, determining a characteristic point in each area, extracting the characteristic point by using a Moravac operator or other operators, and obtaining the homonymous point of the right image through image matching.
And the precise matching of the feature points is realized by adopting a pyramid and least square matching method. The method comprises the steps of averagely generating a 3-layer pyramid image by taking 3 x 3 pixels as one pixel, sequentially calling the image from bottom to top as a first layer image, a second layer image and a third layer image, sequentially extracting feature points on a third layer image, a second layer image and a first layer image pair by adopting a discrimination method with the maximum correlation coefficient and matching the feature points with homologous points, wherein a search window is a rectangular area determined by the maximum parallax of left and right image rows and columns. And after pyramid matching is completed, accurate matching of the images is realized by adopting a least square image matching method.
If geometric distortion and radiation distortion are considered, the homonymy point should satisfy the surrounding gray function:
wherein h is0 h1Representing a radiation distortion parameter, a0 a1 a2 b0 b1 b2Representing a geometric distortion parameter; the radiation parameters and the geometric parameters can be solved by adopting an indirect adjustment method and utilizing a maximum correlation coefficient discrimination method, so that the homonymous point of one image point is accurately determined; the maximum correlation coefficient judging method is that when the correlation coefficient after iteration is smaller than the correlation coefficient after the last iteration, the iteration is stopped.
Further, the semi-automatic gross error judging and eliminating method comprises the following steps:
(1) and searching a companion point for each feature point, and removing gross errors according to the parallax errors.
For each feature point, a point with the most texture information, namely an accompanying point, is searched within the range of r being more than or equal to 5 and less than or equal to 7, and the method is adopted to match the point with the same name. Since the sea surface has no cliff type fluctuation, the parallax difference between the feature point and the accompanying point is not very large and should be within 5 pixels. According to the judgment, the part gross errors can be automatically removed.
(2) And eliminating gross errors according to the parallax change of the feature points on the same line.
Because the feature points of the left image are approximately uniformly distributed and the sea surface fluctuation is small relative to the photographic height, the parallax of the feature points on the same row does not change violently, and the discrepancy can be eliminated manually according to the criterion.
Further, in step five, the determining of the control equation set 1 based on the first image pair includes:
like other photogrammetry, the relative directional governing equation in the ocean wave measurement is also a coplanar equation. The main optical axis of the left camera is taken as the z axis, the x axis and the y axis are respectively parallel to the row direction and the column direction of the CCD to establish a coordinate system o-xyz, which is called as a left camera coordinate system hereinafter, and a right camera coordinate system o '-x' y 'z' is established in the same way. One point A (x) on the object to be measuredA,yA,zA) With a picture point a (x) on each of the left and right camerasa,ya,-f)、a'(x'a',y'a'-f '), then the straight lines oa, o' a 'and oo' are coplanar, i.e.:
let o-xyz rotate a continuously around the y, x axesy、αxAfter angle, z-axis and base line
Overlapping; then, a left camera coordinate system o-xyz is set to rotate beta around the y, x and z axes continuouslyy、βx、βzAfter the angle, the coordinate axes are parallel to the coordinate axes corresponding to the coordinate system o '-x' y 'z' of the right camera; the relative orientation is the solution parameter αy、αx、βy、βx、βzThe value of (c). Thus the baseline in the left camera coordinate system can be expressed as:
wherein the content of the first and second substances,
the unit vector of x-axis, y-axis and z-axis directions is shown, and D represents the length of the base line, which needs to be actually measured. Thus, formula (2) may represent an in-line form:
wherein (xi)a',ηa',ζa') Satisfy the relation:
wherein R is0Represents a transformation matrix between the coordinate systems o-xyz and o '-x' y 'z':
therefore, the formula (4) contains only alphay、αx、βy、βx、βzAnd 5 relative orientation parameters are equal, so that theoretically, the relative orientation parameters can be calculated by only obtaining 5 homonymous point pairs, and the formula (4) is the relative orientation control equation.
Because the nonlinear equation system only contains 5 parameters and the matching precision of the sea surface image feature points is low, practice shows that the initial value provided by the relative orientation direct solution is not ideal sometimes. As the main optical axes of the two cameras are approximately parallel when the stereo photogrammetry is used for measuring the sea waves, the practice shows that the initial values of the relative orientation parameters can be selected as follows:
[αx αy βx βy βz]＝[0 -1.5 0 0 0] (7)。
further, in step six, the preliminary relative orientation includes:
and (4) taking the coordinates of the same-name point pairs as observed quantities, taking 5 relative orientation parameters as adjustment parameters, adopting the initial values obtained in the step five, and solving the relative orientation parameter values by using a conditional adjustment method with parameters, wherein the initial values are called initial relative orientation.
Further, in step seven, the feature point pair extraction and gross error elimination of the sequence image pair includes:
after the characteristic points and the homonymous points of the subsequent image pairs are sequentially determined, n homonymous point pairs are obtained in the first image pair; wherein n is much greater than 5; after an equation set consisting of n control equations is obtained according to the formula (4), the coordinates of the same-name point pairs are used as observed quantities, 5 relative orientation parameters are used as adjustment parameters, and a conditional adjustment method with parameters is adopted to obtain relative orientation parameter values, namely approximate values.
And calculating the epipolar line y coordinate corresponding to each homonymous point x coordinate, and solving the difference delta y between the epipolar line y coordinate and the homonymous point y coordinate. Theoretically, the point where the delta y is zero is considered as an error matching point and is removed, wherein the larger the absolute value of the delta y is, the larger the matching error is.
Further, in step eight, the precise relative orientation includes:
(1) removing homonymous point pairs with delta y more than or equal to 6 to obtain residual homonymous point pair set A1Using A1Using the first image calibration result as the initial value, recalculating the relative orientation parameter to obtain X1；
(2) By using X1Computing a set A1The epipolar line y coordinate corresponding to the x coordinate of each homonymy point is calculated, and the difference delta y between the epipolar line y coordinate and the y coordinate of the homonymy point is calculated;
(3) culling set A1Obtaining the residual homonymous point pair set A by homonymous point pairs with the middle delta y more than or equal to 52Using A2Recalculating the relative orientation parameters to obtain X2The initial value is X1；
(4) And (3) repeating the step (2) and the step (3), sequentially removing homonymous point pairs of which the delta y is more than or equal to [4321], and finally calculating an accurate and stable relative orientation parameter value.
Another object of the present invention is to provide an automated relative orientation system based on sequential sea surface images, which applies the automated relative orientation method based on sequential sea surface images, the automated relative orientation system based on sequential sea surface images comprising:
the image sequence reading module is used for installing the double cameras on the observation platform, increasing the area of an image pair overlapping area, acquiring a row of sea surface image sequences of not less than 50 pairs, reading the images by using an imread command by adopting a tiff image format;
the distortion calibration module is used for calibrating each image in the image sequence according to the camera internal orientation parameters and the distortion coefficients, and the calibrated images conform to the pinhole imaging model;
the maximum parallax determining module is used for selecting a first image pair of the image sequence and determining the maximum parallax in the selected matching area;
the characteristic point pair extraction module is used for extracting characteristic point pairs on the left image and eliminating the difference to obtain n more than or equal to 5 residual characteristic point pairs;
the control equation set determining module is used for determining a control equation set 1 based on the first image pair;
the preliminary relative orientation module is used for calculating a relative orientation parameter value by using the initial value obtained in the fifth step and a conditional adjustment method with parameters, wherein the initial value is the coordinate of the homonymous point pair as an observed quantity, and 5 relative orientation parameters are adjustment parameters, and the relative orientation parameter value is called preliminary relative orientation;
the gross error elimination module is used for extracting the characteristic point pairs of the sequence image pairs and eliminating gross errors;
and the precise relative orientation module is used for determining a precise solution of the relative orientation parameter.
By combining all the technical schemes, the invention has the advantages and positive effects that: the automatic relative orientation method based on the sequence sea surface images only extracts a small number of characteristic point pairs with most abundant texture information and highest matching precision from one image pair, increases the number of the characteristic point pairs by increasing the number of the image pairs, and finally realizes accurate relative orientation. In order to increase the practicability, the invention provides a semi-automatic gross error point eliminating method of a first image pair, a gross error point and error larger point eliminating method of an image sequence, and an initial value obtaining method. Practice shows that the method can realize stable and quick relative orientation of the stereo camera and meet the requirements of stereo photography wave measurement.
The invention mainly comprises the following innovations: 1) a first image characteristic point pair obtaining and gross error eliminating method; 2) obtaining a characteristic point pair of a sequence image pair and removing gross errors; 3) establishing a relative orientation control equation set and selecting an initial value method; 4) and gradually eliminating feature point pairs with low matching quality in an iterative mode to gradually approach the true value of the parameter to be solved, and finally obtaining the accurate relative orientation parameter value.
Meanwhile, the relative orientation calibration in the three-dimensional photography ocean wave measurement can be quickly and accurately realized only by relying on the sequence sea image pairs, and the method has the greatest advantage that the number of the homonymous point pairs which are accurately matched can be increased without limit by increasing the number of the sequence image pairs, so that stable and accurate relative orientation parameters are obtained, and the relative orientation problem in the photography ocean wave measurement is solved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the embodiments of the present invention will be briefly described below, and it is obvious that the drawings described below are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a flowchart of an automated relative orientation method based on sequential sea surface images according to an embodiment of the present invention.
Fig. 2 is a schematic diagram of an automated relative orientation method based on sequential sea surface images according to an embodiment of the present invention.
FIG. 3 is a block diagram of an automated relative orientation system based on sequential sea surface images according to an embodiment of the present invention;
in the figure: 1. an image sequence reading module; 2. a distortion calibration module; 3. a maximum disparity determination module; 4. a characteristic point pair extraction module; 5. a control equation set determining module; 6. a preliminary relative orientation module; 7. a gross error rejection module; 8. and precisely relatively orienting the modules.
Fig. 4 is a schematic diagram of relative orientation of stereography provided by an embodiment of the present invention.
Fig. 5(a) -5 (b) are parallax change diagrams of 10 × 10 feature points and corresponding points of the first image in the sequence of bosch experiment images according to the embodiment of the present invention.
Fig. 6 is a difference layout diagram of y coordinates of a right image homonymous point and y coordinates of a corresponding epipolar line in a bosch experiment image sequence according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
In view of the problems in the prior art, the present invention provides an automated relative orientation method and system based on sequential sea surface images, which will be described in detail with reference to the accompanying drawings.
As shown in fig. 1, the automated relative orientation method based on sequential sea surface images provided by the embodiment of the present invention includes the following steps:
s101, reading a group of sea surface image sequences of not less than 50 pairs;
s102, distortion calibration is carried out on the image sequence;
s103, selecting a first image pair of the image sequence, and determining the maximum parallax in a selected matching area;
s104, extracting the feature point pairs on the left image, and eliminating the difference to obtain n more than or equal to 5 residual feature point pairs;
s105, determining a control equation set 1 based on the first image pair;
s106, primary relative orientation;
s107, extracting characteristic point pairs of the sequence image pairs and removing gross errors;
and S108, precise relative orientation.
Those skilled in the art can also implement the method of the present invention based on the sequence sea surface image by using other steps, and the method of the present invention based on the sequence sea surface image provided in fig. 1 is only one specific embodiment.
A schematic diagram of an automated relative orientation method based on sequential sea surface images according to an embodiment of the present invention is shown in fig. 2.
As shown in fig. 3, the automated relative orientation system based on sequential sea surface images provided by the embodiment of the present invention includes:
the image sequence reading module 1 is used for installing the double cameras on an observation platform, increasing the area of an image pair overlapping area, acquiring a row of sea surface image sequences of not less than 50 pairs, reading the images by using an imread command by adopting a tiff image format;
the distortion calibration module 2 is used for calibrating each image in the image sequence according to the camera internal orientation parameters and the distortion coefficients, and the calibrated images conform to the pinhole imaging model;
the maximum parallax determining module 3 is used for selecting a first image pair of the image sequence and determining the maximum parallax in the selected matching area;
the characteristic point pair extraction module 4 is used for extracting characteristic point pairs on the left image and eliminating the difference to obtain n more than or equal to 5 residual characteristic point pairs;
the control equation set determining module 5 is used for determining a control equation set 1 based on the first image pair;
a preliminary relative orientation module 6, configured to use the coordinates of the same-name point pairs as observed quantities, use 5 relative orientation parameters as adjustment parameters, and use the initial values obtained in step five to calculate relative orientation parameter values by using a conditional adjustment method accompanied by parameters, which is called preliminary relative orientation;
the gross error elimination module 7 is used for extracting the characteristic point pairs of the sequence image pairs and eliminating gross errors;
and an accurate relative orientation module 8 for determining an accurate solution of the relative orientation parameter.
The technical solution of the present invention is further described with reference to the following examples.
Example 1
1. Overview
The present invention has previously proposed a stereo photographic relative orientation method based on a single sea surface image, however practice has shown that the y rotation angle characterizing the baseline direction may have a difference of 2 ° to 3 ° when relative orientation is performed using different images. This error is relatively large and a large measurement error may come. The reason for this is that, due to the specular reflection image, a single sea surface image has fewer characteristic points with rich texture information, and when the number of homonymous point pairs increases to a certain number, if more characteristic point pairs are artificially extracted, the relative orientation accuracy may be reduced.
In view of this, the invention provides a relative orientation method of photogrammetric sea waves based on sequential sea-surface image pairs; the method is mainly characterized in that only a few characteristic point pairs with most abundant texture information and highest matching precision are extracted from one image pair, and the number of the characteristic point pairs is increased by increasing the number of the image pairs, so that accurate relative orientation is finally realized. In order to increase the practicability, the invention provides a semi-automatic gross error point eliminating method of a first image pair, a gross error point and error larger point eliminating method of an image sequence, and an initial value obtaining method. Practice shows that the method can realize stable and quick relative orientation of the stereo camera and meet the requirements of stereo photography wave measurement.
2. The invention aims to provide a stereo photography relative orientation method for ocean wave measurement, which has the advantages that relative orientation calibration in the stereo photography ocean wave measurement can be quickly and accurately realized only by relying on sequence sea-surface image pairs, and the maximum advantage is that the number of identical point pairs which are accurately matched can be increased without limit by increasing the number of the sequence image pairs, so that stable and accurate relative orientation parameters are obtained, and the problem of relative orientation in the photography ocean wave measurement is solved.
The technical scheme adopted by the invention is as follows:
in the shooting wave measurement, firstly, the calibration of the internal orientation elements and the distortion coefficients of the camera is completed, and then the camera is erected for wave measurement, so that the camera can be considered to be imaged by a small hole imaging model when the relative orientation is carried out.
1. Control equation set and initial value of relative orientation parameter
Like other photogrammetry, the relative directional governing equation in the ocean wave measurement is also a coplanar equation. As shown in fig. 4, a coordinate system o-xyz, hereinafter referred to as a left camera coordinate system, is established with a left camera principal optical axis as a z-axis, and an x-axis and a y-axis parallel to a row direction and a column direction of the CCD, respectively, and a right camera coordinate system o '-x' y 'z' is established similarly. One point A (x) on the object to be measuredA,yA,zA) With a picture point a (x) on each of the left and right camerasa,ya,-f)、a'(x'a',y'a'-f '), then the straight lines oa, o' a 'and oo' are coplanar, i.e.:
let o-xyz rotate a continuously around the y, x axesy、αxAfter angle, z-axis and base line
Overlapping; then, a left camera coordinate system o-xyz is set to rotate beta around the y, x and z axes continuouslyy、βx、βzAfter the angle, the coordinate axes are parallel to the coordinate axes corresponding to the coordinate system o '-x' y 'z' of the right camera; the relative orientation is the solution parameter αy、αx、βy、βx、βzThe value of (c). Thus the baseline in the left camera coordinate system can be expressed as:
wherein the content of the first and second substances,
the unit vector of x-axis, y-axis and z-axis directions is shown, and D represents the length of the base line, which needs to be actually measured. Thus, formula (1) can be expressed in the form of an array:
wherein (xi)a',ηa',ζa') Satisfy the relation:
wherein R is0Represents a transformation matrix between the coordinate systems o-xyz and o '-x' y 'z':
(3) in which only alpha is containedy、αx、βy、βx、βzAnd 5 relative orientation parameters are equal, so theoretically, the relative orientation parameters can be calculated by only obtaining 5 pairs of same-name points. (3) The formula is the relative directional control equation of the present invention.
Because the nonlinear equation system only contains 5 parameters and the matching precision of the sea surface image feature points is low, practice shows that the initial value provided by the relative orientation direct solution is not ideal sometimes. As the main optical axes of the two cameras are approximately parallel when the stereo photogrammetry is used for measuring the sea waves, the practice shows that the initial values of the relative orientation parameters can be selected as follows:
[αx αy βx βy βz]＝[0 -1.5 0 0 0] (5)
2. characteristic point pair extraction and gross error elimination of first image pair
This section is one of the core innovations of the present invention. The left image overlapping area is uniformly divided into a plurality of areas, for example, 10 × 10 areas, a feature point is determined in each area, the feature point can be extracted by a Moravac operator or other operators, and the homologous point of the right image is obtained through image matching.
And the precise matching of the feature points is realized by adopting a pyramid and least square matching method. The method comprises the steps of averagely generating a 3-layer pyramid image by taking 3 x 3 pixels as one pixel, sequentially calling the image from bottom to top as a first layer image, a second layer image and a third layer image, sequentially extracting feature points on a third layer image, a second layer image and a first layer image pair by adopting a discrimination method with the maximum correlation coefficient and matching the feature points with homologous points, wherein a search window is a rectangular area determined by the maximum parallax of left and right image rows and columns. And after pyramid matching is completed, accurate matching of the images is realized by adopting a least square image matching method.
If geometric distortion and radiation distortion are considered, the homonymy point should satisfy the surrounding gray function:
wherein h is0 h1Representing a radiation distortion parameter, a0 a1 a2 b0 b1 b2Representing the geometric distortion parameter. In addition, the formula (7) is taken as an observation equation, an indirect adjustment method is adopted, and the radiation parameter and the geometric parameter can be obtained by utilizing a correlation coefficient maximum discrimination method, so that the homonymous point of one image point can be accurately determined. The maximum correlation coefficient discrimination method means that iteration is stopped when the correlation coefficient after iteration is smaller than the correlation coefficient after the last iteration.
The relative relation of the image pair before relative orientation is not determined, and the pyramid matching can only search the homonymy point according to the rectangular area determined by the maximum parallax of the image rows and columns. The search range is large, sea surface image texture information is little, and the rough difference is inevitable often because of specular reflection.
In fact, at present, no method for rapidly and automatically judging and eliminating gross error points exists, and the invention innovatively provides a semi-automatic gross error judging and eliminating method. The method comprises the following specific steps:
firstly, finding an accompanying point for each feature point, and removing gross errors according to the parallax errors.
Specifically, for each feature point, a point with the most texture information, namely an accompanying point, is searched within the range of r being more than or equal to 5 and less than or equal to 7, and the method is adopted to match the point with the same name. Since the sea surface has no cliff type fluctuation, the parallax difference between the feature point and the accompanying point is not very large and is generally within 5 image elements. According to the judgment, the part gross errors can be automatically removed.
And step two, rejecting gross errors according to the parallax change of the feature points on the same line.
Because the feature points of the left image are approximately uniformly distributed and the sea surface fluctuation is small relative to the photographic height, the parallax of the feature points on the same row does not change violently, and the discrepancy can be eliminated manually according to the criterion.
The following will further explain 50 image pairs (hereinafter referred to as "image sequence of bosch experiment") obtained from a stereographic experiment performed on a bosch sea and meteorological observation tower in the open sea of maken, guangdong province. Fig. 5 is a view of the disparity change between 10 × 10 feature points and their corresponding points in the first image pair calculated according to the above method. In the figure, solid lines red, orange, yellow, green, blue, indigo and purple sequentially represent characteristic points of rows 1 to 7, and dotted lines red, orange and yellow sequentially represent characteristic points of rows 8 to 10; the abscissa represents a column of feature points; (a) the ordinate represents the row parallax of the feature point and the homonymy point thereof, (b) the ordinate represents the column parallax of the feature point and the homonymy point thereof; the dots with black circles in the figure indicate that they have been identified and culled by the first step discrimination.
As can be seen from fig. 5, the feature points in the rows 4, 5 and 9 of the 3 rd row and the 4 th row and the 9 th column are matched incorrectly and need to be removed, while the incorrectly matched points in the rows 4 and the 5 th row and the 4 th row and the 9 th column of the 3 rd row and the 9 th column need to be removed manually in the first step.
3. Characteristic point pair extraction and gross error elimination of sequence image pair
This section is also one of the core innovations of the present invention. The characteristic points and the homologous points of the subsequent image pairs are determined in sequence by adopting the method. There is no question that these pairs are mixed with the pairs with the matching error, but the gross error discrimination and elimination method is different from the first image pair.
Since n (n is far larger than 5) homonymous point pairs are obtained in the first image pair, an equation set consisting of n control equations can be obtained according to the formula (3). The coordinates of the same-name point pairs are used as observed quantities, 5 relative orientation parameters are used as adjustment parameters, and the relative orientation parameter values (approximate values) can be obtained by adopting a conditional adjustment method with parameters.
And calculating the epipolar line y coordinate corresponding to each homonymous point x coordinate, and solving the difference delta y between the epipolar line y coordinate and the homonymous point y coordinate. Theoretically, the point where the delta y is zero is considered as an error matching point and is removed, wherein the larger the absolute value of the delta y is, the larger the matching error is. Fig. 6 shows 5000 Δ y from 50 image pairs of the bosch experiment, where a large number of mismatching points are present.
4. Exact solution of relative orientation parameters
The main idea of the part is to gradually remove the homonymous point pairs with larger errors through an iteration process, reserve the homonymous point pairs with high matching precision and gradually enable the relative orientation parameters to approach to a true value, and the process is another core innovation of the invention. The following is also a description of the Bohe experiment as an example.
The first step, removing homonymous point pairs with delta y more than or equal to 6 to obtain a residual homonymous point pair set A1Using A1Using the first image calibration result as the initial value, recalculating the relative orientation parameter to obtain X1。
Second step, using X1Computing a set A1And calculating the difference delta y between the epipolar line y coordinate corresponding to the x coordinate of each homonymous point and the y coordinate of the homonymous point.
Third, rejecting set A1Obtaining the residual homonymous point pair set A by homonymous point pairs with the middle delta y more than or equal to 52Using A2Recalculating the relative orientation parameters to obtain X2The initial value is X1。
And fourthly, repeating the second step and the third step to sequentially remove homonymous point pairs with delta y being more than or equal to 4321, and finally calculating accurate and stable relative orientation parameter values.
Example 2
The present invention will be described in detail with reference to specific embodiments, and the present invention can be implemented by using all languages with computing functions such as Matlab, etc., and Matlab is used as an example for description. It should be understood that the Matlab language is described herein as an example only to explain the present invention and is not intended to limit the present invention. FIG. 2 is a schematic flow diagram of the process of the present invention, which is described in detail below:
the first step is as follows: reading a group of sea surface image sequences not less than 50 pairs
Installing a double camera on an observation platform according to the requirement of close-range photogrammetry, increasing the area of an image pair overlapping area as much as possible, acquiring a list of sea surface image sequences of not less than 50 pairs, preferably adopting a tiff picture format, and reading the pictures by using an imread command.
The second step is that: distortion calibration of image sequences
And calibrating each image in the image sequence according to the orientation parameters and the distortion coefficients in the camera, wherein the calibrated images conform to the pinhole imaging model.
The third step: and selecting a first image pair of the image sequence, and determining the maximum parallax in the selected matching area.
Selecting a first image pair of an image sequence, displaying the first image pair through 'image', artificially and approximately selecting a homonymous point pair near each vertex in a quadrilateral overlapping region, and determining a maximum matching region, a maximum row parallax and a maximum column parallax of a rectangle through the four homonymous point pairs.
The fourth step: and extracting the feature point pairs on the left image, and removing the difference to obtain n more than or equal to 5 residual feature point pairs.
According to the technical scheme of '2', 10 multiplied by 10 feature points are extracted from the left image, the same-name points are obtained through matching, the feature points and the same-name points are called feature point pairs, and the feature point pairs with gross errors are removed.
The fifth step: determining a set of governing equations 1 for an initial image-based pair
According to the technical scheme '1', each characteristic point pair obtained in the fourth step can be listed by a formula (3), so that a control equation set consisting of n equations is obtained.
And a sixth step: preliminary relative orientation
The coordinates of the same-name point pairs are used as observed quantities, 5 relative orientation parameters are used as adjustment parameters, the initial values in the technical scheme 1 are adopted, and the relative orientation parameter values are obtained by using a conditional adjustment method with parameters, so that the initial relative orientation is called as initial relative orientation.
The seventh step: characteristic point pair extraction and gross error elimination of sequence image pair
And (3) sequentially extracting the characteristic point pairs of each image pair in the image sequence according to the method of the technical scheme 2, and removing the rough difference points according to the technical scheme 3.
Eighth step: precise relative orientation
And calculating accurate relative orientation parameters according to the technical scheme '4'.
In order to verify the technology of the invention, the experiment data of the open sea experiment of the Bohe meteorological-oceanic comprehensive observation platform in Guangdong province on 11 months and 11 days in 2015 is used for carrying out the technology verification experiment. Two American SI-6600CL industrial cameras are adopted in the experiment, and are provided with 12mm lenses, the pixel size of the camera is 3.5 mu m, and the resolution is 2208 multiplied by 3002.
Fig. 5(a) -5 (b) are parallax change diagrams of 10 × 10 feature points and their corresponding points of the first image in the sequence of the bosch experimental images obtained in step 5 according to the present invention, from which feature points [3,9], i.e., the third row and the 9 th column, which are gross error points, can be easily determined and need to be removed. Table 1 the 1 st action is calculated according to the sixth step, and the 2 nd action is calculated according to the eighth step to obtain the accurate relative orientation parameters of 50 pairs of sequential images.
Table 1 relative orientation parameter of first image and relative image parameter unit of sequence image: rad (radius of curvature)
In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When used in whole or in part, can be implemented in a computer program product that includes one or more computer instructions. When loaded or executed on a computer, cause the flow or functions according to embodiments of the invention to occur, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, the computer instructions may be transmitted from one website site, computer, server, or data center to another website site, computer, server, or data center via wire (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL), or wireless (e.g., infrared, wireless, microwave, etc.)). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that includes one or more of the available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., Solid State Disk (SSD)), among others.
The above description is only for the purpose of illustrating the present invention and the appended claims are not to be construed as limiting the scope of the invention, which is intended to cover all modifications, equivalents and improvements that are within the spirit and scope of the invention as defined by the appended claims.
Claims (10)
1. An automatic relative orientation method based on a sequence sea surface image is characterized by comprising the following steps:
reading a group of sea surface image sequences of not less than 50 pairs;
carrying out distortion calibration on the image sequence;
selecting a first image pair of the image sequence, and determining the maximum parallax in the selected matching area;
extracting characteristic point pairs on the left image, and eliminating out-of-tolerance to obtain residual characteristic point pairs;
determining a control equation set 1 based on the first image pair;
preliminary relative orientation;
extracting characteristic point pairs of the sequence image pairs and rejecting gross errors;
the precise relative orientation.
2. The method of claim 1, wherein reading a set of at least 50 pairs of sea surface image sequences comprises: the method comprises the steps of installing a double camera on an observation platform, increasing the area of an image pair overlapping area, collecting a row of sea surface image sequences of not less than 50 pairs, adopting a tiff picture format and reading pictures by using an imread command.
3. The method of claim 1, wherein the distortion calibration of the image sequence comprises: and calibrating each image in the image sequence according to the orientation parameters and the distortion coefficients in the camera, wherein the calibrated images conform to the pinhole imaging model.
4. The method of claim 1, wherein the determining a maximum disparity for a first image pair of the selected image sequence in the selected matching region comprises: selecting a first image pair of an image sequence, displaying the first image pair through 'image', artificially and approximately selecting a homonymous point pair near each vertex in a quadrilateral overlapping region, and determining a maximum matching region, a maximum row parallax and a maximum column parallax of a rectangle through the four homonymous point pairs.
5. The method of claim 1, wherein the extracting pairs of feature points on the left image and rejecting differences to obtain n ≧ 5 remaining pairs of feature points comprises: extracting characteristic point pairs of the first image pair and rejecting gross errors:
extracting 10 multiplied by 10 characteristic points from the left image, matching to obtain homologous points of the characteristic points, wherein the characteristic points and the homologous points are called characteristic point pairs, and rejecting the characteristic point pairs with gross errors, and the method comprises the following steps:
uniformly dividing the left image overlapping area into a plurality of areas, determining a characteristic point in each area, extracting the characteristic point by using a Moravac operator or other operators, and obtaining the homonymous point of the right image through image matching;
precise matching of the feature points is realized by adopting a pyramid and least square matching method; generating a 3-layer pyramid image by averaging 3 multiplied by 3 pixels into one pixel, sequentially calling the image as a first layer image, a second layer image and a third layer image from bottom to top, sequentially extracting feature points on the third layer image, the second layer image and the first layer image by adopting a discrimination method with the maximum correlation coefficient and matching the feature points with the corresponding points, wherein a search window is a rectangular area determined by the maximum parallax of the rows and the columns of the left image and the right image; after pyramid matching is completed, accurate matching of images is achieved by adopting a least square image matching method;
if geometric distortion and radiation distortion are considered, the homonymy point should satisfy the surrounding gray function:
wherein h is0 h1Representing a radiation distortion parameter, a0 a1 a2 b0 b1 b2Representing a geometric distortion parameter; the radiation parameters and the geometric parameters can be solved by adopting an indirect adjustment method and utilizing a maximum correlation coefficient discrimination method, so that the homonymous point of one image point is accurately determined; the maximum correlation coefficient judging method is that when the correlation coefficient after iteration is smaller than the correlation coefficient after the last iteration, the iteration is stopped;
the semi-automatic gross error judging and eliminating method comprises the following steps:
(1) searching a companion point for each feature point, and removing gross errors according to the parallax errors;
for each feature point, a point with the most texture information, namely an accompanying point, is searched within the range of r being more than or equal to 5 and less than or equal to 7, and the method is adopted to match the point with the same name; because the sea surface does not have cliff type fluctuation, the parallax difference between the characteristic points and the accompanying points is not very large and is within 5 pixels; part of gross errors can be automatically removed according to the judgment;
(2) removing gross errors according to the parallax change of the feature points on the same line;
because the feature points of the left image are approximately uniformly distributed and the sea surface fluctuation is small relative to the photographic height, the parallax of the feature points on the same row does not change violently, and the discrepancy can be eliminated manually according to the criterion.
6. The method of claim 1, wherein said determining a set of governing equations 1 based on a first image pair comprises: as with other photogrammetry, the relative orientation control equation in the sea wave measurement is also a coplanar equation; establishing a coordinate system o-xyz, hereinafter referred to as a left camera coordinate system, by taking a main optical axis of the left camera as a z axis, and respectively parallel to the row direction and the column direction of the CCD, and establishing a right camera coordinate system o '-x' y 'z' in the same way; one point A (x) on the object to be measuredA,yA,zA) With a picture point a (x) on each of the left and right camerasa,ya,-f)、a'(x'a',y'a'-f '), then the straight lines oa, o' a 'and oo' are coplanar, i.e.:
let o-xyz rotate a continuously around the y, x axesy、αxAfter angle, z-axis and base line
Overlapping; then, a left camera coordinate system o-xyz is set to rotate beta around the y, x and z axes continuouslyy、βx、βzAfter the angle, the coordinate axes are parallel to the coordinate axes corresponding to the coordinate system o '-x' y 'z' of the right camera; the relative orientation is the solution parameter αy、αx、βy、βx、βzA value of (d); thus the baseline in the left camera coordinate system can be expressed as:
wherein the content of the first and second substances,
the unit vectors in the directions of the x axis, the y axis and the z axis are represented, D represents the length of a base line and needs actual measurement; thus, formula (2) may represent an in-line form:
wherein (xi)a',ηa',ζa') Satisfy the relation:
wherein R is0Represents a transformation matrix between the coordinate systems o-xyz and o '-x' y 'z':
thus, (containing only alpha)y、αx、βy、βx、βz5 relative orientation parameters are equal, so that theoretically, the relative orientation parameters can be calculated as the relative orientation control equation only by obtaining 5 homonymous point pairs;
because the nonlinear equation set only contains 5 parameters and the matching precision of the sea surface image feature points is low, practice shows that the initial value provided by the direct solution of relative orientation sometimes has not ideal effect; as the main optical axes of the two cameras are approximately parallel when the stereo photogrammetry is used for measuring the sea waves, the practice shows that the initial values of the relative orientation parameters can be selected as follows:
[αx αy βx βy βz]＝[0 -1.5 0 0 0]。
7. the automated method of relative orientation based on sequential sea surface images of claim 1, wherein said preliminary relative orientation comprises: n homonymous point pairs are obtained in the first image pair; wherein n is much greater than 5; and after an equation set consisting of n control equations is obtained, taking the coordinates of the same-name point pairs as observed quantities, taking 5 relative orientation parameters as adjustment parameters, adopting the initial values obtained in the fifth step, and solving the relative orientation parameter values by using a conditional adjustment method with parameters, wherein the initial relative orientation is called as initial relative orientation.
8. The method of claim 1, wherein the extracting and gross-rejection of the feature point pairs of the sequence image pairs comprises: after sequentially determining the characteristic points and the homonymous points of the subsequent image pairs, calculating the epipolar line y coordinate corresponding to the x coordinate of each homonymous point, and solving the difference delta y between the epipolar line y coordinate and the y coordinate of the homonymous point; theoretically, the point where the delta y is zero is considered as an error matching point and is removed, wherein the larger the absolute value of the delta y is, the larger the matching error is.
9. The automated method of relative orientation based on sequential sea surface images of claim 1, wherein said precise relative orientation comprises:
(1) removing homonymous point pairs with delta y more than or equal to 6 to obtain residual homonymous point pair set A1Using A1Using the first image calibration result as the initial value, recalculating the relative orientation parameter to obtain X1；
(2) By using X1Computing a set A1The epipolar line y coordinate corresponding to the x coordinate of each homonymy point is calculated, and the difference delta y between the epipolar line y coordinate and the y coordinate of the homonymy point is calculated;
(3) culling set A1Obtaining the residual homonymous point pair set A by homonymous point pairs with the middle delta y more than or equal to 52Using A2Recalculating the relative orientation parameters to obtain X2The initial value is X1；
(4) And (3) repeating the step (2) and the step (3), sequentially removing homonymous point pairs of which the delta y is more than or equal to [4321], and finally calculating an accurate and stable relative orientation parameter value.
10. An automated relative orientation system based on sequential sea surface images for implementing the automated relative orientation method based on sequential sea surface images according to any one of claims 1 to 9, wherein the automated relative orientation system based on sequential sea surface images comprises:
the image sequence reading module is used for installing the double cameras on the observation platform, increasing the area of an image pair overlapping area, acquiring a row of sea surface image sequences of not less than 50 pairs, reading the images by using an imread command by adopting a tiff image format;
the distortion calibration module is used for calibrating each image in the image sequence according to the camera internal orientation parameters and the distortion coefficients, and the calibrated images conform to the pinhole imaging model;
the maximum parallax determining module is used for selecting a first image pair of the image sequence and determining the maximum parallax in a selected matching area;
the characteristic point pair extraction module is used for extracting characteristic point pairs on the left image and eliminating the difference to obtain n more than or equal to 5 residual characteristic point pairs;
the control equation set determining module is used for determining a control equation set 1 based on the first image pair;
the preliminary relative orientation module is used for calculating a relative orientation parameter value by using the initial value obtained in the fifth step and a conditional adjustment method with parameters, wherein the initial value is the coordinate of the homonymous point pair as an observed quantity, and 5 relative orientation parameters are adjustment parameters, and the relative orientation parameter value is called preliminary relative orientation;
the gross error elimination module is used for extracting the characteristic point pairs of the sequence image pairs and eliminating gross errors;
and the precise relative orientation module is used for determining a precise solution of the relative orientation parameter.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110234110.4A CN113029108B (en) | 2021-03-03 | 2021-03-03 | Automatic relative orientation method and system based on sequence sea surface images |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110234110.4A CN113029108B (en) | 2021-03-03 | 2021-03-03 | Automatic relative orientation method and system based on sequence sea surface images |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113029108A true CN113029108A (en) | 2021-06-25 |
| CN113029108B CN113029108B (en) | 2022-08-26 |
Family
ID=76465666
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110234110.4A Active CN113029108B (en) | 2021-03-03 | 2021-03-03 | Automatic relative orientation method and system based on sequence sea surface images |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113029108B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113658170A (en) * | 2021-08-30 | 2021-11-16 | 杭州柳叶刀机器人有限公司 | Joint registration point generation method and device, electronic equipment and storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105761257A (en) * | 2016-02-05 | 2016-07-13 | 中测新图(北京)遥感技术有限责任公司 | Elimination method for gross error in unmanned aerial vehicle image matching on cross air strip and device thereof |
| CN106767899A (en) * | 2016-11-22 | 2017-05-31 | 航天恒星科技有限公司 | It is a kind of that the Incorporate scaling method and device of satellite are painted based on day |
-
2021
- 2021-03-03 CN CN202110234110.4A patent/CN113029108B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105761257A (en) * | 2016-02-05 | 2016-07-13 | 中测新图(北京)遥感技术有限责任公司 | Elimination method for gross error in unmanned aerial vehicle image matching on cross air strip and device thereof |
| CN106767899A (en) * | 2016-11-22 | 2017-05-31 | 航天恒星科技有限公司 | It is a kind of that the Incorporate scaling method and device of satellite are painted based on day |
Non-Patent Citations (3)
| Title |
|---|
| 王琳: "相对定向中错误匹配的剔除方法研究", 《城市勘测》 * |
| 王英霞: "基于海面影像的立体摄影相对定向方法", 《海洋科学进展》 * |
| 王英霞: "岸基数字摄影海浪三维观测技术及应用研究", 《中国优秀博硕士学位论文全文数据库(博士)》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113658170A (en) * | 2021-08-30 | 2021-11-16 | 杭州柳叶刀机器人有限公司 | Joint registration point generation method and device, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113029108B (en) | 2022-08-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109903227B (en) | Panoramic image splicing method based on camera geometric position relation | |
| CN107833181B (en) | Three-dimensional panoramic image generation method based on zoom stereo vision | |
| WO2018209968A1 (en) | Camera calibration method and system | |
| CN107492069B (en) | Image fusion method based on multi-lens sensor | |
| WO2014024579A1 (en) | Optical data processing device, optical data processing system, optical data processing method, and optical data processing-use program | |
| CN109580649B (en) | Engineering structure surface crack identification and projection correction method and system | |
| CN109727278B (en) | Automatic registration method for airborne LiDAR point cloud data and aerial image | |
| CN110874854B (en) | Camera binocular photogrammetry method based on small baseline condition | |
| CN106595700A (en) | Target channel space reference calibration method based on three-point coordinate measurement | |
| CN111709985B (en) | Underwater target ranging method based on binocular vision | |
| CN112929626B (en) | Three-dimensional information extraction method based on smartphone image | |
| CN108759788B (en) | Unmanned aerial vehicle image positioning and attitude determining method and unmanned aerial vehicle | |
| CN113947638B (en) | Method for correcting orthographic image of fish-eye camera | |
| Rüther et al. | A comparison of close-range photogrammetry to terrestrial laser scanning for heritage documentation | |
| EP1484576A2 (en) | Apparatus and method for calibrating zoom lens | |
| CN112288848A (en) | Method for calculating engineering quantity through three-dimensional modeling of unmanned aerial vehicle aerial photography | |
| CN113298947A (en) | Multi-source data fusion-based three-dimensional modeling method medium and system for transformer substation | |
| CN110807815A (en) | Rapid underwater calibration method based on two groups of mutually orthogonal parallel lines corresponding vanishing points | |
| CN114549746A (en) | High-precision true color three-dimensional reconstruction method | |
| CN108269234A (en) | A kind of lens of panoramic camera Attitude estimation method and panorama camera | |
| CN113029108B (en) | Automatic relative orientation method and system based on sequence sea surface images | |
| CN116681827A (en) | Defect-free three-dimensional point cloud reconstruction method and device based on multi-monitoring camera and point cloud fusion | |
| CN111968182B (en) | Calibration method for nonlinear model parameters of binocular camera | |
| CN112819900B (en) | Method for calibrating internal azimuth, relative orientation and distortion coefficient of intelligent stereography | |
| CN115326025B (en) | Binocular image measurement and prediction method for sea waves |
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 |