CN110487254A - A kind of submarine target size method for fast measuring for ROV - Google Patents
A kind of submarine target size method for fast measuring for ROV Download PDFInfo
- Publication number
- CN110487254A CN110487254A CN201910622509.2A CN201910622509A CN110487254A CN 110487254 A CN110487254 A CN 110487254A CN 201910622509 A CN201910622509 A CN 201910622509A CN 110487254 A CN110487254 A CN 110487254A
- Authority
- CN
- China
- Prior art keywords
- rov
- binocular
- algorithm
- matching
- submarine target
- 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
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
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/50—Depth or shape recovery
- G06T7/55—Depth or shape recovery from multiple images
- G06T7/593—Depth or shape recovery from multiple images from stereo images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
- G06T7/85—Stereo camera calibration
-
- 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
- G06T2207/10012—Stereo images
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Image Processing (AREA)
Abstract
The present invention is to provide a kind of submarine target size method for fast measuring for ROV.Step 1 carries out underwater calibrating parameters correction and corrects with Bouguet polar curve to the binocular to distort under water or so view;Step 2 selects an endpoint in the corresponding two-end-point of object to be measured on the binocular left view in ROV water surface monitoring system, determines the match point in right figure using thick-smart Stereo Matching Algorithm;Step 3 recovers the three-dimensional coordinate of institute's reconnaissance according to binocular measuring principle;Step 4 repeats step 2 and step 3 obtains second endpoint, then calculates the Euclidean distance of the two points again, obtains the actual distance between two points.Relative to traditional measurement method based on global registration algorithm, method therefor of the present invention be can be realized more rapidly with more accurate submarine target dimensional measurement, can be widely applied to ROV subaqueous survey task, have extraordinary practicability.
Description
Technical field
The present invention relates to a kind of submarine target measurement methods, specifically a kind of to be suitable for monitoring in the ROV water surface
It is more accurate more quickly to the interesting target arbitrarily manually selected on underwater two-dimension image even any two ends point in system
The method that air line distance measures.
Background technique
The developing direction of modern range-measurement system tends to be intelligent, and wide application platform is provided for stereoscopy passive ranging,
As passive ranging, stereovision technique is not necessarily to emit any signal to object, only need to collect figure by imaging sensor
As right, by identification, the monitoring to image information, the intelligence of range-measurement system will be effectively improved.Therefore, research is based on vertical
The real time distance system of body vision technology has very important application value.
At present in ROV subaqueous survey research field, main solution be all based on BM (Block Matching) or
SGM (Semiglobal Matching) algorithm carries out three-dimensional reconstruction after carrying out Stereo matching, such method needs to carry out global view
Difference figure calculates, and the time is longer.In addition to this, since underwater picture contrast low noise is big, feature is unintelligible, traditional measurement method
The depth information error calculated by global registration is larger, leads to the low measurement accuracy of submarine target size.
Summary of the invention
Measuring speed can be improved the purpose of the present invention is to provide one kind and that improves measurement accuracy is used for ROV's
Submarine target size method for fast measuring.
The object of the present invention is achieved like this:
Step 1 carries out underwater calibrating parameters correction and rectifys with Bouguet polar curve to the binocular to distort under water or so view
Just;
Step 2 is selected in the corresponding two-end-point of object to be measured on the binocular left view in ROV water surface monitoring system
One endpoint determines the match point in right figure using thick-smart Stereo Matching Algorithm;
Step 3 recovers the three-dimensional coordinate of institute's reconnaissance according to binocular measuring principle;
Step 4 repeats step 2 and step 3 obtains second endpoint, then calculates the Euclidean distance of the two points again,
Obtain the actual distance between two points.
The present invention may also include:
1. described determine that the match point in right figure specifically includes using thick-smart Stereo Matching Algorithm:
(1) box that a 9*9 pixel is outlined centered on the reconnaissance of left figure institute, is matched in right figure using thick matching algorithm
Optimal 9*9 pixel region out;
(2) smart matching is carried out in the thick matching area of right view, is traversed the region and is found optimal match point.
2. the thick matching algorithm selects SAD algorithm.
3. the progress essence matching in the thick matching area of right view specifically includes:
Match point p2 in right view should meet formula: p2=min (ρ (I1(Ωr(p1)),I2(Ωr(p2)))),
Middle Ω r (p1) indicates that centered on p1, r is the radius of neighbourhood, In(x) it indicates gray value, the difference of left and right matching area is represent with ρ
The opposite sex,
Similitude is calculated using cross correlation algorithm and Census normalization blending algorithm:
Wherein α is the weight coefficient of two kinds of similarity calculations of equilibrium of ε (f, g), and ε (f, g) and φ (f, g) are NCC respectively
What (f, g) and Census (f, g) were obtained by normalization, the output valve range of NCC (f, g) is [- 1,1], Census's (f, g)
Output valve range is [0, r2];
The output valve of two kinds of similarity algorithms is normalized between [0,1] using linear function normalization algorithm, normalizing
It is as follows to change function:
Wherein max is the maximum value of sample data, and min is the minimum value of sample data.
4. described according to binocular measuring principle, the three-dimensional coordinate for recovering institute's reconnaissance is specifically included:
For the point P (X, Y, Z) in space, p1 (x1, y1) and p2 (x2, y2) and y-axis are respectively formed on the view of left and right
Coordinate is identical, corrects gain of parameter focal length f, optical center coordinate (cx, cy), camera lens parallax range B according to binocular, disparity computation is as follows
Formula is
D=x1-x2
Using binocular imaging principle, the three-dimensional coordinate of P is recovered:
The present invention provides a kind of submarine target size quick high accuracy measurement methods for ROV, three-dimensional using thick-essence
The match point that matching process only solves tested two o'clock obtains local optimum, avoids existing global registration algorithm and calculates the overall situation most
The case where figure of merit, improves measuring speed and precision.The present invention can be applicable to the monitoring of the water surface in ROV subaqueous survey task system
System carries out space distance measurement just for the two-end-point of the interesting target arbitrarily manually selected on two dimensional image.Of the invention
Main feature includes: the measurement that (1) is directed to submarine target size, need to locally solve tested end two o'clock at algorithm design upper
Left and right views registered point, without carrying out global registration, so as to preferably avoid error hiding.(2) binocular solid matches Shi Xiantong
It crosses SAD (Sun of Absolute Differences) algorithm slightly to be matched, and utilizes NCC in quasi- matching area
The high-precision matching algorithm that (Normalized Cross Correlation) is merged with Census normalization finds best match
Point.(3) coordinate formula of two endpoints in target is recovered using three-dimensional reconstruction, then is calculated two o'clock Euclidean distance and acquired under water
The real space dimension of any two endpoint, the three-dimensional coordinate without all the points in reconstruction image.(4) using thick-essence three-dimensional
Method of completing the square need to only obtain local optimum, therefore have better insensitivity to picture noise and illumination effect.(5) of the invention
It can be applicable in the water surface monitoring system in ROV subaqueous survey task, it is emerging just for the sense arbitrarily manually selected on two dimensional image
Two endpoints of interesting target carry out space distance measurement.
Compared with the prior art, the invention has the following beneficial effects:
1. the advantages of thick-smart Stereo Matching Algorithm that the present invention designs has merged a variety of matching algorithms, has to illumination not
Sensitive characteristic can also preferably extract the feature of edge and angle point;
2. the present invention only solves the match point of tested two o'clock, solves existing global registration algorithm and only demand global optimum rather than office
The problem of portion's optimal value, improve measuring speed and precision;
3. even Stereo Matching Algorithm of the invention can also compare calibration chessboard is this to repeat the more pattern of texture
Good finds match point, overcomes the drawbacks of existing matching algorithm is affected by pixel value.
4. the present invention need to only obtain local optimum using thick-smart solid matching method, therefore to picture noise and illumination
Influence that there is better insensitivity.
Detailed description of the invention
Fig. 1 is the flow chart of the submarine target size quick high accuracy measurement method for ROV.
Fig. 2 a- Fig. 2 d gives underwater original image and correcting image contrast effect figure, in which: Fig. 2 a is original left view
Figure;Fig. 2 b is correction left view;Fig. 2 c is original right view;Fig. 2 d is correction right view.
Fig. 3 gives the exemplary diagram of Census algorithm operational process.
Fig. 4 gives binocular measuring principle figure.
Fig. 5 gives the measurement effect figure of Different matching algorithm, wherein from left to right: first row is original image;Second
Column are BM;Third column are SGM;4th column are measurement figures of the present invention.
Fig. 6 is the measured value of Different matching algorithm described in step S4 and error contrast table.
Specific embodiment
It illustrates below and the present invention is described in more detail.
In conjunction with Fig. 1, the submarine target size quick high accuracy measurement method for ROV of the invention includes the following steps:
S1. it for the binocular to distort under water or so view, carries out underwater calibrating parameters correction and is corrected with Bouguet polar curve;
First binocular camera is demarcated using Zhang Zhengyou algorithm under water, the left and right view of shooting is joined according to calibration
Number carries out distortion correction, then obtains the left and right view after the correction of row alignment using Bouguet polar curve correction algorithm.Two after correction
The optical axis of a camera is substantially parallel, and height of the picture point on left images is consistent, only need to be same in subsequent Stereo matching
Search left and right can be such that efficiency greatly improves as the match point of plane on row, facilitate successive depths image and calculate and three-dimensional
It rebuilds.Original image and correcting image are demarcated as shown in Fig. 2 a to Fig. 2 d.
S2. it is manually selected in the corresponding two-end-point of object to be measured on the binocular left view in ROV water surface monitoring system
One endpoint determines the match point in right figure using thick-smart Stereo Matching Algorithm of proposition;
Size due to obtaining target only needs to click through both ends on binocular vision left view in ROV water surface monitoring system
Row is chosen manually, therefore can obtain match point on right view using thick-smart local matching method.
Specifically comprise the following steps:
S21. the box that a 9*9 pixel is outlined centered on the reconnaissance of left figure institute, using thick matching algorithm in right figure on
Allot optimal 9*9 pixel region.SAD algorithm is a kind of template matching algorithm based on grey scale pixel value, which has meter
Calculate the characteristics of not needing complicated multiplication and division arithmetic simply, the preliminary screening suitable for multistep treatment.Therefore, of the invention
Quasi- matching area is found in thick matching using SAD algorithm, the end point p1 of object to be measured is manually selected in left view, automatically
The neighborhood frame for generating a 9*9 pixel, scans for Block- matching to right figure.It is as follows that SAD matches cost function:
The basic thought of the algorithm is the absolute value of the difference summation that each pixel is corresponded to numerical value, assesses two images accordingly
The similarity of block.It finds out absolute value and the smallest image block is exactly and the most matched region of template.
S22. smart matching is carried out in the thick matching area of right view, is traversed the region and is found optimal match point.
Match point p2 in right view should meet formula:
p2=min (ρ (I1(Ωr(p1)),I2(Ωr(p2))))
Ω r (p1) indicates that centered on p1, r is the radius of neighbourhood (passing through experimental verification, it is preferable to be set as 5 effects), In(x)
It indicates gray value, the otherness of left and right matching area is represent with ρ.Present invention employs cross correlation algorithms and Census to normalize
Blending algorithm is for calculating similitude:
Wherein α is the weight coefficient of two kinds of similarity calculations of equilibrium of ε (f, g), is set as 0.5 in experiment.ε (f, g) and
φ (f, g) is that NCC (f, g) and Census (f, g) is obtained by normalization respectively.The output valve range of NCC (f, g) be [- 1,
1], the output valve range of Census (f, g) is [0, r2].The present invention uses linear function normalization algorithm by two kinds of similarity operators
The output valve of method normalizes between [0,1], and normalized function is as follows:
Wherein max is the maximum value of sample data, and min is the minimum value of sample data.
NCC is a kind of algorithm based on two groups of sample data correlations of statistical calculations, and for image, each pixel
Point can be regarded as RGB numerical value, and such entire image can regard the set of a sample data as, if it has one
It is 1 that subset and another sample data, which are mutually matched then its NCC value, indicates that correlation is very high, has indicated if it is -1
It is complete uncorrelated, matching degree between the two is calculated by normalized relativity measurement formula.The formula of NCC (f, g) is such as
Under:
WhereinWithIt is the pixel average in the region f and g, σfAnd σgBe f and g pixel criterion it is poor.
The calculation formula of Census (f, g) is as follows:
Census (f, g)=Hamming (η (f), η (g))
Wherein m is the center pixel of f, and n is the other positions pixel of f, after the function representation f (n) is compared with f (m)
Return value.If f (m) is greater than f (n), 1 is returned, otherwise returns to 0.These 0 and 1 are grouped together to form a sequence in order
Column.Similarly, for the region a g also available sequence.The similitude of Census (f, g) is exactly the corresponding sequence by f and g
Column do an exclusive or to calculate.Sample calculation process is as shown in Figure 3.
S3. according to binocular measuring principle, the three-dimensional coordinate of institute's reconnaissance is recovered:
Binocular measuring principle is respectively formed p1 on the view of left and right as shown in figure 4, for the point P (X, Y, Z) in space
(x1, y1) and p2 (x2, y2) and y-axis coordinate is identical correct parameter according to binocular, can obtain focal length f, optical center coordinate (cx,
Cy), camera lens parallax range B, shown in the following formula of disparity computation.
D=x1-x2
Using binocular imaging principle, the three-dimensional coordinate of P can be recovered:
S4. it repeats S2 and S3 and obtains second endpoint, calculate Euclidean distances of the two points, that is, acquire true between two points
Actual distance from.
S2 and S3, the three dimensional space coordinate of available two points are repeated, then calculates the Euclidean distance of two points
Obtain actual distance.
In order to verify the validity of Measurement Algorithm of the present invention, underwater scaling board and bottle target length are measured,
It is compared again with traditional global registration algorithm BM and SGM algorithm, measurement effect figure is as shown in figure 5, measurement result and error
Such as Fig. 6.Can be seen that the completely black region in part in BM or SGM depth map from Fig. 5 front two row effect indicates not match gridiron pattern
Son is unable to measure size between two grid at all, and it is very big to illustrate that BM or SGM are influenced by image pixel value, cannot be to there is repetition
The characteristics of image of texture is correctly matched.Three row effects can be seen that traditional matching process due to position of bottleneck after Fig. 5
Matching effect official post obtain final measurement error greater than 5%.The characteristics of due to NCC and Census, the present invention can well
With the scaling board with repetition texture.For measurement result of the same object in different distance also very close to error is smaller.Root
It is found according to experimental data, measurement length is bigger, and error is bigger, but global error also can control in lesser range.Therefore, originally
The method of invention is substantially better than traditional global registration algorithm, can be applied in ROV water surface monitoring system, manually emerging to sense
The two-end-point of interesting target carries out space distance measurement, has extraordinary practicability.
The foregoing is merely presently preferred embodiments of the present invention, is not intended to limit the invention, it is all in spirit of the invention and
Any modifications, equivalent replacements, and improvements etc. done within principle, should all be included in the protection scope of the present invention.
Claims (5)
1. a kind of submarine target size method for fast measuring for ROV, it is characterized in that:
Step 1 carries out underwater calibrating parameters correction and corrects with Bouguet polar curve to the binocular to distort under water or so view;
Step 2 selects one in the corresponding two-end-point of object to be measured on the binocular left view in ROV water surface monitoring system
Endpoint determines the match point in right figure using thick-smart Stereo Matching Algorithm;
Step 3 recovers the three-dimensional coordinate of institute's reconnaissance according to binocular measuring principle;
Step 4 repeats step 2 and step 3 obtains second endpoint, then calculates the Euclidean distance of the two points again, obtains
Actual distance between two points.
2. the submarine target size method for fast measuring according to claim 1 for ROV, it is characterized in that the utilization
Slightly-essence Stereo Matching Algorithm determines that the match point in right figure specifically includes:
(1) box that a 9*9 pixel is outlined centered on the reconnaissance of left figure institute, is matched most in right figure using thick matching algorithm
Good 9*9 pixel region;
(2) smart matching is carried out in the thick matching area of right view, is traversed the region and is found optimal match point.
3. the submarine target size method for fast measuring according to claim 2 for ROV, it is characterized in that: described thick
With algorithms selection SAD algorithm.
4. the submarine target size method for fast measuring according to claim 3 for ROV, it is characterized in that described in right view
Scheme progress essence matching in thick matching area to specifically include:
Match point p2 in right view should meet formula: p2=min (ρ (I1(Ωr(p1)),I2(Ωr(p2)))), wherein Ω r
(p1) it indicates centered on p1, r is the radius of neighbourhood, In(x) it indicates gray value, the otherness of left and right matching area is represent with ρ,
Similitude is calculated using cross correlation algorithm and Census normalization blending algorithm:
Wherein α is the weight coefficient of two kinds of similarity calculations of equilibrium of ε (f, g), and ε (f, g) and φ (f, g) are NCC (f, g) respectively
It is obtained with Census (f, g) by normalization, the output valve range of NCC (f, g) is [- 1,1], the output of Census (f, g)
Value range is [0, r2];
The output valve of two kinds of similarity algorithms is normalized between [0,1] using linear function normalization algorithm, normalizes letter
Number is as follows:
Wherein max is the maximum value of sample data, and min is the minimum value of sample data.
5. according to claim 1 to the submarine target size method for fast measuring for being used for ROV described in 4 any one, feature
It is described according to binocular measuring principle, the three-dimensional coordinate for recovering institute's reconnaissance specifically includes:
For the point P (X, Y, Z) in space, p1 (x1, y1) and p2 (x2, y2) and y-axis coordinate are respectively formed on the view of left and right
It is identical, gain of parameter focal length f, optical center coordinate (cx, cy), camera lens parallax range B, the following formula of disparity computation are corrected according to binocular
For
D=x1-x2
Using binocular imaging principle, the three-dimensional coordinate of P is recovered:
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910622509.2A CN110487254B (en) | 2019-07-11 | 2019-07-11 | Rapid underwater target size measuring method for ROV |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910622509.2A CN110487254B (en) | 2019-07-11 | 2019-07-11 | Rapid underwater target size measuring method for ROV |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110487254A true CN110487254A (en) | 2019-11-22 |
CN110487254B CN110487254B (en) | 2021-11-16 |
Family
ID=68546967
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910622509.2A Active CN110487254B (en) | 2019-07-11 | 2019-07-11 | Rapid underwater target size measuring method for ROV |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110487254B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111461079A (en) * | 2020-05-18 | 2020-07-28 | 江苏电力信息技术有限公司 | Binocular image-based method for detecting personnel under suspension arm |
CN111563921A (en) * | 2020-04-17 | 2020-08-21 | 西北工业大学 | Underwater point cloud acquisition method based on binocular camera |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103679699A (en) * | 2013-10-16 | 2014-03-26 | 南京理工大学 | Stereo matching method based on translation and combined measurement of salient images |
CN107588721A (en) * | 2017-08-28 | 2018-01-16 | 武汉科技大学 | The measuring method and system of a kind of more sizes of part based on binocular vision |
CN109191513A (en) * | 2018-08-28 | 2019-01-11 | 江苏久创电气科技有限公司 | Power equipment solid matching method based on global optimization |
-
2019
- 2019-07-11 CN CN201910622509.2A patent/CN110487254B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103679699A (en) * | 2013-10-16 | 2014-03-26 | 南京理工大学 | Stereo matching method based on translation and combined measurement of salient images |
CN107588721A (en) * | 2017-08-28 | 2018-01-16 | 武汉科技大学 | The measuring method and system of a kind of more sizes of part based on binocular vision |
CN109191513A (en) * | 2018-08-28 | 2019-01-11 | 江苏久创电气科技有限公司 | Power equipment solid matching method based on global optimization |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111563921A (en) * | 2020-04-17 | 2020-08-21 | 西北工业大学 | Underwater point cloud acquisition method based on binocular camera |
CN111563921B (en) * | 2020-04-17 | 2022-03-15 | 西北工业大学 | Underwater point cloud acquisition method based on binocular camera |
CN111461079A (en) * | 2020-05-18 | 2020-07-28 | 江苏电力信息技术有限公司 | Binocular image-based method for detecting personnel under suspension arm |
Also Published As
Publication number | Publication date |
---|---|
CN110487254B (en) | 2021-11-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107833181B (en) | Three-dimensional panoramic image generation method based on zoom stereo vision | |
CN107077743B (en) | System and method for dynamic calibration of an array camera | |
CN106683070B (en) | Height measuring method and device based on depth camera | |
Johannsen et al. | On the calibration of focused plenoptic cameras | |
KR100513055B1 (en) | 3D scene model generation apparatus and method through the fusion of disparity map and depth map | |
CN108510551B (en) | Method and system for calibrating camera parameters under long-distance large-field-of-view condition | |
Gedge et al. | Refractive epipolar geometry for underwater stereo matching | |
Pizarro et al. | Relative Pose Estimation for Instrumented, Calibrated Imaging Platforms. | |
US9025862B2 (en) | Range image pixel matching method | |
CN110009690A (en) | Binocular stereo vision image measuring method based on polar curve correction | |
TW201717613A (en) | An image calibrating, composing and depth rebuilding method of a panoramic fish-eye camera and a system thereof | |
CN107560592B (en) | Precise distance measurement method for photoelectric tracker linkage target | |
JP5672112B2 (en) | Stereo image calibration method, stereo image calibration apparatus, and computer program for stereo image calibration | |
CN109493384B (en) | Camera pose estimation method, system, device and storage medium | |
CN112184811B (en) | Monocular space structured light system structure calibration method and device | |
CN107729893A (en) | A kind of vision positioning method of clapper die spotting press, system and storage medium | |
CN109785370A (en) | A kind of weak texture image method for registering based on space time series model | |
CN107679542B (en) | Double-camera stereoscopic vision identification method and system | |
CN110487254A (en) | A kind of submarine target size method for fast measuring for ROV | |
US8340399B2 (en) | Method for determining a depth map from images, device for determining a depth map | |
CN117197333A (en) | Space target reconstruction and pose estimation method and system based on multi-view vision | |
CN114923665B (en) | Image reconstruction method and image reconstruction test system for wave three-dimensional height field | |
CN115719320A (en) | Tilt correction dense matching method based on remote sensing image | |
CN111768448A (en) | Spatial coordinate system calibration method based on multi-camera detection | |
CN109741389A (en) | One kind being based on the matched sectional perspective matching process of region base |
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 |