CN110060202A - A kind of initial method and system of monocular SLAM algorithm - Google Patents
A kind of initial method and system of monocular SLAM algorithm Download PDFInfo
- Publication number
- CN110060202A CN110060202A CN201910319610.0A CN201910319610A CN110060202A CN 110060202 A CN110060202 A CN 110060202A CN 201910319610 A CN201910319610 A CN 201910319610A CN 110060202 A CN110060202 A CN 110060202A
- Authority
- CN
- China
- Prior art keywords
- pixel
- image
- slam algorithm
- monocular
- camera
- 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
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 124
- 238000000034 method Methods 0.000 title claims abstract description 45
- 238000006243 chemical reaction Methods 0.000 claims description 18
- 239000011159 matrix material Substances 0.000 claims description 8
- 230000008859 change Effects 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 5
- 238000013519 translation Methods 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 3
- 238000012546 transfer Methods 0.000 claims description 2
- 230000008569 process Effects 0.000 description 8
- 230000008901 benefit Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 230000000007 visual effect Effects 0.000 description 4
- 230000004888 barrier function Effects 0.000 description 3
- 241000251468 Actinopterygii Species 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 239000012491 analyte Substances 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 235000015243 ice cream Nutrition 0.000 description 1
- 238000011423 initialization method Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Classifications
-
- G06T3/047—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
- G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
- G06T7/75—Determining position or orientation of objects or cameras using feature-based methods involving models
Abstract
The present invention provides the initial methods and system of a kind of monocular SLAM algorithm, this method comprises: choosing the pixel in image crossover region from acquisition image, and after the pixel of selection to be converted to the pixel of corresponding binocular SLAM algorithm, the three-dimensional coordinate of the pixel after converting in image crossover region is calculated based on binocular SLAM algorithm, and the three-dimensional coordinate of pixel is converted to the coordinate information of camera coordinates system, and then can initialize according to the coordinate information of camera coordinates system to monocular SLAM algorithm.Thus, the embodiment of the present invention can be by assisting monocular SLAM algorithm to realize initialization according to binocular SLAM algorithm, so as to make the corresponding monocular system of monocular SLAM algorithm faster, preferably realize initialization, improve the stability and computational accuracy of monocular SLAM algorithm.
Description
Technical field
The present invention relates to automobile technical fields, more particularly to the initial method and system of a kind of monocular SLAM algorithm.
Background technique
At present in the solution of automatic parking auxiliary system (Auto Parking Assist, APA), 4 can be generally carried
The fish-eye camera of a low cost, realizing full vehicle monitoring system (Around View Monitoring, AVM) function, or
Person reaches visual display effect relevant to APA.Vehicle usually needs during finding parking stall using APA in the prior art
Parking stall and static-obstacle thing, such as ice-cream cone, wheel shelves, pole and wire net etc. are detected, or even needs to detect some uncommon
Barrier, such as bicycle, chair etc..Static-obstacle thing is detected using currently used machine learning method, is needed
Be known in advance barrier classification and a large amount of training data, this method make find parking stall versatility and flexibility compared with
Difference.
In addition, current newest research hotspot is point of use cloud to realize static-obstacle analyte detection, but usually require
Using relatively good global camera (Global Shutter Camera) or binocular camera, and largely configured on existing automobile
Fisheye camera can not realize well.The existing vision SLAM (Simultaneous based on monocular fisheye camera
Localization And Mapping) point cloud generation method is primarily present that stability is poor, precision is not high, is difficult to obtain static barrier
The problems such as hindering object true scale.Wherein, stability difference, which is mainly manifested in a cloud generating process often, there is initialization failure, point
Cloud generate that deviation is larger, noise is larger and the point cloud for that can generate that there are dimensional deviations is also larger etc..Precision is asked
Topic and true scale problem be mainly manifested in monocular SLAM algorithm generate a point cloud do not have true scale, scale have larger randomness,
Other aspects such as it is varied with scene and key frame interval etc..
Summary of the invention
In view of the above problems, it proposes on the present invention overcomes the above problem or at least be partially solved in order to provide one kind
State the initial method and system of the monocular SLAM algorithm of problem.
According to the present invention on the one hand, a kind of initial method of monocular SLAM algorithm is provided, comprising:
The pixel in image crossover region is chosen from acquisition image, wherein described image crossover region is to pass through different angles
Spend the content overlapping region of two images of acquisition；
The pixel of selection is converted to the pixel of corresponding binocular SLAM algorithm, is calculated based on the binocular SLAM algorithm
The three-dimensional coordinate of pixel after being converted in image crossover region；
The three-dimensional coordinate of the pixel is converted to the coordinate information of camera coordinates system；
Coordinate information according to the camera coordinates system initializes monocular SLAM algorithm.
Optionally, the pixel in image crossover region is chosen from acquisition image, comprising:
Acquisition image is obtained, the image that will acquire is input in default camera model；
Described image after camera model processing, from treated image according to monocular SLAM algorithm picks image
Pixel in crossover region.
Optionally, the pixel of selection is converted to the pixel of corresponding binocular SLAM algorithm, comprising:
The pixel of selection is converted to the pixel of corresponding binocular SLAM algorithm based on image coordinate system.
Optionally, the three-dimensional coordinate of the pixel after converting in image crossover region is calculated based on the binocular SLAM algorithm,
Include:
Obtain the parallax range between the corresponding camera of acquisition image measured in advance；
Based on the binocular SLAM algorithm and according to the parallax range between the camera, calculates and converted in image crossover region
The three-dimensional coordinate of pixel afterwards.
Optionally, it after being initialized according to the coordinate information of the camera coordinates system to monocular SLAM algorithm, also wraps
It includes:
Pixel out of monocular SLAM algorithm picks image crossover region in acquisition image according to initialization, wherein institute
State the content overlapping region that image crossover region is two images acquired by different angle；
The pixel of selection is converted to the pixel of corresponding binocular SLAM algorithm, is calculated based on the binocular SLAM algorithm
The three-dimensional coordinate of pixel after being converted in image crossover region；
The coordinate information that the three-dimensional coordinate of the pixel is converted to camera coordinates system utilizes the list after the initialization
Mesh SLAM algorithm calculates the pose variation from the camera of an angle acquisition image based on the coordinate information of the camera coordinates system；
Change the three-dimensional coordinate for calculating the pixel in the non-overlapping area of image according to the pose, wherein described image is non-
Crossover region is the content not overlapping region of two images acquired by different angle；
Three-dimensional coordinate according to the non-overlapping area of image and the pixel in crossover region generates point cloud.
Optionally, change the three-dimensional coordinate for calculating the pixel in the non-overlapping area of image according to the pose, comprising:
Monocular SLAM algorithm after passing through initialization simultaneously changes according to the pose, calculates from angle acquisition image
Shift length of the camera between any two different moments；
By the monocular SLAM algorithm after initialization according to the pixel in the shift length calculating non-overlapping area of image
Three-dimensional coordinate.
Optionally, the pose variation of the camera, comprising: the spin matrix and/or translation matrix of the camera.
According to the present invention on the other hand, a kind of initialization system of monocular SLAM algorithm, including monocular subsystem are additionally provided
System, binocular subsystem, conversion subsystem, wherein
The monocular subsystem chooses the pixel in image crossover region, wherein described image is overlapping from acquisition image
Area is the content overlapping region of two images acquired by different angle；
The conversion subsystem is converted to the pixel of selection suitable for the binocular subsystem and corresponding binocular
The pixel of SLAM algorithm；
The binocular subsystem calculates three of the pixel after converting in image crossover region based on the binocular SLAM algorithm
Tie up coordinate；
The three-dimensional coordinate of the pixel is converted to the camera suitable for the monocular subsystem by the conversion subsystem
The coordinate information of coordinate system；
The monocular subsystem, the coordinate information according to the camera coordinates system initialize monocular SLAM algorithm.
According to the present invention in another aspect, additionally providing a kind of computer storage medium, the computer storage medium storage
There is computer program code, when the computer program code is run on the computing device, the calculating equipment is caused to execute
The initial method of monocular SLAM algorithm in any embodiment above.
In embodiments of the present invention, the pixel in image crossover region is being chosen from acquisition image, and by the picture of selection
After vegetarian refreshments is converted to the pixel of corresponding binocular SLAM algorithm, calculated based on binocular SLAM algorithm after being converted in image crossover region
The three-dimensional coordinate of pixel, and the three-dimensional coordinate of pixel is converted to the coordinate information of camera coordinates system, and then can foundation
The coordinate information of camera coordinates system initializes monocular SLAM algorithm.The embodiment of the present invention can be by according to double as a result,
The coordinate information of pixel more acurrate, stable in image crossover region is calculated in mesh SLAM algorithm, to utilize these pictures
The coordinate information auxiliary monocular SLAM algorithm of vegetarian refreshments realizes initialization, so as to make the corresponding monocular system of monocular SLAM algorithm
Faster, it preferably realizes initialization, improves the stability and computational accuracy of monocular SLAM algorithm.
The above description is only an overview of the technical scheme of the present invention, in order to better understand the technical means of the present invention,
And it can be implemented in accordance with the contents of the specification, and in order to allow above and other objects of the present invention, feature and advantage can
It is clearer and more comprehensible, the followings are specific embodiments of the present invention.
According to the following detailed description of specific embodiments of the present invention in conjunction with the accompanying drawings, those skilled in the art will be brighter
The above and other objects, advantages and features of the present invention.
Detailed description of the invention
By reading the following detailed description of the preferred embodiment, various other advantages and benefits are common for this field
Technical staff will become clear.The drawings are only for the purpose of illustrating a preferred embodiment, and is not considered as to the present invention
Limitation.And throughout the drawings, the same reference numbers will be used to refer to the same parts.In the accompanying drawings:
Fig. 1 shows the flow diagram of the initial method of monocular SLAM algorithm according to an embodiment of the invention；
Fig. 2 shows the acquisition visual field schematic diagrames of fisheye camera on vehicle according to an embodiment of the invention；
Fig. 3 shows the structural schematic diagram of the initialization system of monocular SLAM algorithm according to an embodiment of the invention.
Specific embodiment
Exemplary embodiments of the present disclosure are described in more detail below with reference to accompanying drawings.Although showing the disclosure in attached drawing
Exemplary embodiment, it being understood, however, that may be realized in various forms the disclosure without should be by embodiments set forth here
It is limited.On the contrary, these embodiments are provided to facilitate a more thoroughly understanding of the present invention, and can be by the scope of the present disclosure
It is fully disclosed to those skilled in the art.
In order to solve the above technical problems, the embodiment of the invention provides a kind of initial methods of monocular SLAM algorithm.Fig. 1
Show the flow diagram of the initial method of monocular SLAM algorithm according to an embodiment of the invention.Referring to Fig. 1, the party
Method includes at least step S102 to step S108.
Step S102 chooses the pixel in image crossover region from acquisition image, wherein image crossover region is by not
With the content overlapping region of two images of angle acquisition.
In the step, the pixel in image crossover region is actually to be activated a little, the selected pixels in image crossover region
Select points equally is preferably carried out when point, and the density of reconnaissance is slightly larger, to meet the demand of the initialization of monocular SLAM algorithm.
The pixel of selection is converted to the pixel of corresponding binocular SLAM algorithm by step S104, is calculated based on binocular SLAM
Method calculates the three-dimensional coordinate of the pixel after converting in image crossover region.
In the step, the three-dimensional coordinate of the pixel after conversion refers to the three-dimensional coordinate information based on world coordinate system.
The three-dimensional coordinate of pixel is converted to the coordinate information of camera coordinates system by step S106.
Step S108, the coordinate information according to camera coordinates system initialize monocular SLAM algorithm.
In embodiments of the present invention, the pixel in image crossover region is being chosen from acquisition image, and by the picture of selection
After vegetarian refreshments is converted to the pixel of corresponding binocular SLAM algorithm, calculated based on binocular SLAM algorithm after being converted in image crossover region
The three-dimensional coordinate of pixel, and the three-dimensional coordinate of pixel is converted to the coordinate information of camera coordinates system, and then can foundation
The coordinate information of camera coordinates system initializes monocular SLAM algorithm.The embodiment of the present invention is calculated by binocular SLAM as a result,
The coordinate information of pixel more acurrate, stable in image crossover region is calculated in method, to utilize the seat of these pixels
It marks information auxiliary monocular SLAM algorithm and realizes initialization, so as to make the corresponding monocular system of monocular SLAM algorithm faster, more
Good realization initialization, improves the stability and computational accuracy of monocular SLAM algorithm.
Since current vision SLAM algorithm requires the correction figure based on image, the field angle FOV for correcting figure is smaller, adjacent
The overlapping region of two cameras is not big enough.And it can be very good to solve what monocular algorithm was encountered using binocular vision SLAM algorithm
Stability and dimensional accuracy problem.Therefore, the embodiment of the present invention needs to do binocular using at least two cameras, just has enough
Field angle overlapping region is used to binocular vision SLAM algorithm, and at least two cameras can be located at the different angle on vehicle
Any position.
Step S102 is seen above, in an embodiment of the present invention, image crossover region can be mounted in vehicle difference position
The content overlapping region in two images that any two camera set acquires respectively.It is handed over choosing image from acquisition image
When folding the pixel in area, figure can be chosen in a collected frame or multiple image in initialization from one of camera
As the pixel of arbitrary objects in crossover region.Certainly, the equipment of the acquisition image in the embodiment of the present invention is not limited only to camera,
It can also be other image capture devices, the embodiment of the present invention does not do specific restriction to this.
The camera of the acquisition image of camera low cost and big visual angle in order to balance, the installation of vehicle different location angle can be adopted
With the biggish fisheye camera of angular field of view of acquisition image.
For example, with reference to Fig. 2, the acquisition field range for the fisheye camera being mounted on automobile is indicated with semicircle, shown in Fig. 2
It is mounted with 4 fisheye cameras in the automobile of embodiment, the front, rear, left and right four direction of body of a motor car is located at, in this hair
In bright embodiment, two parts can be divided into for the image of two fisheye cameras of arbitrary neighborhood, acquisition, a part is image
Crossover region, another part are the non-overlapping areas of image, wherein image crossover region is that two neighboring fisheye camera acquires picture material phase
Same region, the non-overlapping area of image are that two neighboring fisheye camera acquires the different region of picture material.Assuming that two neighboring fish
Eye camera includes preposition fisheye camera and side fisheye camera, is choosing arbitrary objects from the corresponding image of side fisheye camera
Multiple pixels when, need to carry out select points equally respectively in image crossover region and the non-overlapping area of image, and in image crossover region
The density of the point selected in domain is slightly larger.
Step S102 is seen above, in an embodiment of the present invention, the picture in image crossover region is chosen from acquisition image
During vegetarian refreshments, in order to the image for directly camera being used to acquire, it can will be obtained after the image for getting camera acquisition
The image got is input in default camera model, is handled by camera model image, and then from treated image
According to the pixel in monocular SLAM algorithm picks image crossover region.
The embodiment of the present invention uses monocular pinhole camera model and fisheye camera model (Omnidirectio nal respectively
Camera Model) image of fisheye camera acquisition is handled, by respectively treated image effectively as
Vegetarian refreshments compares it is found that the coverage area for the pixel that is activated in the image of fisheye camera model treatment is larger and quantity is more.
Therefore, the embodiment of the present invention preferably uses fisheye camera model, so as to without the fish-eye image acquired to fish-eye camera
Distortion correction (i.e. it is not necessary that fish-eye image is converted to plan view) is done, the content of flake picture directly can be applied to vision SLAM
In algorithm, the consistency and quality of maximal end point cloud are improved.Also, field angle can be increased using fisheye camera model, mentioned
Rise matching precision.
If choosing fisheye camera model as default camera model, then when executing above step S102, in order to
The image directly acquired using fisheye camera avoids the picture by the image flame detection of fish-eye camera acquisition at non-flake, to prevent
Only field of view reduces.The embodiment of the present invention can first obtain two images that two fisheye cameras acquire respectively, then by two
A image is separately input into fisheye camera model, and then in image after fisheye camera model treatment, from treated image
The corresponding image of any fisheye camera of middle selection, and choose from selection treated image multiple pixels of arbitrary objects
Point.
Under normal conditions, the distance between two neighboring fisheye camera being installed on vehicle farther out, and towards difference, is adopted
The angle for collecting image is different.For example, the two distance is farther out for preposition fisheye camera and side fisheye camera adjacent on automobile
And it is respectively facing front side, left/right.In the corresponding image of any fisheye camera of selection, need according to practical application scene
It makes a return journey and determines that the corresponding image of which fisheye camera of selection is more particularly suitable.
For example, in automatic parking scene, since driver can be intuitive to see the scene of vehicle front, and can not
Intuitively see the scene of automobile side.It is thereby possible to select the visual field of side fisheye camera is target area, that is, choose side
Fisheye camera corresponding image detects the static-obstacle thing of automobile side when facilitating driver to search parking stall.
Step S104 is seen above, in an embodiment of the present invention, image crossover region will be located in the pixel of selection
During pixel is converted to the pixel of corresponding binocular SLAM algorithm, it can come to carry out pixel in conjunction with image coordinate system
Conversion, the process prior art specifically converted may be implemented, and details are not described herein again.
In an alternative embodiment of the invention, the pixel after converting in image crossover region is calculated based on binocular SLAM algorithm
During three-dimensional coordinate, then the parallax range that can first obtain between the corresponding camera of acquisition image measured in advance uses
Binocular SLAM algorithm and the three-dimensional coordinate that the pixel after converting in image crossover region is calculated according to the parallax range between camera.
Three-dimensional coordinate information in the embodiment further comprises the depth information of pixel.
In this embodiment, if binocular is formed using the fisheye camera for being mounted on vehicle front and side, it is possible in advance
The parallax range between vehicle front and the fisheye camera of side is first measured, and then using binocular SLAM algorithm and according to two fishes
Parallax range between eye camera calculates the three-dimensional coordinate of the pixel after converting in image crossover region.
After being initialized to monocular SLAM algorithm, after initialization can also be assisted using binocular SLAM algorithm
Monocular SLAM algorithm faster, more accurately obtains the clinkering point in acquisition image in real time, and according to the non-overlapping area of image and overlaps
The three-dimensional coordinate of pixel in area generates the point cloud of object.The embodiment of the invention also provides a kind of binocular point Yun Sheng as a result,
At method, this method mainly includes the following steps 1 to step 5.
Step 1, from the pixel in the monocular SLAM algorithm picks image crossover region in acquisition image according to initialization.
In this step, image crossover region is the content overlapping region of two images acquired by different angle.Also,
Since the camera being installed on vehicle can acquire image in real time, this method can be selected from the image acquired in real time
Take the pixel in image crossover region.
In this step, the multiple pixels chosen from acquisition image refer to clinkering point, are preliminary from image
After selecting pixel, three-dimensional can be calculated subsequent in the stable pixel further screened, clinkering point
Coordinate information.
Step 2, the pixel that the pixel of selection is converted to corresponding binocular SLAM algorithm, based on binocular SLAM algorithm
Three-dimensional coordinate of the nomogram as the pixel after being converted in crossover region.
In the step, when the pixel of selection to be converted to the pixel of corresponding binocular SLAM algorithm, it can also be based on
The pixel of selection is converted to the pixel of corresponding binocular SLAM algorithm by image coordinate system.
Step 3, the coordinate information that the three-dimensional coordinate of pixel is converted to camera coordinates system utilizes the monocular after initialization
SLAM algorithm calculates the pose variation from the camera of an angle acquisition image based on the coordinate information of camera coordinates system.
Step 4, the three-dimensional coordinate of the pixel in the non-overlapping area of image is calculated according to pose variation, wherein the non-friendship of image
Folded area is the content not overlapping region of two images acquired by different angle.
Step 5, the three-dimensional coordinate according to the non-overlapping area of image and the pixel in crossover region generates point cloud.
The embodiment of the present invention constitutes biocular systems by the camera installed in vehicle different location angle, utilizes biocular systems
In the binocular SLAM algorithm pixel that calculates image crossover region three-dimensional coordinate (i.e. world coordinate system coordinate) after, Ke Yigen
The three-dimensional coordinate of the pixel of image crossover region is converted into a camera according to seat value target conversion regime between different coordinates
Camera coordinates system coordinate information, and then according to monocular SLAM algorithm and based on the coordinate for the camera coordinates system being converted to believe
Breath calculates the three-dimensional coordinate of the pixel in the non-overlapping area of image, and the example scheme can be according to the figure of acquisition image as a result,
As the reckoning of the pixel information progress scale in overlapping region, the pixel information in the non-overlapping region of image is obtained.Due to double
Mesh SLAM algorithm has many advantages, such as not need that initialization, can to obtain the more accurate dimensional information of object, stability more preferable, because
This, can assist monocular SLAM algorithm faster, more accurately to obtain the clinkering point in acquisition image using binocular SLAM algorithm, from
And improve the stability that a cloud generates and the accuracy that the true scale of object obtains.
Step 1 is seen above, in an embodiment of the present invention, the process for choosing the pixel in image crossover region can join
According to foregoing embodiments, after the image for getting camera acquisition, the image that will acquire is input in default camera model, by phase
Machine model handles image, and then overlapping according to the monocular SLAM algorithm picks image of initialization from treated image
Pixel in area.
Step 4 is seen above, in an embodiment of the present invention, the pose variation according to camera calculates in the non-overlapping area of image
The detailed process of three-dimensional coordinate of pixel be, firstly, passing through the monocular SLAM algorithm after initialization and the position according to camera
Appearance variation, calculates shift length of the camera between any two different moments.Then, pass through the monocular SLAM after initialization
Algorithm and the three-dimensional coordinate that the pixel in the non-overlapping area of image is calculated according to shift length.
In this embodiment, the camera of pose variation refers to the corresponding camera of an image of selected pixels point.Therefore,
Calculate the corresponding camera of image that the corresponding camera of shift length between any two different moments is selected pixels point.Example
Such as, the camera of any two different angle is separately mounted to the front and side of vehicle, and foregoing embodiments are adopted from the camera of side
The pixel that arbitrary objects in image crossover region are chosen in the image of collection, in this embodiment it is possible to pass through monocular SLAM algorithm
And change according to the pose of side camera, calculate shift length of the side camera between any two different moments.
In the embodiment, shift length of the camera between any two different moments is it also will be understood that at the camera
Parallax range between any two different moments.
In an embodiment of the present invention, the pose variation of camera may include the spin matrix and/or translation matrix of camera.
Based on the same inventive concept, the embodiment of the invention also provides a kind of initialization systems of monocular SLAM algorithm.Fig. 3
Show the structural schematic diagram of the initialization system of monocular SLAM algorithm according to an embodiment of the invention.Referring to Fig. 3, monocular
The initialization system 300 of SLAM algorithm includes monocular subsystem 320, conversion subsystem 330, binocular subsystem 340.
Now introduce the initialization system 300 of the monocular SLAM algorithm of the embodiment of the present invention each composition or device function with
And the connection relationship between each section:
Monocular subsystem 320 chooses the pixel in image crossover region, wherein image crossover region is from acquisition image
Pass through the content overlapping region for two images that different angle acquires；
Conversion subsystem 330 is coupled with monocular subsystem 320, and the pixel of selection is converted to suitable for binocular subsystem
The pixel of system 340 and corresponding binocular SLAM algorithm；
Binocular subsystem 340 is coupled with conversion subsystem 330, calculates the transfer of image crossover region based on binocular SLAM algorithm
The three-dimensional coordinate of pixel after changing；
The three-dimensional coordinate of pixel is converted to the camera coordinates system suitable for monocular subsystem 320 by conversion subsystem 330
Coordinate information；
Monocular subsystem 320, the coordinate information according to camera coordinates system initialize monocular SLAM algorithm.
In an embodiment of the present invention, monocular subsystem 320 can also obtain acquisition image, and the image that will acquire is defeated
Enter into default camera model, when image is after camera model is handled, can be calculated from treated image according to monocular SLAM
Method chooses the pixel in image crossover region.
In an embodiment of the present invention, the pixel of selection is being converted to corresponding binocular SLAM calculation by conversion subsystem 330
When the pixel of method, the pixel of selection can be converted to the pixel of corresponding binocular SLAM algorithm based on image coordinate system.
In an embodiment of the present invention, binocular subsystem 340 can first obtain the corresponding phase of acquisition image measured in advance
Parallax range between machine, is then based on binocular SLAM algorithm and according to the parallax range between camera, calculates in image crossover region
The three-dimensional coordinate of pixel after conversion.
In an embodiment of the present invention, when monocular subsystem 320 according to camera coordinates system coordinate information to monocular SLAM
After algorithm is initialized, corresponding monocular subsystem 320, binocular subsystem 340, conversion subsystem 330 be can also be performed
The process of object point cloud is generated, specific process is as follows:
Monocular subsystem 320, out of, monocular SLAM algorithm picks image crossover region in acquisition image according to initialization
Pixel.
Wherein, image crossover region is the content overlapping region of two images acquired by the camera of different angle.Also,
Since the camera being installed on vehicle can acquire image in real time, monocular subsystem 320 can be from the figure acquired in real time
The pixel in image crossover region is chosen as in.
The pixel of selection is converted to suitable for binocular subsystem 340 and corresponding binocular SLAM is calculated by conversion subsystem 330
The pixel of method.
Binocular subsystem 340 calculates the three-dimensional of the pixel after converting in image crossover region based on binocular SLAM algorithm and sits
Mark.
The three-dimensional coordinate of pixel is converted to the camera coordinates system suitable for monocular subsystem 320 by conversion subsystem 330
Coordinate information.
Monocular subsystem 320 is calculated based on the coordinate information of camera coordinates system from an angle using monocular SLAM algorithm
The pose variation of the camera of image is acquired, and changes the three-dimensional for calculating the arbitrary objects pixel in the non-overlapping area of image according to pose
Coordinate.Wherein, the non-overlapping area of image is the content not overlapping region of two images acquired by different angle.
Monocular subsystem 320, the three-dimensional coordinate according to the non-overlapping area of image and the pixel in crossover region generate point cloud.
In an embodiment of the present invention, monocular subsystem 320 is calculating the picture in the non-overlapping area of image according to pose variation
When the three-dimensional coordinate of vegetarian refreshments, it can change by the monocular SLAM algorithm after initialization and according to pose, calculate from an angle
Acquire shift length of the camera of image between any two different moments.Then, it is calculated by the monocular SLAM after initialization
Method calculates the three-dimensional coordinate of the pixel in the non-overlapping area of image according to shift length.
In an embodiment of the present invention, the pose variation of camera includes the spin matrix and/or translation matrix of camera.
The embodiment of the invention also provides a kind of computer storage medium, computer storage medium is stored with computer program
Code causes calculating equipment to execute the list in any embodiment above when computer program code is run on the computing device
The initial method of mesh SLAM algorithm.
According to the combination of any one above-mentioned preferred embodiment or multiple preferred embodiments, the embodiment of the present invention can reach
It is following the utility model has the advantages that
In embodiments of the present invention, the pixel in image crossover region is being chosen from acquisition image, and by the picture of selection
After vegetarian refreshments is converted to the pixel of corresponding binocular SLAM algorithm, calculated based on binocular SLAM algorithm after being converted in image crossover region
The three-dimensional coordinate of pixel, and the three-dimensional coordinate of pixel is converted to the coordinate information of camera coordinates system, and then can foundation
The coordinate information of camera coordinates system initializes monocular SLAM algorithm.The embodiment of the present invention can be by according to double as a result,
Mesh SLAM algorithm initializes to assist monocular SLAM algorithm to realize, so as to make the corresponding monocular system of monocular SLAM algorithm more
Fastly, it preferably realizes initialization, improves the stability and computational accuracy of monocular SLAM algorithm.
It is apparent to those skilled in the art that the specific work of the system of foregoing description, device and unit
Make process, can refer to corresponding processes in the foregoing method embodiment, for brevity, does not repeat separately herein.
In addition, each functional unit in each embodiment of the present invention can be physically independent, can also two or
More than two functional units integrate, and can be all integrated in a processing unit with all functional units.It is above-mentioned integrated
Functional unit both can take the form of hardware realization, can also be realized in the form of software or firmware.
Those of ordinary skill in the art will appreciate that: if the integrated functional unit is realized and is made in the form of software
It is independent product when selling or using, can store in a computer readable storage medium.Based on this understanding,
Technical solution of the present invention is substantially or all or part of the technical solution can be embodied in the form of software products,
The computer software product is stored in a storage medium comprising some instructions, with so that calculating equipment (such as
Personal computer, server or network equipment etc.) various embodiments of the present invention the method is executed when running described instruction
All or part of the steps.And storage medium above-mentioned includes: USB flash disk, mobile hard disk, read-only memory (ROM), random access memory
Device (RAM), the various media that can store program code such as magnetic or disk.
Alternatively, realizing that all or part of the steps of preceding method embodiment can be (all by the relevant hardware of program instruction
Such as personal computer, the calculating equipment of server or network equipment etc.) it completes, described program instruction can store in one
In computer-readable storage medium, when described program instruction is executed by the processor of calculating equipment, the calculating equipment is held
The all or part of the steps of row various embodiments of the present invention the method.
Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention., rather than its limitations；To the greatest extent
Present invention has been described in detail with reference to the aforementioned embodiments for pipe, those skilled in the art should understand that: at this
Within the spirit and principle of invention, it is still possible to modify the technical solutions described in the foregoing embodiments or right
Some or all of the technical features are equivalently replaced；And these are modified or replaceed, and do not make corresponding technical solution de-
From protection scope of the present invention.
Claims (9)
1. a kind of initial method of monocular SLAM algorithm, comprising:
The pixel in image crossover region is chosen from acquisition image, wherein described image crossover region is to adopt by different angle
The content overlapping region of two images of collection；
The pixel of selection is converted to the pixel of corresponding binocular SLAM algorithm, image is calculated based on the binocular SLAM algorithm
The three-dimensional coordinate of pixel after being converted in crossover region；
The three-dimensional coordinate of the pixel is converted to the coordinate information of camera coordinates system；
Coordinate information according to the camera coordinates system initializes monocular SLAM algorithm.
2. according to the method described in claim 1, wherein, the pixel in image crossover region is chosen from acquisition image, comprising:
Acquisition image is obtained, the image that will acquire is input in default camera model；
Described image is after camera model processing, from overlapping according to monocular SLAM algorithm picks image in treated image
Pixel in area.
3. method according to claim 1 or 2, wherein the pixel of selection is converted to corresponding binocular SLAM algorithm
Pixel, comprising:
The pixel of selection is converted to the pixel of corresponding binocular SLAM algorithm based on image coordinate system.
4. method according to claim 1 or 2, wherein calculate the transfer of image crossover region based on the binocular SLAM algorithm
The three-dimensional coordinate of pixel after changing, comprising:
Obtain the parallax range between the corresponding camera of acquisition image measured in advance；
Based on the binocular SLAM algorithm and according to the parallax range between the camera, calculate after being converted in image crossover region
The three-dimensional coordinate of pixel.
5. according to the method described in claim 1, wherein, the coordinate information according to the camera coordinates system is to monocular SLAM algorithm
After being initialized, further includes:
Pixel out of monocular SLAM algorithm picks image crossover region in acquisition image according to initialization, wherein the figure
As the content overlapping region that crossover region is two images acquired by different angle；
The pixel of selection is converted to the pixel of corresponding binocular SLAM algorithm, image is calculated based on the binocular SLAM algorithm
The three-dimensional coordinate of pixel after being converted in crossover region；
The coordinate information that the three-dimensional coordinate of the pixel is converted to camera coordinates system utilizes the monocular after the initialization
SLAM algorithm calculates the pose variation from the camera of an angle acquisition image based on the coordinate information of the camera coordinates system；
Change the three-dimensional coordinate for calculating the pixel in the non-overlapping area of image according to the pose, wherein described image is non-overlapping
Area is the content not overlapping region of two images acquired by different angle；
Three-dimensional coordinate according to the non-overlapping area of image and the pixel in crossover region generates point cloud.
6. according to the method described in claim 5, wherein, changing the pixel calculated in the non-overlapping area of image according to the pose
Three-dimensional coordinate, comprising:
Change by the monocular SLAM algorithm after initialization and according to the pose, calculates the camera from an angle acquisition image
Shift length between any two different moments；
By initialization after monocular SLAM algorithm according to the shift length calculate the non-overlapping area of image in pixel three
Tie up coordinate.
7. method according to claim 5 or 6, wherein
The pose of the camera changes, comprising: the spin matrix and/or translation matrix of the camera.
8. a kind of initialization system of monocular SLAM algorithm, including monocular subsystem, binocular subsystem, conversion subsystem, wherein
The monocular subsystem chooses the pixel in image crossover region, wherein described image crossover region is from acquisition image
Pass through the content overlapping region for two images that different angle acquires；
The pixel of selection is converted to suitable for the binocular subsystem and corresponding binocular SLAM is calculated by the conversion subsystem
The pixel of method；
The binocular subsystem calculates the three-dimensional of the pixel after converting in image crossover region based on the binocular SLAM algorithm and sits
Mark；
The three-dimensional coordinate of the pixel is converted to the camera coordinates suitable for the monocular subsystem by the conversion subsystem
The coordinate information of system；
The monocular subsystem, the coordinate information according to the camera coordinates system initialize monocular SLAM algorithm.
9. a kind of computer storage medium, the computer storage medium is stored with computer program code, when the computer
When program code is run on the computing device, the calculating equipment perform claim is caused to require the described in any item monoculars of 1-7
The initial method of SLAM algorithm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910319610.0A CN110060202B (en) | 2019-04-19 | 2019-04-19 | Monocular SLAM algorithm initialization method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910319610.0A CN110060202B (en) | 2019-04-19 | 2019-04-19 | Monocular SLAM algorithm initialization method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110060202A true CN110060202A (en) | 2019-07-26 |
| CN110060202B CN110060202B (en) | 2021-06-08 |
Family
ID=67319832
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910319610.0A Active CN110060202B (en) | 2019-04-19 | 2019-04-19 | Monocular SLAM algorithm initialization method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110060202B (en) |
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111311684A (en) * | 2020-04-01 | 2020-06-19 | 亮风台（上海）信息科技有限公司 | Method and equipment for initializing SLAM |
| CN112067007A (en) * | 2020-11-12 | 2020-12-11 | 湖北亿咖通科技有限公司 | Map generation method, computer storage medium, and electronic device |
| CN112348868A (en) * | 2020-11-06 | 2021-02-09 | 养哇(南京)科技有限公司 | Method and system for recovering monocular SLAM scale through detection and calibration |
| CN112380963A (en) * | 2020-11-11 | 2021-02-19 | 东软睿驰汽车技术（沈阳）有限公司 | Depth information determination method and device based on panoramic all-round looking system |
| WO2021128314A1 (en) * | 2019-12-27 | 2021-07-01 | 深圳市大疆创新科技有限公司 | Image processing method and device, image processing system and storage medium |
| CN113763252A (en) * | 2021-09-16 | 2021-12-07 | 中国电子科技集团公司第五十四研究所 | Method for converting geodetic coordinate system and SLAM coordinate system for unmanned aerial vehicle |
| US11403069B2 (en) | 2017-07-24 | 2022-08-02 | Tesla, Inc. | Accelerated mathematical engine |
| US11409692B2 (en) | 2017-07-24 | 2022-08-09 | Tesla, Inc. | Vector computational unit |
| US11487288B2 (en) | 2017-03-23 | 2022-11-01 | Tesla, Inc. | Data synthesis for autonomous control systems |
| US11537811B2 (en) | 2018-12-04 | 2022-12-27 | Tesla, Inc. | Enhanced object detection for autonomous vehicles based on field view |
| US11561791B2 (en) | 2018-02-01 | 2023-01-24 | Tesla, Inc. | Vector computational unit receiving data elements in parallel from a last row of a computational array |
| US11562231B2 (en) | 2018-09-03 | 2023-01-24 | Tesla, Inc. | Neural networks for embedded devices |
| US11567514B2 (en) | 2019-02-11 | 2023-01-31 | Tesla, Inc. | Autonomous and user controlled vehicle summon to a target |
| US11610117B2 (en) | 2018-12-27 | 2023-03-21 | Tesla, Inc. | System and method for adapting a neural network model on a hardware platform |
| US11636333B2 (en) | 2018-07-26 | 2023-04-25 | Tesla, Inc. | Optimizing neural network structures for embedded systems |
| US11665108B2 (en) | 2018-10-25 | 2023-05-30 | Tesla, Inc. | QoS manager for system on a chip communications |
| US11681649B2 (en) | 2017-07-24 | 2023-06-20 | Tesla, Inc. | Computational array microprocessor system using non-consecutive data formatting |
| US11734562B2 (en) | 2018-06-20 | 2023-08-22 | Tesla, Inc. | Data pipeline and deep learning system for autonomous driving |
| US11748620B2 (en) | 2019-02-01 | 2023-09-05 | Tesla, Inc. | Generating ground truth for machine learning from time series elements |
| US11790664B2 (en) | 2019-02-19 | 2023-10-17 | Tesla, Inc. | Estimating object properties using visual image data |
| US11816585B2 (en) | 2018-12-03 | 2023-11-14 | Tesla, Inc. | Machine learning models operating at different frequencies for autonomous vehicles |
| US11841434B2 (en) | 2018-07-20 | 2023-12-12 | Tesla, Inc. | Annotation cross-labeling for autonomous control systems |
| CN117434570A (en) * | 2023-12-20 | 2024-01-23 | 绘见科技(深圳)有限公司 | Visual measurement method, measurement device and storage medium for coordinates |
| US11893393B2 (en) | 2017-07-24 | 2024-02-06 | Tesla, Inc. | Computational array microprocessor system with hardware arbiter managing memory requests |
| US11893774B2 (en) | 2018-10-11 | 2024-02-06 | Tesla, Inc. | Systems and methods for training machine models with augmented data |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110125323A1 (en) * | 2009-11-06 | 2011-05-26 | Evolution Robotics, Inc. | Localization by learning of wave-signal distributions |
| WO2016021252A1 (en) * | 2014-08-05 | 2016-02-11 | ソニー株式会社 | Information processing device, information processing method, and image display system |
| CN106570913A (en) * | 2016-11-04 | 2017-04-19 | 上海玄彩美科网络科技有限公司 | Feature based monocular SLAM (Simultaneous Localization and Mapping) quick initialization method |
| CN106595639A (en) * | 2016-12-27 | 2017-04-26 | 纳恩博（北京）科技有限公司 | Positioning system and positioning method and device thereof and robot |
| CN107657640A (en) * | 2017-09-30 | 2018-02-02 | 南京大典科技有限公司 | Intelligent patrol inspection management method based on ORB SLAM |
| CN109493378A (en) * | 2018-10-29 | 2019-03-19 | 宁波研新工业科技有限公司 | A kind of measuring for verticality method combined based on monocular vision with binocular vision |
-
2019
- 2019-04-19 CN CN201910319610.0A patent/CN110060202B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110125323A1 (en) * | 2009-11-06 | 2011-05-26 | Evolution Robotics, Inc. | Localization by learning of wave-signal distributions |
| WO2016021252A1 (en) * | 2014-08-05 | 2016-02-11 | ソニー株式会社 | Information processing device, information processing method, and image display system |
| CN106570913A (en) * | 2016-11-04 | 2017-04-19 | 上海玄彩美科网络科技有限公司 | Feature based monocular SLAM (Simultaneous Localization and Mapping) quick initialization method |
| CN106595639A (en) * | 2016-12-27 | 2017-04-26 | 纳恩博（北京）科技有限公司 | Positioning system and positioning method and device thereof and robot |
| CN107657640A (en) * | 2017-09-30 | 2018-02-02 | 南京大典科技有限公司 | Intelligent patrol inspection management method based on ORB SLAM |
| CN109493378A (en) * | 2018-10-29 | 2019-03-19 | 宁波研新工业科技有限公司 | A kind of measuring for verticality method combined based on monocular vision with binocular vision |
Non-Patent Citations (2)
| Title |
|---|
| A. SHAOBO 等: "Fast initialization for feature-based monocular slam", 《2017 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP)》 * |
| 王晓华 等: "室内环境下结合里程计的双目视觉SLAM研究", 《西安理工大学学报》 * |
Cited By (30)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11487288B2 (en) | 2017-03-23 | 2022-11-01 | Tesla, Inc. | Data synthesis for autonomous control systems |
| US11403069B2 (en) | 2017-07-24 | 2022-08-02 | Tesla, Inc. | Accelerated mathematical engine |
| US11681649B2 (en) | 2017-07-24 | 2023-06-20 | Tesla, Inc. | Computational array microprocessor system using non-consecutive data formatting |
| US11893393B2 (en) | 2017-07-24 | 2024-02-06 | Tesla, Inc. | Computational array microprocessor system with hardware arbiter managing memory requests |
| US11409692B2 (en) | 2017-07-24 | 2022-08-09 | Tesla, Inc. | Vector computational unit |
| US11561791B2 (en) | 2018-02-01 | 2023-01-24 | Tesla, Inc. | Vector computational unit receiving data elements in parallel from a last row of a computational array |
| US11797304B2 (en) | 2018-02-01 | 2023-10-24 | Tesla, Inc. | Instruction set architecture for a vector computational unit |
| US11734562B2 (en) | 2018-06-20 | 2023-08-22 | Tesla, Inc. | Data pipeline and deep learning system for autonomous driving |
| US11841434B2 (en) | 2018-07-20 | 2023-12-12 | Tesla, Inc. | Annotation cross-labeling for autonomous control systems |
| US11636333B2 (en) | 2018-07-26 | 2023-04-25 | Tesla, Inc. | Optimizing neural network structures for embedded systems |
| US11562231B2 (en) | 2018-09-03 | 2023-01-24 | Tesla, Inc. | Neural networks for embedded devices |
| US11893774B2 (en) | 2018-10-11 | 2024-02-06 | Tesla, Inc. | Systems and methods for training machine models with augmented data |
| US11665108B2 (en) | 2018-10-25 | 2023-05-30 | Tesla, Inc. | QoS manager for system on a chip communications |
| US11816585B2 (en) | 2018-12-03 | 2023-11-14 | Tesla, Inc. | Machine learning models operating at different frequencies for autonomous vehicles |
| US11537811B2 (en) | 2018-12-04 | 2022-12-27 | Tesla, Inc. | Enhanced object detection for autonomous vehicles based on field view |
| US11908171B2 (en) | 2018-12-04 | 2024-02-20 | Tesla, Inc. | Enhanced object detection for autonomous vehicles based on field view |
| US11610117B2 (en) | 2018-12-27 | 2023-03-21 | Tesla, Inc. | System and method for adapting a neural network model on a hardware platform |
| US11748620B2 (en) | 2019-02-01 | 2023-09-05 | Tesla, Inc. | Generating ground truth for machine learning from time series elements |
| US11567514B2 (en) | 2019-02-11 | 2023-01-31 | Tesla, Inc. | Autonomous and user controlled vehicle summon to a target |
| US11790664B2 (en) | 2019-02-19 | 2023-10-17 | Tesla, Inc. | Estimating object properties using visual image data |
| WO2021128314A1 (en) * | 2019-12-27 | 2021-07-01 | 深圳市大疆创新科技有限公司 | Image processing method and device, image processing system and storage medium |
| CN111311684A (en) * | 2020-04-01 | 2020-06-19 | 亮风台（上海）信息科技有限公司 | Method and equipment for initializing SLAM |
| CN112348868A (en) * | 2020-11-06 | 2021-02-09 | 养哇(南京)科技有限公司 | Method and system for recovering monocular SLAM scale through detection and calibration |
| CN112380963A (en) * | 2020-11-11 | 2021-02-19 | 东软睿驰汽车技术（沈阳）有限公司 | Depth information determination method and device based on panoramic all-round looking system |
| CN112067007B (en) * | 2020-11-12 | 2021-01-29 | 湖北亿咖通科技有限公司 | Map generation method, computer storage medium, and electronic device |
| CN112067007A (en) * | 2020-11-12 | 2020-12-11 | 湖北亿咖通科技有限公司 | Map generation method, computer storage medium, and electronic device |
| CN113763252B (en) * | 2021-09-16 | 2022-12-09 | 中国电子科技集团公司第五十四研究所 | Geodetic coordinate system and SLAM coordinate system conversion method for unmanned aerial vehicle |
| CN113763252A (en) * | 2021-09-16 | 2021-12-07 | 中国电子科技集团公司第五十四研究所 | Method for converting geodetic coordinate system and SLAM coordinate system for unmanned aerial vehicle |
| CN117434570A (en) * | 2023-12-20 | 2024-01-23 | 绘见科技(深圳)有限公司 | Visual measurement method, measurement device and storage medium for coordinates |
| CN117434570B (en) * | 2023-12-20 | 2024-02-27 | 绘见科技(深圳)有限公司 | Visual measurement method, measurement device and storage medium for coordinates |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110060202B (en) | 2021-06-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110060202A (en) | A kind of initial method and system of monocular SLAM algorithm | |
| CN110910453B (en) | Vehicle pose estimation method and system based on non-overlapping view field multi-camera system | |
| CN105957007B (en) | Image split-joint method based on characteristic point plane similarity | |
| KR101475583B1 (en) | Image generation device and operation support system | |
| JP6310652B2 (en) | Video display system, video composition device, and video composition method | |
| US11407363B2 (en) | Method for calculating a tow hitch position | |
| CN103155551B (en) | Video generation device and operations support system | |
| CN106997579B (en) | Image splicing method and device | |
| CN106803899B (en) | Merge the method and apparatus of image | |
| CN106856000B (en) | Seamless splicing processing method and system for vehicle-mounted panoramic image | |
| CN105745122A (en) | Driver assistance system for displaying surroundings of a vehicle | |
| JP6024581B2 (en) | Image processing apparatus for vehicle | |
| CN105894448B (en) | The generation method of mask matrix, the synthetic method for image of parking and device | |
| CN102855649A (en) | Method for splicing high-definition image panorama of high-pressure rod tower on basis of ORB (Object Request Broker) feature point | |
| CN107843251A (en) | The position and orientation estimation method of mobile robot | |
| CN102291541A (en) | Virtual synthesis display system of vehicle | |
| CN104966318A (en) | A reality augmenting method having image superposition and image special effect functions | |
| CN110245199B (en) | Method for fusing large-dip-angle video and 2D map | |
| CN107798702A (en) | A kind of realtime graphic stacking method and device for augmented reality | |
| JP2011259152A (en) | Driving assistance device | |
| WO2016129612A1 (en) | Method for reconstructing a three-dimensional (3d) scene | |
| CN110084851A (en) | A kind of binocular point cloud generation method and system | |
| GB2513703B (en) | Method and apparatus for three-dimensional imaging of at least a partial region of a vehicle environment | |
| CN107209930A (en) | Look around image stability method and device | |
| CN115272494A (en) | Calibration method and device for camera and inertial measurement unit and computer equipment |
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 | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20220323 Address after: 430051 No. b1336, chuanggu startup area, taizihu cultural Digital Creative Industry Park, No. 18, Shenlong Avenue, Wuhan Economic and Technological Development Zone, Wuhan, Hubei Province Patentee after: Yikatong (Hubei) Technology Co.,Ltd. Address before: No.c101, chuanggu start up area, taizihu cultural Digital Industrial Park, No.18 Shenlong Avenue, Wuhan Economic Development Zone, Hubei Province Patentee before: HUBEI ECARX TECHNOLOGY Co.,Ltd. |
|
| TR01 | Transfer of patent right |