CN109257527A - The acquisition device and method of image mosaic data - Google Patents
The acquisition device and method of image mosaic data Download PDFInfo
- Publication number
- CN109257527A CN109257527A CN201811176600.8A CN201811176600A CN109257527A CN 109257527 A CN109257527 A CN 109257527A CN 201811176600 A CN201811176600 A CN 201811176600A CN 109257527 A CN109257527 A CN 109257527A
- Authority
- CN
- China
- Prior art keywords
- image
- camera
- acquisition component
- cameras
- plane mirror
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 41
- 230000003287 optical effect Effects 0.000 claims description 45
- 238000003384 imaging method Methods 0.000 claims description 8
- 230000001537 neural effect Effects 0.000 claims 1
- 238000010586 diagram Methods 0.000 description 14
- 230000000007 visual effect Effects 0.000 description 9
- 238000013135 deep learning Methods 0.000 description 5
- 238000013527 convolutional neural network Methods 0.000 description 3
- 239000012141 concentrate Substances 0.000 description 2
- 230000007812 deficiency Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 239000004744 fabric Substances 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 description 1
- 239000010931 gold Substances 0.000 description 1
- 229910052737 gold Inorganic materials 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000002203 pretreatment Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/50—Constructional details
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/50—Constructional details
- H04N23/55—Optical parts specially adapted for electronic image sensors; Mounting thereof
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/90—Arrangement of cameras or camera modules, e.g. multiple cameras in TV studios or sports stadiums
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N5/00—Details of television systems
- H04N5/222—Studio circuitry; Studio devices; Studio equipment
- H04N5/262—Studio circuits, e.g. for mixing, switching-over, change of character of image, other special effects ; Cameras specially adapted for the electronic generation of special effects
- H04N5/265—Mixing
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Studio Devices (AREA)
Abstract
This disclosure relates to a kind of acquisition device and method of image mosaic data.The image collecting device includes reflection subassembly and acquisition component, and reflection subassembly includes at least three plane mirrors, and at least three plane mirrors are bonded into inverted pyramid type；Acquisition component, including at least three cameras, at least three cameras correspondingly arrange that acquisition component is for acquiring multiple images of the horizontal scene after reflection subassembly reflects at least three plane mirrors.The image collecting device that the embodiment of the present disclosure provides, true value image seamless and without ghost image can be obtained by the device and there are the plurality of input images of parallax, the problems such as obtaining difficulty for image mosaic data set shortage, image mosaic true value image provides feasible solution, to meet the use demand of image mosaic data.
Description
Technical field
This disclosure relates to the acquisition device and side of technical field of computer vision more particularly to a kind of image mosaic data
Method.
Background technique
Current depth study development is maked rapid progress, from Text region, speech recognition, image classification, object detection to image
Semantic segmentation etc..With deep learning network proposition and continuously improve, deep learning be used to generate nature high definition true to nature
Image.But current merging algorithm for images can not be spliced into true value image of the satisfied panorama sketch as deep learning network,
Exist simultaneously that will lead to panorama sketch after the multiple image splicing of parallax fuzzy or have ghost image.Therefore, how to generate clearly and without weight
The panorama sketch of shadow is a problem to be solved.
Summary of the invention
In view of this, the present disclosure proposes the acquisition devices and method of a kind of image mosaic data.
According to the one side of the disclosure, a kind of image collecting device is provided.
In one possible implementation, described device includes:
Reflection subassembly, including at least three plane mirrors, at least three plane mirror are bonded into inverted pyramid type；
Acquisition component, including at least three cameras, at least three camera and at least three plane mirrors phase
It accordingly arranges, the acquisition component is used to acquire multiple images of the horizontal scene after reflection subassembly reflection.
In one possible implementation, each plane mirror of the reflection subassembly and the angle of horizontal plane remain
One angle, and the shape of each plane mirror is isosceles triangle,
Wherein, the first angle is greater than 0 degree and less than 90 degree.
In one possible implementation, each camera of the acquisition component is located at below corresponding plane mirror
Middle position, and the height of each camera and at a distance from the central axis of the reflection subassembly it is identical.
In one possible implementation, focal centre length altogether is at least three cameras of the acquisition component
When, it is the true value image in seamless and non-overlapping region to the spliced panorama sketch of described multiple images,
Wherein, at least three take the photograph described in when the focal centre length altogether is the virtual optical center coincidence of at least three camera
As the height of head, the virtual optical center of each camera be the optical center of each camera on corresponding plane mirror at image position
It sets.
According to another aspect of the present disclosure, a kind of image-pickup method is provided.
In one possible implementation, the method is applied to above-mentioned image collecting device, described image acquisition
Device includes reflection subassembly and acquisition component, which comprises
Adjust position and the shooting direction of at least three cameras of the acquisition component；
Multiple images of the horizontal scene after reflection subassembly reflection are acquired by the acquisition component.
In one possible implementation, the position and shooting side of at least three cameras of the acquisition component are adjusted
To, comprising:
The position of the acquisition component is adjusted, so that at least three cameras of the acquisition component are in optical center height altogether
Degree, wherein at least three cameras described in when the focal centre length altogether is the virtual optical center coincidence of at least three camera
Height, the virtual optical center of each camera is imaging position of the optical center of each camera on corresponding plane mirror,
Wherein, the method also includes:
Splicing is carried out to described multiple images, obtains the true value image in seamless and non-overlapping region.
In one possible implementation, position and the shooting direction of the acquisition component are adjusted, comprising:
The position of the acquisition component is adjusted, so that at least three cameras of the acquisition component are higher than optical center height altogether
Degree,
Wherein, the method also includes: using the multiple images with overlapping region as input picture.
In one possible implementation, the method also includes:
According to the true value image and multiple input pictures, training data of the building for training image splicing network
Collection.
In one possible implementation, the method also includes:
According to true value image and multiple input pictures that the training data is concentrated, training described image splices network.
In one possible implementation, described image splicing network includes convolutional neural networks, generates confrontation network
At least one of.
Image collecting device and method provided by the embodiment of the present disclosure, image collecting device include reflection subassembly and acquisition
Component, reflection subassembly include at least three plane mirrors for being bonded into inverted pyramid type, and acquisition component includes and at least three planes
At least three cameras that mirror is correspondingly arranged, for acquiring multiple images of the horizontal scene after reflection subassembly reflects.Root
According to embodiment of the disclosure, the seamless and true value image without ghost image and the plurality of input images there are parallax can be obtained, is
The problems such as image mosaic data set lacks, image mosaic true value image obtains difficulty, provides feasible solution, thus full
The use demand of sufficient image mosaic data.
According to below with reference to the accompanying drawings to detailed description of illustrative embodiments, the other feature and aspect of the disclosure will become
It is clear.
Detailed description of the invention
Comprising in the description and constituting the attached drawing of part of specification and specification together illustrates the disclosure
Exemplary embodiment, feature and aspect, and for explaining the principles of this disclosure.
Fig. 1 shows the structural schematic diagram of the image collecting device according to one embodiment of the disclosure；
Fig. 2 a, Fig. 2 b and Fig. 2 c are shown according to the plane mirror of the image collecting device of one embodiment of the disclosure and horizontal plane
The schematic diagram of angle；
Fig. 3 shows the flow chart of the image-pickup method according to one embodiment of the disclosure；
Fig. 4 a shows the schematic diagram of the acquisition scene of the image collecting device according to one embodiment of the disclosure；
Fig. 4 b shows the schematic diagram of the true value image of the image acquisition device according to one embodiment of the disclosure；
Fig. 5 a shows the schematic diagram of the acquisition scene of the image collecting device according to one embodiment of the disclosure；
Figure 5b shows that the schematic diagrames according to the input picture of the image acquisition device of one embodiment of the disclosure.
Specific embodiment
Various exemplary embodiments, feature and the aspect of the disclosure are described in detail below with reference to attached drawing.It is identical in attached drawing
Appended drawing reference indicate element functionally identical or similar.Although the various aspects of embodiment are shown in the attached drawings, remove
It non-specifically points out, it is not necessary to attached drawing drawn to scale.
Dedicated word " exemplary " means " being used as example, embodiment or illustrative " herein.Here as " exemplary "
Illustrated any embodiment should not necessarily be construed as preferred or advantageous over other embodiments.
In addition, giving numerous details in specific embodiment below to better illustrate the disclosure.
It will be appreciated by those skilled in the art that without certain details, the disclosure equally be can be implemented.In some instances, for
Method, means, element and circuit well known to those skilled in the art are not described in detail, in order to highlight the purport of the disclosure.
Fig. 1 shows the structural schematic diagram of the image collecting device according to one embodiment of the disclosure.Image collecting device includes
Reflection subassembly 1 and acquisition component 2, reflection subassembly 1 include at least three plane mirror A1, and at least three plane mirrors are bonded into down golden word
Tower；Acquisition component 2, including at least three camera A2, at least three camera A2 are corresponding at least three plane mirror A1
Ground arrangement, acquisition component 1 are used to acquire multiple images of the horizontal scene after the reflection of reflection subassembly 2.
In accordance with an embodiment of the present disclosure, image collecting device can get seamless and true value image without ghost image and there is view
The plurality of input images of difference, providing for the problems such as image mosaic data set shortage, image mosaic true value image acquisition difficulty can
Capable solution, to meet the data requirements of image mosaic.
In the present embodiment, at least three plane mirror A1 can be bonded by inverted pyramid type using the method for bonding, it can
At least three edges plane mirror A1 are polished, enable adjacent plane mirror surface A1 seamless combination.At least three plane mirror A1 mirrors
It faces outwardly, the horizontal scene of different directions can be reflected.Each camera A2 can arrange correspondingly one by one with each plane mirror A1,
The quantity of camera A2 can be identical with the quantity of plane mirror A1, so that acquisition component 2 can collect horizontal scene through reflection subassembly 1
Multiple images after reflection, can be according to the side of actual acquisition image needed to camera A2 arrangement corresponding with plane mirror A1
Formula and camera A2 and the quantity of plane mirror A1 are configured, the disclosure to this with no restriction.
In one possible implementation, each plane mirror of reflection subassembly and the angle of horizontal plane remain first jiao
Degree, and the shape of each plane mirror is isosceles triangle, wherein first angle is greater than 0 degree less than 90 degree.
Fig. 2 a, Fig. 2 b and Fig. 2 c are shown according to the plane mirror of the image collecting device of one embodiment of the disclosure and horizontal plane
The schematic diagram of angle.
In this implementation, as shown in Fig. 2 a, Fig. 2 b and Fig. 2 c, the mirror surface and water of each plane mirror A1 of reflection subassembly
The angle of plane remains first angle α, and first angle α is greater than 0 degree and less than 90 degree, so that reflection subassembly can be reflected into
Ambient level scene, if first angle is greater than 90 degree, reflection subassembly can not reflection levels scene.Wherein, Fig. 2 a is first jiao
The schematic diagram of image collecting device of the degree equal to 45 degree, Fig. 2 b are the signal of image collecting device of the first angle greater than 45 degree
Figure, Fig. 2 c are the schematic diagram of image collecting device of the first angle less than 45 degree.It can be according to the demand of actual image acquisition to
The size of one angle [alpha] is configured, the disclosure to this with no restriction.The shape of each plane mirror can be isosceles triangle, also
Can be easy to be combined into the triangular shaped of the tower of falling gold with other, the disclosure to this with no restriction.
In one possible implementation, each camera of acquisition component is located at the centre below corresponding plane mirror
Position, and the height of each camera and at a distance from the central axis of reflection subassembly it is identical.
In this implementation, each camera of acquisition component is located at the middle position below corresponding plane mirror, often
A camera is identical at a distance from reflection subassembly central axis, so that plane mirror mirror surface can cover most of visual field of camera.
The height of each camera is identical and height is adjustable, so that camera can obtain the image of different visual field sizes, and then can get
Different input pictures and true value image can need to be configured the height of camera according to actual image acquisition, this
It discloses to this with no restriction.Camera is specifically related with the first angle of reflection subassembly towards angle towards plane mirror, Ke Yigen
According to the size of the first angle of reflection subassembly and needing to camera towards being configured for actual acquisition image, the disclosure pair
This is with no restriction.
In one possible implementation, when at least three cameras of acquisition component are in focal centre length altogether, to more
Panorama sketch after a image mosaic is the true value image in seamless and non-overlapping region.
In this implementation, the optical center of each camera is each camera in corresponding plane mirror imaging position
Virtual optical center, the height of the virtual optical center of each camera each camera when being overlapped are total focal centre length.At least the three of acquisition component
It is the true value image in seamless non-overlapping region to the spliced panorama sketch of multiple images when a camera is in focal centre length altogether.
For example, it uses and inverted pyramid type is formed by four sides plane mirror (pyramidal bottom apex is the reflection group of A5)
Part is made of the image collecting device of acquisition component four cameras.Wherein, in image collecting device each plane mirror shape
For isosceles triangle, corresponding arrangement, camera are located at middle position below corresponding plane mirror to camera one by one with plane mirror,
And each camera and reflection subassembly central axis are equidistant, the mirror surface and horizontal plane of each plane mirror keep first
Angle is equal to 45 degree.As shown in Figure 2 a, the optical center A3 of each camera adjusts camera shooting in the A4 that is imaged as of corresponding plane mirror A1
The height of head realizes the total optical center of four cameras, at this time so that the virtual optical center A4 of four cameras is overlapped on a point
The height of camera is total focal centre length.When the height of camera is equal to focal centre length altogether, camera is shot straight up, is passed through
The reflection of plane mirror mirror surface can collect horizontal scene.The visual field of camera does not have an overlapping region, and four cameras it
Between without parallax, the image mosaic that can be acquired each camera by direct method is at a panorama seamless and without ghost image
Figure, and using the panorama sketch as true value image.
The process of Image Acquisition is illustrated below.
Fig. 3 shows the flow chart of the image-pickup method according to one embodiment of the disclosure.This method is applied to above-mentioned image
Acquisition device, image collecting device structure is as shown in Figure 1, include reflection subassembly and acquisition component, as shown in figure 3, this method can
To include step S11 to step S12.
In step s 11, position and the shooting direction of at least three cameras of acquisition component are adjusted；
In step s 12, multiple images of the horizontal scene after reflection subassembly reflects are acquired by acquisition component.
In accordance with an embodiment of the present disclosure, position and the shooting direction of at least three cameras of acquisition component are adjusted；Pass through
Acquisition component acquires multiple images of the horizontal scene after reflection subassembly reflects.Image is acquired by this method, can get nothing
Seam and the true value image without ghost image and the plurality of input images there are parallax, are image mosaic data deficiency, and image mosaic is true
The problems such as acquisition of value image is difficult provides feasible solution, meets the use demand of image mosaic data.
In one possible implementation, position and the shooting direction of at least three cameras of acquisition component are adjusted.
It in this implementation, can be according to the first angle that plane mirror each in reflection subassembly and horizontal plane are kept to extremely
The shooting direction of few three cameras is configured, so that camera can collect horizontal scene by the reflection of plane mirror.
The position of at least three cameras of adjustable acquisition component, so that flat mirror face can cover most of visual field of camera.
In one possible implementation, it to form reflection subassembly by four sides plane mirror, is made of and adopts four cameras
It is illustrated for the image collecting device of collection component.Wherein, the shape of each plane mirror is isosceles three in image collecting device
Angular (vertex A5), corresponding arrangement, camera are located at interposition below corresponding plane mirror to camera one by one with plane mirror
It sets, and each camera and reflection subassembly central axis are equidistant.As shown in Figure 2 a, when plane mirror each in reflection subassembly
A1 and horizontal plane keep first angle when being 45 degree, and the optical center A3 of each camera is in corresponding plane mirror in acquisition component
Imaging position is the virtual optical center A4 of each camera, and the vertex A5 of camera optical center A3 and reflection subassembly is contour, can make to image
The virtual optical center A4 of head is overlapped on a point, i.e. realization camera is total to optical center, and it is high to be known as optical center altogether for the height of camera at this time
Degree.The height of camera is adjusted, when so that the height of camera being total focal centre length, camera is shot straight up, by flat
The reflection of face mirror A1 can collect the image of horizontal scene.Adjust each camera to reflection subassembly central axis away from
From so that plane mirror can cover most of visual field of camera.
It as shown in Figure 2 b, is greater than 45 degree when plane mirror A1 each in reflection subassembly and horizontal plane keep first angle
When, the optical center A3 of each camera is the virtual optical center A4 of each camera, camera shooting in the imaging position corresponding plane mirror A1
Head optical center A3 is higher than the vertex A5 of reflection subassembly, then the virtual optical center A4 of camera can be made to be overlapped on a point, that is, realize and take the photograph
Picture head is total to optical center, and the height of camera is known as focal centre length altogether at this time.The height for adjusting camera, so that the height of camera is
Altogether when focal centre length, the direction of camera needs that horizontal scene figure could be collected by the reflection of plane mirror close to plane mirror
Picture.Adjust each camera to reflection subassembly central axis distance so that plane mirror can cover most of view of camera
It is wild.
As shown in Figure 2 c, when it is less than 45 degree that each plane mirror and horizontal plane, which keep first angle, each camera
Optical center corresponding plane mirror imaging position be each camera virtual optical center, camera optical center A3 be lower than reflection subassembly
Vertex A5, then the virtual optical center A4 of camera can be made to be overlapped on a point, i.e. realization camera is total to optical center, camera at this time
Height be known as altogether focal centre length.When the height for adjusting camera is total focal centre length, the direction of camera is needed far from plane
Mirror could collect the image of horizontal scene by the reflection of plane mirror.The position of at least three camera of acquisition component is adjusted,
So that plane mirror can cover most of visual field of camera.
In one possible implementation, the position of at least three cameras of acquisition component is adjusted in step s 11
And shooting direction, it may include:
The position of acquisition component is adjusted, so that at least three cameras of acquisition component are in focal centre length altogether,
Wherein, this method further include:
Splicing is carried out to multiple images, obtains the true value image in seamless and non-overlapping region.
Fig. 4 a shows the schematic diagram of the acquisition scene of the image collecting device according to one embodiment of the disclosure；
Fig. 4 b shows the schematic diagram of the true value image of the image acquisition device according to one embodiment of the disclosure.
In this implementation, the optical center of each camera is each camera in corresponding plane mirror imaging position
Virtual optical center, the height of at least three cameras is total focal centre length when the virtual optical center of at least three cameras is overlapped.Acquisition
It is seamless non-overlapping region to the spliced panorama sketch of multiple images when at least three cameras of component are in focal centre length altogether
True value image.
In one possible implementation, it to form reflection subassembly by four sides plane mirror, is made of and adopts four cameras
It is illustrated for the image collecting device of collection component.Wherein, the shape of each plane mirror is isosceles three in image collecting device
Angular, camera with plane mirror each take the photograph positioned at corresponding plane mirror lower section middle position one by one by corresponding arrangement, camera
As being equidistant for head and reflection subassembly central axis.When the first angle etc. that the mirror surface and horizontal plane of each plane mirror are kept
When 45 degree, as shown in fig. 4 a, the optical center A3 of each camera is imaged as A4 in corresponding plane mirror A1, adjusts the height of camera
Degree realizes the total optical center of multiple camera A3, at this time camera A3 so that the A4 of at least three cameras is overlapped on a point
Height be altogether focal centre length.When the height of camera A3 is equal to focal centre length altogether, camera A3 is shot straight up, by flat
The reflection of face mirror mirror surface can collect horizontal scene image.The visual field of camera A3 does not have overlapping region, and multiple camera shootings
There is no parallax between head, multiple image can be spliced by direct splicing method by a panorama sketch seamless and without ghost image, it should
Panorama sketch is as shown in Figure 4 b, can be used as true value image.
It should be appreciated that can also according to actual needs using other merging algorithm for images, the disclosure to this with no restriction.
In one possible implementation, the position of at least three cameras of acquisition component is adjusted in step s 11
And shooting direction, comprising:
The position of acquisition component is adjusted, so that at least three cameras of acquisition component are higher than focal centre length altogether,
Wherein, this method further include: using the multiple images with overlapping region as input picture.
Fig. 5 a shows the schematic diagram of the acquisition scene of the image collecting device according to one embodiment of the disclosure；
Figure 5b shows that the schematic diagrames according to the input picture of the image acquisition device of one embodiment of the disclosure.
In this implementation, when the position of at least three cameras of acquisition device is higher than focal centre length altogether, virtual optical
The heart is no longer overlapped, and the visual field of adjacent camera has certain overlapping region, and there is also certain between multiple cameras
Parallax.At this point, the multiple images that can acquire at least three cameras are directly as input picture.
To form reflection subassembly by four sides plane mirror, it is set to by the image collector that four cameras form acquisition component
Example.Wherein, the shape of each plane mirror is isosceles triangle in image collecting device, camera and plane mirror corresponding cloth one by one
It sets, camera is located at middle position below corresponding plane mirror, and each camera phase at a distance from reflection subassembly central axis
Deng.When the first angle that the mirror surface of each plane mirror and horizontal plane are kept is equal to 45 degree, as shown in Figure 5 a, each camera
Optical center A3 be imaged as virtual optical center A4 in corresponding plane mirror A1, the height of camera is adjusted, when the height of camera is higher than
Focal centre length altogether, so that the virtual optical center A4 of at least three cameras is not overlapped, camera is shot straight up, passes through plane mirror
The reflection of A1 mirror surface can collect horizontal scene image.The visual field of camera has certain overlapping region, multiple to take the photograph
As there is also certain parallaxes between head.At this point, four cameras are rendered as the arrangement mode of common full-view camera just, it can
With by multiple image (as shown in Figure 5 b) directly as input picture.
Repeat the above steps S11-S12, by repeatedly adjusting the position of acquisition component and being acquired, obtains seamless and nothing
The true value image (as shown in Figure 4 b) of overlapping region and multiple input pictures (as shown in Figure 5 b) with overlapping region.
In one possible implementation, the method may also include that according to true value image and multiple input pictures,
Training dataset of the building for training image splicing network.In this implementation, according to the true value image of acquisition and more
A input picture can construct the data set for training image splicing network, can also construct for training other deep learnings
The data set of model, the disclosure to this with no restriction.
In one possible implementation, image mosaic network includes convolutional neural networks, generates in confrontation network
It is at least one.
In this implementation, image mosaic network includes convolutional neural networks, generates and fight at least one of network,
It can also be other image mosaic networks, image mosaic network can be selected according to actual needs, the disclosure does not limit this
System.
In one possible implementation, the method may also include that according to training data concentrate true value image with
And multiple input pictures, training image splice network.In this implementation, can to training data concentrate true value image with
And multiple reasonable pretreatments of input picture addition, reasonable network model and loss function are selected to construct image mosaic net
Network, and then the image mosaic network is trained to realize to obtain satisfactory image mosaic network based on deep learning
Merging algorithm for images.
Image collecting device and method provided by the embodiment of the present disclosure, image collecting device include reflection subassembly and acquisition
Component, reflection subassembly include at least three plane mirrors for being bonded into inverted pyramid type, and acquisition component includes and at least three planes
At least three cameras that mirror is correspondingly arranged, for acquiring multiple images of the horizontal scene after reflection subassembly reflects.This
The image collecting device that open embodiment provides, true value image and presence seamless and without ghost image can be obtained by the device
The plurality of input images of parallax provides for the problems such as image mosaic data set shortage, image mosaic true value image acquisition difficulty
Feasible solution, to meet the use demand of image mosaic data.
It should be noted that although describing image collecting device and method, ability using above-described embodiment as example
Field technique personnel it is understood that the disclosure answer it is without being limited thereto.In fact, user can answer completely according to personal preference and/or actually
Each step is flexibly set with scene, as long as meeting the technical solution of the disclosure.
The presently disclosed embodiments is described above, above description is exemplary, and non-exclusive, and
It is not limited to disclosed each embodiment.Without departing from the scope and spirit of illustrated each embodiment, for this skill
Many modifications and changes are obvious for the those of ordinary skill in art field.The selection of term used herein, purport
In principle, the practical application or to the technological improvement in market for best explaining each embodiment, or make the art its
Its those of ordinary skill can understand each embodiment disclosed herein.
Claims (10)
1. a kind of image collecting device, which is characterized in that described device includes:
Reflection subassembly, including at least three plane mirrors, at least three plane mirror are bonded into inverted pyramid type；
Acquisition component, including at least three cameras, at least three camera are corresponding at least three plane mirror
Ground arrangement, the acquisition component are used to acquire multiple images of the horizontal scene after reflection subassembly reflection.
2. the apparatus according to claim 1, which is characterized in that each plane mirror of the reflection subassembly and the folder of horizontal plane
Angle remains first angle, and the shape of each plane mirror is isosceles triangle,
Wherein, the first angle is greater than 0 degree and less than 90 degree.
3. the apparatus according to claim 1, which is characterized in that each camera of the acquisition component is located at corresponding
Middle position below plane mirror, and the height of each camera and the phase at a distance from the central axis of the reflection subassembly
Together.
4. the apparatus according to claim 1, which is characterized in that be in altogether at least three cameras of the acquisition component
It is the true value image in seamless and non-overlapping region to the spliced panorama sketch of described multiple images when focal centre length,
Wherein, at least three cameras described in when the focal centre length altogether is the virtual optical center coincidence of at least three camera
Height, the virtual optical center of each camera is imaging position of the optical center of each camera on corresponding plane mirror.
5. a kind of image-pickup method, which is characterized in that the method is applied to according to claim 1 described in any one of -4
Image collecting device, described image acquisition device includes reflection subassembly and acquisition component, which comprises
Adjust position and the shooting direction of at least three cameras of the acquisition component；
Multiple images of the horizontal scene after reflection subassembly reflection are acquired by the acquisition component.
6. according to the method described in claim 5, it is characterized in that, adjusting the position of at least three cameras of the acquisition component
It sets and shooting direction, comprising:
The position of the acquisition component is adjusted, so that at least three cameras of the acquisition component are in focal centre length altogether,
In, it is described altogether focal centre length be at least three camera virtual optical center be overlapped when described at least three cameras height
Degree, the virtual optical center of each camera are imaging position of the optical center of each camera on corresponding plane mirror,
Wherein, the method also includes:
Splicing is carried out to described multiple images, obtains the true value image in seamless and non-overlapping region.
7. according to the method described in claim 6, wrapping it is characterized in that, adjust position and the shooting direction of the acquisition component
It includes:
The position of the acquisition component is adjusted, so that at least three cameras of the acquisition component are higher than focal centre length altogether,
Wherein, the method also includes: using the multiple images with overlapping region as input picture.
8. the method according to the description of claim 7 is characterized in that the method also includes:
According to the true value image and multiple input pictures, training dataset of the building for training image splicing network.
9. according to the method described in claim 8, it is characterized in that, the method also includes:
According to true value image and multiple input pictures that the training data is concentrated, training described image splices network.
10. method according to claim 7 or 8, which is characterized in that it includes convolutional Neural net that described image, which splices network,
Network generates and fights at least one of network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811176600.8A CN109257527A (en) | 2018-10-10 | 2018-10-10 | The acquisition device and method of image mosaic data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811176600.8A CN109257527A (en) | 2018-10-10 | 2018-10-10 | The acquisition device and method of image mosaic data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109257527A true CN109257527A (en) | 2019-01-22 |
Family
ID=65045821
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811176600.8A Pending CN109257527A (en) | 2018-10-10 | 2018-10-10 | The acquisition device and method of image mosaic data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109257527A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111193920A (en) * | 2019-12-31 | 2020-05-22 | 重庆特斯联智慧科技股份有限公司 | Video picture three-dimensional splicing method and system based on deep learning network |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CA2174157A1 (en) * | 1995-04-28 | 1996-10-29 | Vishvjit Singh Nalwa | Panoramic viewing apparatus |
| US6128143A (en) * | 1998-08-28 | 2000-10-03 | Lucent Technologies Inc. | Panoramic viewing system with support stand |
| CN101521745A (en) * | 2009-04-14 | 2009-09-02 | 王广生 | Multi-lens optical center superposing type omnibearing shooting device and panoramic shooting and retransmitting method |
| CN101783883A (en) * | 2009-12-26 | 2010-07-21 | 华为终端有限公司 | Adjusting method in co-optical-center videography and co-optical-center camera system |
-
2018
- 2018-10-10 CN CN201811176600.8A patent/CN109257527A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CA2174157A1 (en) * | 1995-04-28 | 1996-10-29 | Vishvjit Singh Nalwa | Panoramic viewing apparatus |
| US6128143A (en) * | 1998-08-28 | 2000-10-03 | Lucent Technologies Inc. | Panoramic viewing system with support stand |
| CN101521745A (en) * | 2009-04-14 | 2009-09-02 | 王广生 | Multi-lens optical center superposing type omnibearing shooting device and panoramic shooting and retransmitting method |
| CN101783883A (en) * | 2009-12-26 | 2010-07-21 | 华为终端有限公司 | Adjusting method in co-optical-center videography and co-optical-center camera system |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111193920A (en) * | 2019-12-31 | 2020-05-22 | 重庆特斯联智慧科技股份有限公司 | Video picture three-dimensional splicing method and system based on deep learning network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Adelson et al. | Single lens stereo with a plenoptic camera | |
| EP3101624B1 (en) | Image processing method and image processing device | |
| US10015469B2 (en) | Image blur based on 3D depth information | |
| CN103945210B (en) | A kind of multi-cam image pickup method realizing shallow Deep Canvas | |
| EP0680638B1 (en) | Process for estimating disparities between monoscopic images forming a stereoscopic image | |
| CN111355878A (en) | Terminal, shooting method and storage medium | |
| CN107454377B (en) | A kind of algorithm and system carrying out three-dimensional imaging using camera | |
| US20220272261A1 (en) | Imaging system, method, and applications | |
| US20010020976A1 (en) | Stereo panoramic camera arrangements for recording panoramic images useful in a stereo panoramic image pair | |
| CA3019936A1 (en) | Three-dimensional, 360-degree virtual reality camera exposure control | |
| CN108093237A (en) | High spatial resolution optical field acquisition device and image generating method | |
| JP6971084B2 (en) | Methods and devices for generating data that expresses the blur associated with light field data | |
| CN109147027A (en) | Monocular image three-dimensional rebuilding method, system and device based on reference planes | |
| Wu et al. | Geometry based three-dimensional image processing method for electronic cluster eye | |
| Chen et al. | Bionic mosaic method of panoramic image based on compound eye of fly | |
| CN104915919B (en) | Image processing apparatus and image processing method | |
| CN109257527A (en) | The acquisition device and method of image mosaic data | |
| CN108051979A (en) | Panoramic imaging system and method thereof | |
| Kang et al. | Fast dense 3D reconstruction using an adaptive multiscale discrete-continuous variational method | |
| KR20200117507A (en) | Camera module and image operating method performed therein | |
| US11170514B2 (en) | Image processing apparatus, image processing method, 3D printing system, and storage medium | |
| KR101469361B1 (en) | Apparatus for panorama image acquisition | |
| CN109426823B (en) | AR device photographing method and device and AR device | |
| CN106846469B (en) | Based on tracing characteristic points by the method and apparatus of focusing storehouse reconstruct three-dimensional scenic | |
| CN109618095A (en) | Double 8K image virtual reality processing methods and system |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190122 |