CN108734679A - A kind of computer vision system - Google Patents
A kind of computer vision system Download PDFInfo
- Publication number
- CN108734679A CN108734679A CN201810504070.9A CN201810504070A CN108734679A CN 108734679 A CN108734679 A CN 108734679A CN 201810504070 A CN201810504070 A CN 201810504070A CN 108734679 A CN108734679 A CN 108734679A
- Authority
- CN
- China
- Prior art keywords
- module
- image
- layer
- unit
- tiling
- 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
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 14
- 238000005070 sampling Methods 0.000 claims abstract description 13
- 238000000034 method Methods 0.000 claims abstract description 12
- HTFVKMHFUBCIMH-UHFFFAOYSA-N 1,3,5-triiodo-1,3,5-triazinane-2,4,6-trione Chemical compound IN1C(=O)N(I)C(=O)N(I)C1=O HTFVKMHFUBCIMH-UHFFFAOYSA-N 0.000 claims abstract description 9
- 238000004891 communication Methods 0.000 claims abstract description 7
- 230000003321 amplification Effects 0.000 claims abstract description 4
- 238000003199 nucleic acid amplification method Methods 0.000 claims abstract description 4
- 238000011084 recovery Methods 0.000 claims abstract description 3
- 230000005284 excitation Effects 0.000 claims description 7
- 239000011159 matrix material Substances 0.000 claims description 4
- 238000012986 modification Methods 0.000 claims description 4
- 230000004048 modification Effects 0.000 claims description 4
- 210000002569 neuron Anatomy 0.000 claims description 4
- 210000004027 cell Anatomy 0.000 claims description 2
- 238000013528 artificial neural network Methods 0.000 claims 1
- 238000005516 engineering process Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 239000000686 essence Substances 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30212—Military
Abstract
The invention discloses a kind of computer vision systems, including：Display module, Image Intensified System, the display module includes a frame of reference and an identifying system, the display module connect coding module, the coding module connection tiling convolutional neural networks, the tiling convolutional neural networks use TICA methods, the tiling convolutional neural networks connection described image increases system, it includes a characteristic extracting module and an image amplification module that described image, which enhances system, described image enhances module and connects sampling module, the sampling module includes an area image recovery module and a communication module, the sampling module connects output module.Advantages of the present invention：Computer is solved to image interpretation, interpretation accuracy is improved so that it can carry out large-scale image procossing, improves the efficiency of image procossing.
Description
Technical field
The invention belongs to field of computer technology, and in particular to a kind of computer vision system.
Background technology
With the development of science and technology, the means such as missile-borne scouting, unmanned plane scouting are constantly progressive perfect, these reconnaissance means acquisition
The recognition methods of great amount of images information, traditional images information is artificial interpretation.Artificial interpretation one is that speed is slow, by it is artificial because
Element is affected, it is difficult to meet battlefield operation needs, the especially needs of precision strike operation, second is that with the hair of reconnaissance means
Exhibition, the image information of acquisition is more and more, is influenced by interpretation professional's quantity, it is difficult to complete interpretation task.Therefore, it uses
Computer carries out interpretation to image, at being an urgent problem to be solved currently.
Invention content
In order to solve above-mentioned the deficiencies in the prior art, the present invention provides a kind of computer vision systems, solve computer
To image interpretation, interpretation accuracy is improved.
In order to achieve the above object, the technical solution adopted by the present invention is：
A kind of computer vision system, including：Display module, Image Intensified System, the display module include a ginseng
Test system and an identifying system, the display module connect coding module, the coding module connection tiling convolutional Neural net
Network, the tiling convolutional neural networks use TICA methods, the tiling convolutional neural networks connection described image to increase system,
It includes a characteristic extracting module and an image amplification module that described image, which enhances system, and described image enhancing module connection is adopted
Egf block, the sampling module include an area image recovery module and a communication module, and the sampling module connection is defeated
Go out module.
To further description of the present invention, the frame of reference is in communication in the identifying system, and the identifying system passes through
Thick identification carries out simple classification, is then communicated to tiling convolutional neural networks.
To further description of the present invention, the TICA methods two-tier network, weight W is obtained by study in first layer,
Weight V is fixed in the second layer, only by hard coded indicate preceding layer in neuron spatial relationship, select quadratic sum square root for
The excitation value of simple unit and pond unit, the fraction list that each second layer pond unit pi his simple to first layer closes on
First pond gives an input pattern x(t), the excitation value of each the second layer unit is：
To further description of the present invention, the parameter W is indicated by finding sparse features in the second layer, solves formula
In, input pattern isHere W ∈ Rm×nWith V ∈ Rm×n, wherein n is the size of input picture or characteristic pattern, and m is one layer
The number of middle Hidden unit, V are a fixed matrix (Vij=1or 0) it is used for indicating Hidden unit hiTwo-dimensional spatial relationship,
Specifically, hiUnit is located at a two-dimensional grid, wherein each piIt is connected to a continuous hiCell block, orthogonality constraint
WWT=I ensures that the feature learnt is varied.
To further description of the present invention, the output module includes that an optical projection device is filled with a projection modification
It sets.
The present invention's has the advantages that：Computer is solved to image interpretation, improve interpretation accuracy so that it can
Large-scale image procossing is carried out, the efficiency of image procossing is improved.
Description of the drawings
Fig. 1 is the module map of the present invention
Specific implementation mode
Below in conjunction with attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that is retouched
The embodiment stated is only a part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, originally
The every other embodiment that field those of ordinary skill is obtained without creative efforts, belongs to the present invention
The range of protection.
Embodiment 1, as shown in Figure 1：
A kind of computer vision system, including：Display module, Image Intensified System, display module include a referential
System and an identifying system, display module connect coding module, coding module connection tiling convolutional neural networks, tiling convolution god
TICA methods, tiling convolutional neural networks connection image is used to increase system through network, Image Intensified System is carried including a feature
Modulus block and an image amplification module, image enhancement module connect sampling module, and sampling module includes that an area image is extensive
Multiple module and a communication module, sampling module connect output module.
A kind of optimal technical scheme of the present embodiment, frame of reference are in communication in identifying system, and identifying system is by slightly identifying
Simple classification is carried out, tiling convolutional neural networks are then communicated to.
A kind of optimal technical scheme of the present embodiment, TICA method two-tier networks, weight W is by learning in first layer
It arrives, weight V is fixed in the second layer, and the spatial relationship of neuron in preceding layer is only indicated by hard coded, selects quadratic sum square
Root is the excitation value of simple unit and pond unit, the small portion that each second layer pond unit pi his simple to first layer closes on
Subdivision pond gives an input pattern x(t), the excitation value of each the second layer unit is：
A kind of optimal technical scheme of the present embodiment, parameter W are indicated by finding sparse features in the second layer, solve formula
In, input pattern isHere W ∈ Rm×nWith V ∈ Rm×n, wherein n is the size of input picture or characteristic pattern, and m is one layer
The number of middle Hidden unit, V are a fixed matrix (Vij=1or 0) it is used for indicating Hidden unit hiTwo-dimensional spatial relationship,
Specifically, hiUnit is located at a two-dimensional grid, wherein each piIt is connected to a continuous hiCell block, orthogonality constraint
WWT=I ensures that the feature learnt is varied.
A kind of optimal technical scheme of the present embodiment, output module include that an optical projection device is filled with a projection modification
It sets.
In the present embodiment, operation is carried out using TICA methods by the convolutional neural networks that tile, then brings solution formula into,
The data transmission to generate to Image Intensified System is enhanced, sampling processing is then carried out by sampling module, finally by defeated
Go out module and image output is carried out to image, thick identification can be carried out using identifying system in image processing process and simply divided
Then class is communicated to tiling convolutional neural networks, increase the precision of image recognition, according to TICA method two-tier networks, first
Weight W is obtained by study in layer, and weight V is fixed in the second layer, and the space of neuron in preceding layer is only indicated by hard coded
Relationship selects quadratic sum square root for the excitation value of simple unit and pond unit, each pi pairs of second layer pond unit
The fraction unit cells that one layer of simple hi closes on give an input pattern x(t), detailed calculation process is carried out to image.Together
Shi Liyong parameters W indicates that in solution formula, input pattern is by finding sparse features in the second layerHere W ∈ Rm ×nWith V ∈ Rm×n, wherein n is the size of input picture or characteristic pattern, and m is the number of Hidden unit in one layer, and V is a fixation
Matrix (Vij=1or 0) it is used for indicating Hidden unit hiTwo-dimensional spatial relationship, specifically, hiUnit is located at a two dimension
Grid, wherein each piIt is connected to a continuous hiCell block, orthogonality constraint WWT=I ensures that the feature learnt is a variety of more
Sample, it is ensured that the accuracy of image procossing and treatment effeciency.
It is obvious to a person skilled in the art that invention is not limited to the details of the above exemplary embodiments, Er Qie
In the case of without departing substantially from spirit or essential attributes of the invention, the present invention can be realized in other specific forms.Therefore, no matter
From the point of view of which point, the present embodiments are to be considered as illustrative and not restrictive, and the scope of the present invention is by appended power
Profit requires rather than above description limits, it is intended that all by what is fallen within the meaning and scope of the equivalent requirements of the claims
Variation is included within the present invention.Any reference signs in the claims should not be construed as limiting the involved claims.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all essences in the present invention
With within principle, any modification, equivalent replacement, improvement and so on should all be included in the protection scope of the present invention god.
Claims (5)
1. a kind of computer vision system, including：Display module, Image Intensified System, which is characterized in that the display module packet
Include a frame of reference and an identifying system, the display module connect coding module, the coding module connection tiling volume
Product neural network, the tiling convolutional neural networks use TICA methods, the tiling convolutional neural networks to connect described image
Increasing system, it includes a characteristic extracting module and an image amplification module that described image, which enhances system, and described image enhances mould
Block connects sampling module, and the sampling module includes an area image recovery module and a communication module, the sampling mould
Block connects output module.
2. a kind of computer vision system according to claim 1, which is characterized in that the frame of reference is in communication in described
Then identifying system, the identifying system are communicated to tiling convolutional neural networks by slightly identifying carry out simple classification.
3. a kind of computer vision system according to claim 1, which is characterized in that the TICA methods two-tier network,
Weight W is obtained by study in first layer, and weight V is fixed in the second layer, only indicates neuron in preceding layer by hard coded
Spatial relationship selects quadratic sum square root for the excitation value of simple unit and pond unit, each second layer pond unit pi
The fraction unit cells that hi simple to first layer closes on give an input pattern x(t), the excitation of each the second layer unit
Value is：
4. a kind of computer vision system according to claim 1, which is characterized in that the parameter W passes through in the second layer
Middle searching sparse features indicate that in solution formula, input pattern isHere W ∈ Rm×nWith V ∈ Rm×n, wherein n is input
The size of image or characteristic pattern, m are the numbers of Hidden unit in one layer, and V is a fixed matrix (Vij=1 or 0) it is used for table
Show Hidden unit hiTwo-dimensional spatial relationship, specifically, hiUnit is located at a two-dimensional grid, wherein each piIt is connected to
One continuous hiCell block, orthogonality constraint WWT=I ensures that the feature learnt is varied.
5. a kind of computer vision system according to claim 1, which is characterized in that the output module includes a light
Projection arrangement and a projection modification device.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810504070.9A CN108734679A (en) | 2018-05-23 | 2018-05-23 | A kind of computer vision system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810504070.9A CN108734679A (en) | 2018-05-23 | 2018-05-23 | A kind of computer vision system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108734679A true CN108734679A (en) | 2018-11-02 |
Family
ID=63936203
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810504070.9A Pending CN108734679A (en) | 2018-05-23 | 2018-05-23 | A kind of computer vision system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108734679A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102881003A (en) * | 2012-08-30 | 2013-01-16 | 暨南大学 | Method for removing cosmic rays in charge-coupled device (CCD) astronomic image |
| CN104598916A (en) * | 2014-09-11 | 2015-05-06 | 单勇 | Establishment method of train recognition system and train recognition method |
| CN106897673A (en) * | 2017-01-20 | 2017-06-27 | 南京邮电大学 | A kind of recognition methods again of the pedestrian based on retinex algorithms and convolutional neural networks |
| CN106991440A (en) * | 2017-03-29 | 2017-07-28 | 湖北工业大学 | A kind of image classification algorithms of the convolutional neural networks based on spatial pyramid |
| CN107481233A (en) * | 2017-08-22 | 2017-12-15 | 广州辰创科技发展有限公司 | A kind of image-recognizing method being applied in FOD foreign bodies detection radars |
-
2018
- 2018-05-23 CN CN201810504070.9A patent/CN108734679A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102881003A (en) * | 2012-08-30 | 2013-01-16 | 暨南大学 | Method for removing cosmic rays in charge-coupled device (CCD) astronomic image |
| CN104598916A (en) * | 2014-09-11 | 2015-05-06 | 单勇 | Establishment method of train recognition system and train recognition method |
| CN106897673A (en) * | 2017-01-20 | 2017-06-27 | 南京邮电大学 | A kind of recognition methods again of the pedestrian based on retinex algorithms and convolutional neural networks |
| CN106991440A (en) * | 2017-03-29 | 2017-07-28 | 湖北工业大学 | A kind of image classification algorithms of the convolutional neural networks based on spatial pyramid |
| CN107481233A (en) * | 2017-08-22 | 2017-12-15 | 广州辰创科技发展有限公司 | A kind of image-recognizing method being applied in FOD foreign bodies detection radars |
Non-Patent Citations (1)
| Title |
|---|
| QUOC V.LE等: "Tiled convolutional neural networks", 《ANNUAL CONFERENCE ON NEURAL INFORMATION PROCESSING SYSTEMS 2010》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113807210B (en) | Remote sensing image semantic segmentation method based on pyramid segmentation attention module | |
| CN110210486B (en) | Sketch annotation information-based generation countermeasure transfer learning method | |
| Hou et al. | C-UNet: Complement UNet for remote sensing road extraction | |
| CN109829399A (en) | A kind of vehicle mounted road scene point cloud automatic classification method based on deep learning | |
| Zhou et al. | Effects of spatial pattern of forest vegetation on urban cooling in a compact megacity | |
| Petrovska et al. | Aerial scene classification through fine-tuning with adaptive learning rates and label smoothing | |
| CN113239954B (en) | Attention mechanism-based image semantic segmentation feature fusion method | |
| CN111401156B (en) | Image identification method based on Gabor convolution neural network | |
| Hung et al. | Remote sensing scene classification and explanation using RSSCNet and LIME | |
| CN112102165B (en) | Light field image angular domain super-resolution system and method based on zero sample learning | |
| CN110704649B (en) | Method and system for constructing flow image data set | |
| CN115457531A (en) | Method and device for recognizing text | |
| CN109523558A (en) | A kind of portrait dividing method and system | |
| CN113256649A (en) | Remote sensing image station selection and line selection semantic segmentation method based on deep learning | |
| Chen et al. | A fusion-based defogging algorithm | |
| CN114399018B (en) | Efficient ientNet ceramic fragment classification method based on sparrow optimization of rotary control strategy | |
| CN116108917A (en) | Non-marked sample enhanced semi-supervised graph neural network method | |
| Wang et al. | Deep-learning-based automatic detection of photovoltaic cell defects in electroluminescence images | |
| Piątkowski et al. | TinyML-based concept system used to analyze whether the face mask is worn properly in battery-operated conditions | |
| Yang et al. | Deep learning-based method for detection of external air conditioner units from street view images | |
| Liu et al. | YOLOv3_ReSAM: A small-target detection method | |
| CN108734679A (en) | A kind of computer vision system | |
| Wu et al. | Accelerating haze removal algorithm using cuda | |
| CN115082800B (en) | Image segmentation method | |
| Min et al. | The Floor Plan Design Method of Exhibition Halls in CGAN-Assisted Museum Architecture |
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 |
Application publication date: 20181102 |
|
| RJ01 | Rejection of invention patent application after publication |