CN107563973A - A kind of image processing method and device - Google Patents
A kind of image processing method and device Download PDFInfo
- Publication number
- CN107563973A CN107563973A CN201710697106.5A CN201710697106A CN107563973A CN 107563973 A CN107563973 A CN 107563973A CN 201710697106 A CN201710697106 A CN 201710697106A CN 107563973 A CN107563973 A CN 107563973A
- Authority
- CN
- China
- Prior art keywords
- image
- edge
- processing
- coloured image
- carried out
- 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
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The present invention proposes a kind of image processing method and device.A kind of image processing method, including：Obtain pending coloured image；Edge denoising is carried out to the coloured image；Gray value conversion processing is carried out to the coloured image after edge denoising, obtains gray level image.Gradation of image conversion processing is realized using the above method, so as to get gray level image more readily identify, beneficial to improve image recognition precision.
Description
Technical field
The present invention relates to digital image processing techniques field, more particularly to a kind of image processing method and device.
Background technology
Image recognition, refer to handle image using computer, analyzed and understood, to identify various different modes
The technology of target and object.Most common image recognition is to identify the word content in image.Because the data volume of image is relative
It is larger, so when carrying out Text region to image, typically coloured image first can be converted into gray level image, this process is claimed
Converted for gray-scale map.
Traditional image recognition technology to character image carry out gray-scale map conversion when, only realize from coloured image to
The conversion of gray level image, the denoising to image border is not related to, when causing subsequently picture material is identified, know
Other precision is not high.
The content of the invention
The defects of based on above-mentioned prior art and deficiency, the present invention propose a kind of image processing method and device, Neng Gou
When carrying out gray scale conversion to image, edge denoising is carried out to image, beneficial to raising image recognition precision.
First aspect present invention proposes a kind of image processing method, including：Obtain pending coloured image；To the coloured silk
Color image carries out edge denoising；Gray value conversion processing is carried out to the coloured image after edge denoising, obtained
Gray level image.Above-mentioned processing procedure, after coloured image is obtained, edge denoising first is carried out to image, then carries out ash again
Angle value conversion processing, the edge noise that the gray-scale map so obtained contains is less, and border is apparent, beneficial to identification.
Second aspect of the present invention proposes a kind of image processing apparatus, including：Image acquisition unit, it is pending for obtaining
Coloured image；Denoising unit, for carrying out edge denoising to the coloured image；Gray scale conversion unit, for pair
The coloured image after edge denoising carries out gray value conversion processing, obtains gray level image.
In one implementation, after pending image is obtained, this method also includes：Identify the coloured image
Image type；If the coloured image is Chinese text image, edge denoising is carried out to the coloured image；If
The coloured image is English text image, then directly carries out gray value conversion processing to the coloured image, obtain gray-scale map
Picture.In this implementation, Chinese text image and English text image are handled differently, side is performed to Chinese text image
Edge denoising, and gray scale conversion processing is directly carried out to English text image, so it is handled differently, can be according to actual conditions
Improve treatment effeciency.
In one implementation, edge denoising is carried out to the coloured image, including：The coloured image is entered
The processing of row edge swell and/or edge corrosion processing.In this implementation, at using edge swell processing or edge corrosion
Reason realizes the edge denoising to coloured image.
In one implementation, edge swell processing is carried out to the coloured image and edge corrosion is handled, including：Press
According to default kernel function dot matrix combined value, edge swell processing is carried out to the coloured image；According to default kernel function, opposite side
The coloured image after edge expansion process carries out edge corrosion processing.In a kind of this implementation, first coloured image is entered
The processing of row edge swell, then edge corrosion processing is carried out, realize the edge denoising to image.
In one implementation, after gray level image is obtained, this method also includes：To each picture of the gray level image
The gray value of vegetarian refreshments is normalized；The gray value of each pixel after normalized is carried out at opening operation
Reason；The gray value of each pixel after split calculation process carries out renormalization processing.In this implementation, exist
After completing gray-scale map conversion, opening operation processing is further carried out to obtained gray-scale map, reduces the sawtooth effect of image border
Should.
Brief description of the drawings
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
There is the required accompanying drawing used in technology description to be briefly described, it should be apparent that, drawings in the following description are only this
The embodiment of invention, for those of ordinary skill in the art, on the premise of not paying creative work, can also basis
The accompanying drawing of offer obtains other accompanying drawings.
Fig. 1 is a kind of schematic flow sheet of image processing method provided in an embodiment of the present invention；
Fig. 2 is the schematic flow sheet of another image processing method provided in an embodiment of the present invention；
Fig. 3 is a kind of structural representation of image processing apparatus provided in an embodiment of the present invention；
Fig. 4 is the structural representation of another image processing apparatus provided in an embodiment of the present invention；
Fig. 5 is the structural representation of another image processing apparatus provided in an embodiment of the present invention.
Embodiment
Technical scheme of the embodiment of the present invention is applied to the scene that coloured image is converted into gray level image.Using of the invention real
A technical scheme is applied, denoising can be carried out to image border, known beneficial to image is improved while gray-scale map conversion is realized
Other precision.
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Site preparation describes, it is clear that described embodiment is only part of the embodiment of the present invention, rather than whole embodiments.It is based on
Embodiment in the present invention, those of ordinary skill in the art are obtained every other under the premise of creative work is not made
Embodiment, belong to the scope of protection of the invention.
Shown in Figure 1 the embodiment of the invention discloses a kind of image processing method, this method includes：
S101, obtain pending coloured image；
Specifically, above-mentioned coloured image, refers to need the image for including word for being converted into gray level image.It is above-mentioned pending
Coloured image, can be obtained by any approach, the colored text image of digital form.Such as from the Internet download
Colored text image, utilize colored text image of digital equipment shooting etc..
S102, edge denoising is carried out to the coloured image；
Specifically, the embodiment of the present invention includes the edge denoising to coloured image.Edge denoising can make image word
Boundary between symbol is apparent, and more conducively image character is identified.The embodiment of the present invention can be can be used for using any one
Edge denoising is so that the apparent edge denoising method of intercharacter boundary.Edge denoising is carried out to the coloured image of acquisition.
S103, gray value conversion processing is carried out to the coloured image after edge denoising, obtain gray level image.
Specifically, the embodiment of the present invention is using most common collapse function Y=0.3R+0.59G+0.11B realizations pair at present
The gray scale conversion of coloured image, wherein R, G, B are respectively the three primary colors of coloured image, and Y is characterized as the gray value of image.Pass through
The 3-dimensional value that above-mentioned linear collapse function realizes coloured picture is converted into 1 dimension gray value.Above-mentioned collapse function is directly encapsulated by python
Realize, so special modification need not be made, and aforementioned proportion value is through prolonged artificial study gained, should not be changed.
Using above-mentioned technical proposal, before gray scale conversion is carried out to coloured image, edge first is carried out to coloured image and gone
Make an uproar processing, make the boundary in image between character apparent, gray scale conversion then is carried out to coloured image again, so convert and obtain
Gray level image be more conducive to identify, gray level image accuracy of identification can be improved.
Fig. 2 shows the more specifically processing procedure of the image processing method shown in Fig. 1.Shown in Figure 2, the present invention is real
Image processing method disclosed in example is applied, is specifically included：
S201, obtain pending coloured image；
Specifically, above-mentioned coloured image, refers to need the image for including word for being converted into gray level image.It is above-mentioned pending
Coloured image, can be obtained by any approach, the colored text image of digital form.Such as from the Internet download
Colored text image, utilize colored text image of digital equipment shooting etc..
The image type for the pending coloured image that S202, identification obtain；
Specifically, above-mentioned image type, including Chinese text image and English text image.Identify pending coloured image
Image type, that is, identify that pending coloured image is Chinese text image, or English text image.
If the Chinese text in the pending coloured image obtained occupies larger proportion, then it is assumed that during the coloured image is
Literary text image；On the contrary, if the English text of the pending coloured image obtained occupies larger proportion, then it is assumed that the cromogram
As being English text image.Manual identified, or any available Computer Identification can be specifically used to realize to image class
The identification of type.
Because Chinese text image is relative to English text image, content of text is more complicated, and identification difficulty is higher, because
This, the embodiment of the present invention carries out image type differentiation to the pending image of acquisition, and Chinese text image is introduced below
Comprehensive processing, and for English text image, gray-scale map conversion is carried out using conventional gray scale conversion method.This hair
Bright embodiment introduces above-mentioned comprehensive processing procedure exemplified by carrying out gray-scale map conversion to Chinese text image.It is appreciated that
It is the comprehensive processing method that the embodiment of the present invention proposes, the processing to English text image can also be applied to.
If pending coloured image is Chinese text coloured image, step S203 is performed, according to default core letter
Number dot matrix combined value, edge swell processing is carried out to pending coloured image；
Specifically, above-mentioned edge swell processing, refers to expand character edge, improves the processing of character proportion.Edge
Expansion process can be remarkably reinforced the identified probability of character picture.
The cardinal principle of edge swell processing is the region (character boundary region) larger to the gray scale difference of image, by right
Around boundary point in the region, the pixel in the range of kernel function is sampled and generally increases the side of higher limit
Method, realize the expansion of character boundary.Its parameter mainly configured is kernel function dot matrix combined value.The value is characterized as around pixel
Operating range, its minimum value for (1,1) single-point combine.With the increase of kernel function dot matrix combined value, the section of word is imitated
Fruit is more and more obvious, is more beneficial to identified；But if kernel function dot matrix combined value crosses senior general and causes the word in text expanded
In serious, recognition failures are directly resulted in.In actual production environment, core letter can be flexibly set according to demand and expansion effect
Number dot matrix combined value.
The embodiment of the present invention has tested various kernel function dot matrix combined value, finally finds to use (6,6) kernel function dot matrix group
The best results of conjunction value.Therefore, (6,6) are used as default kernel function dot matrix combined value by the embodiment of the present invention, to pending
Coloured image carries out edge swell processing.
S204, according to default kernel function, edge corrosion processing is carried out to the coloured image after edge expansion process；
Specifically, above-mentioned edge corrosion processing, refers to reduce character edge, strengthens the differentiation degree of intercharacter, improve word
Accord with identification probability.
After performing step S203 and carrying out edge swell processing to coloured image, edge corrosion is continued to coloured image
Processing, realizes the relative reduction to coloured image, reduces in edge swell processing to the zoom degree of character, can strengthen
The differentiation degree of intercharacter.
The implementation process that the implementation process of edge corrosion processing is handled with edge swell is equally to kernel function relatively
In the range of pixel handled, only will be enlarged by computing and be changed to collapse computing.This method can realize character edge
Reduction, the free time between character stroke is improved, strengthen the identification degree of character and intercharacter.
Around the character edge of coloured image after selection edge swell processing of the embodiment of the present invention, in the range of kernel function
Pixel carry out collapse processing.In edge corrosion processing, most important configuration parameter is kernel function, most basic kernel function
For (1,1).When actually implementing technical scheme of the embodiment of the present invention, according to demand and corrosive effect, core letter can be flexibly set
Number.With the increase of kernel function, the single effect of character is that the expansion of character pitch is obvious, the identification more beneficial to single character；
But if kernel function crosses senior general and causes that character corrosion is excessively serious, and it is very few to ultimately result in the pixel quantity of character, on the contrary not
Beneficial to character recognition.
The embodiment of the present invention has tested a variety of different kernel functions, finally found that use (5,5) kernel function, and image is carried out
The effect of edge corrosion processing is best, and therefore, kernel function (5,5) is set as default kernel function, opposite side by the embodiment of the present invention
Coloured image after edge expansion process carries out edge corrosion processing.
The edge swell processing and edge corrosion processing that above-mentioned steps S203 and step S204 is introduced, are realized to cromogram
The edge denoising of picture, it can improve to the character recognition probability in coloured image.
It should be noted that above-mentioned edge swell processing and edge corrosion processing are a pair of relative calculation process processes,
When actually implementing technical scheme of the embodiment of the present invention, edge swell processing can be individually performed, or individually perform edge corrosion
Processing, the resolution of character in coloured image can be improved.It is of course also possible to according to the introduction of the embodiment of the present invention, successively
Edge swell processing and edge corrosion processing are performed, or first carries out edge corrosion processing, it is rear to perform edge swell processing.
When edge swell processing and edge corrosion processing united application, law of communication is unsatisfactory for, that is, first carries out edge swell
Edge corrosion processing is performed after processing, and edge swell processing is performed after first carrying out edge corrosion processing, its effect is different.In reality
Border is in use, can according to demand, flexibly from edge swell processing and edge corrosion processing.The embodiment of the present invention is in many experiments
Middle discovery, edge swell processing is first carried out to image, rear progress edge corrosion processing, the recognizable degree highest of its character, because
This, the embodiment of the present invention is preferentially using first progress edge swell processing, the rear method for carrying out edge corrosion processing, to pending
Coloured image carries out edge denoising.
Although edge swell processing mentioned above and edge corrosion processing do not have exchangeability, appoint and so meet to combine
Rule.Carry out that during image conversion different combinations can be used under conditions of order of operation is not changed, so bring
Effect is identical, and king-sized change need not occur for two kinds of respective kernel functions of processing.Although two kinds of processing are all simultaneously
Need to use kernel function, and it is very close in the definition of respective computing Kernel Function, but because the edge of image is different, respectively
It is not necessarily the same from used optimal kernel function.
It is further to note that the side that above-mentioned steps S203 and step S204 is carried out primarily directed to Chinese text image
Edge denoising.Because Chinese text is relative to English text, its character stroke is more complicated, and identification difficulty is higher, therefore, this hair
Bright embodiment improves the difference degree at Chinese text image character edge by edge swell and the processing method of edge corrosion, with
Make the character recognition precision for the gray level image that gray scale conversion obtains higher.When above-mentioned pending coloured image is English text figure
During picture, because English text character is relatively easy, existing character identifying method can realize effective identification to character, therefore
It is contemplated that step S203 and step S204 are skipped, or selection performs step S203 and step S204 first, being walked to simplify processing
Suddenly.It is of course also possible to regard accuracy of identification demand, above-mentioned steps S203 and step S204 processing is carried out to English text image.
S205, gray value conversion processing is carried out to the coloured image after edge corrosion treatment, obtain gray level image；
Specifically, when actually implementing technical scheme of the embodiment of the present invention, if being selectively skipped over step S203 and step
Rapid S204, or selection performs step S203 and step S204 first, then step S205 is specially to the cromogram after respective handling
Picture, gray value conversion processing is carried out, above-mentioned coloured image is converted into gray level image.
Gray value conversion processing is carried out to coloured image, refers to be directed to 3 color information of each pixel in coloured image, leads to
Collapse computing is crossed, 3-dimensional (three primary colors Vector Groups) pixel value is reduced to 1 dimension (gray value) pixel value, is realized from coloured image to ash
Spend the conversion of image.
Wherein particularly noteworthy is that collapse function is the core for realizing the conversion of gradation of image figure.It is most common at present to collapse
Contracting function is Y=0.3R+0.59G+0.11B, and wherein R, G, B are respectively the three primary colors of coloured image, and Y is characterized as the ash of image
Angle value.The 3-dimensional value for realizing coloured picture by above-mentioned linear collapse function is converted into 1 dimension gray value.Above-mentioned collapse function by
Python directly encapsulates realization, so special modification need not be made, and aforementioned proportion value is through prolonged artificial study institute
, it should not change.
Collapse processing is carried out to coloured image to be realized in the case where information is not lost, and reduces the quality of image, and greatly
Width reduces the data volume size of image, significantly reduces the data volume and operand of image recognition processes, while reduces the disturbance of 3 colors and make an uproar
Sound, realize color lump denoising.The mutation color lump that the process of denoising is mainly shown as in image can quilt when carrying out dimension values conversion
The computing of other dimensions of surrounding color lump is cumulative to be flooded, although being mutated the gray value Y calculated and the value of surrounding of color lump
In the presence of certain difference, but due to the cumulative change that certain dimension can be obviously reduced of various dimensions.
S206, the gray value to each pixel of obtained gray level image are normalized；
Specifically, the gray value of each pixel is normalized, refer to use linear function, by each pixel
Grayvalue transition is between [0,1].The embodiment of the present invention uses general linear normalization method, realizes the gray scale of each pixel
The conversion being worth between [0,1].
S207, the gray value to each pixel after normalization carry out opening operation processing；
Specifically, above-mentioned opening operation processing, refers to carry out extracting operation processing to the gray value of pixel.
Image edge expansion process and edge corrosion processing are completed, and after grayvalue transition, image is converted by coloured picture
For gray-scale map, but now noise be present in image, simultaneously because the collapse computing in conversion process easily causes edge sharp
Existing sawtooth effect is dissolved, is unfavorable for the artificial visual of gray-scale map, it is therefore desirable to opening operation processing is carried out to each pixel, reduced
The sawtooth effect of image edge.
After opening operation processing being carried out to the gray value of pixel, between the gray value of pixel still falls within [0,1], but its
Gray value can increase, and expand the gray value average value of former smaller average, ensure that the gray value branch of image is balanced, realize image
The soft benefit in edge.
S208, the gray value to each pixel after opening operation carry out renormalization processing.
Specifically, renormalization processing makes the grey scale pixel value of image return to original codomain scope, the ash finally given
Image is spent, more conforms to gray level image effect corresponding to coloured image.And due to normalization and renormalization during to image
Opening operation has been carried out, has reached the purpose for reducing edge sawtooth effect.
It should be noted that above-mentioned steps S206~S208 is to realize the effect promoting processing after gray-scale map conversion, its mesh
Be that the soft processing of sawtooth is carried out to obtained gray level image, make image border softer.Actually implementing the embodiment of the present invention
During technical scheme, it can flexibly decide whether to carry out obtained gray level image above-mentioned steps S206~S208 processing.
Fig. 3 shows a kind of structural representation of image processing apparatus disclosed in the embodiment of the present invention, shown in Figure 3,
The device includes：Image acquisition unit 301, for obtaining pending coloured image；Denoising unit 302, for institute
State coloured image and carry out edge denoising；Gray scale conversion unit 303, for the coloured image after edge denoising
Gray value conversion processing is carried out, obtains gray level image.
Wherein, when the denoising unit carries out edge denoising to the coloured image, it is specifically used for：To described
Coloured image carries out edge swell processing and/or edge corrosion processing.
It is specific to use when the denoising unit carries out edge swell processing and edge corrosion processing to the coloured image
In：According to default kernel function dot matrix combined value, edge swell processing is carried out to the coloured image；According to default core letter
Number, edge corrosion processing is carried out to the coloured image after edge expansion process.
Specifically, the specific works content of the unit in the present embodiment, the content of above method embodiment is referred to,
Here is omitted.
Fig. 4 shows the structural representation of another image processing apparatus disclosed in the embodiment of the present invention, shown in Fig. 3
On the architecture basics of image processing apparatus, the image processing apparatus shown in Fig. 4 also includes：Recognition unit 304, it is described for identifying
The image type of coloured image；If the coloured image is Chinese text image, control the denoising unit 302 right
The coloured image carries out edge denoising；If the coloured image is English text image, the ash is directly controlled
Spend conversion unit 303 and gray value conversion processing is carried out to the coloured image, obtain gray level image.
Specifically, the specific works content of the unit in the present embodiment, the content of above method embodiment is referred to,
Here is omitted.
Fig. 5 shows the structural representation of another image processing apparatus disclosed in the embodiment of the present invention, shown in Fig. 3
On the architecture basics of image processing apparatus, the image processing apparatus shown in Fig. 4 also includes：Normalization unit 305, for described
The gray value of each pixel of gray level image is normalized；Opening operation processing unit 306, for normalized
The gray value of each pixel afterwards carries out opening operation processing；Renormalization unit 307, after split calculation process
The gray value of each pixel carries out renormalization processing.
Specifically, the specific works content of the unit in the present embodiment, the content of above method embodiment is referred to,
Here is omitted.
The foregoing description of the disclosed embodiments, professional and technical personnel in the field are enable to realize or using the present invention.
A variety of modifications to these embodiments will be apparent for those skilled in the art, as defined herein
General Principle can be realized in other embodiments without departing from the spirit or scope of the present invention.Therefore, it is of the invention
The embodiments shown herein is not intended to be limited to, and is to fit to and principles disclosed herein and features of novelty phase one
The most wide scope caused.
Claims (10)
- A kind of 1. image processing method, it is characterised in that including：Obtain pending coloured image；Edge denoising is carried out to the coloured image；Gray value conversion processing is carried out to the coloured image after edge denoising, obtains gray level image.
- 2. according to the method for claim 1, it is characterised in that after pending image is obtained, this method also includes：Identify the image type of the coloured image；If the coloured image is Chinese text image, edge denoising is carried out to the coloured image；If the coloured image is English text image, gray value conversion processing directly is carried out to the coloured image, obtained To gray level image.
- 3. method according to claim 1 or 2, it is characterised in that edge denoising, bag are carried out to the coloured image Include：Edge swell processing and/or edge corrosion processing are carried out to the coloured image.
- 4. according to the method for claim 3, it is characterised in that edge swell processing and edge are carried out to the coloured image Corrosion treatment, including：According to default kernel function dot matrix combined value, edge swell processing is carried out to the coloured image；According to default kernel function, edge corrosion processing is carried out to the coloured image after edge expansion process.
- 5. method according to claim 1 or 2, it is characterised in that after gray level image is obtained, this method also includes：The gray value of each pixel of the gray level image is normalized；Opening operation processing is carried out to the gray value of each pixel after normalized；The gray value of each pixel after split calculation process carries out renormalization processing.
- A kind of 6. image processing apparatus, it is characterised in that including：Image acquisition unit, for obtaining pending coloured image；Denoising unit, for carrying out edge denoising to the coloured image；Gray scale conversion unit, for carrying out gray value conversion processing to the coloured image after edge denoising, obtain ash Spend image.
- 7. device according to claim 6, it is characterised in that the device also includes：Recognition unit, for identifying the image type of the coloured image；If the coloured image is Chinese text image, the denoising unit is controlled to carry out side to the coloured image Edge denoising；If the coloured image is English text image, the gray scale conversion unit is directly controlled to enter the coloured image Row gray value conversion processing, obtains gray level image.
- 8. the device according to claim 6 or 7, it is characterised in that the denoising unit enters to the coloured image During the denoising of row edge, it is specifically used for：Edge swell processing and/or edge corrosion processing are carried out to the coloured image.
- 9. device according to claim 8, it is characterised in that the denoising unit carries out side to the coloured image When edge expansion process and edge corrosion are handled, it is specifically used for：According to default kernel function dot matrix combined value, edge swell processing is carried out to the coloured image；According to default kernel function, edge corrosion processing is carried out to the coloured image after edge expansion process.
- 10. the device according to claim 6 or 7, it is characterised in that the device also includes：Normalization unit, the gray value for each pixel to the gray level image are normalized；Opening operation processing unit, for carrying out opening operation processing to the gray value of each pixel after normalized；Renormalization unit, the gray value for each pixel after split calculation process carry out renormalization processing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710697106.5A CN107563973A (en) | 2017-08-15 | 2017-08-15 | A kind of image processing method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710697106.5A CN107563973A (en) | 2017-08-15 | 2017-08-15 | A kind of image processing method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107563973A true CN107563973A (en) | 2018-01-09 |
Family
ID=60974532
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710697106.5A Pending CN107563973A (en) | 2017-08-15 | 2017-08-15 | A kind of image processing method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107563973A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109597592A (en) * | 2018-12-03 | 2019-04-09 | 珠海趣印科技有限公司 | A kind of App text intelligence Enhancement Method |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104809733A (en) * | 2015-05-08 | 2015-07-29 | 中北大学 | Ancient building wall polluted inscription character image edge extraction method |
| CN106960427A (en) * | 2016-01-11 | 2017-07-18 | 中兴通讯股份有限公司 | The method and apparatus of image in 2 D code processing |
-
2017
- 2017-08-15 CN CN201710697106.5A patent/CN107563973A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104809733A (en) * | 2015-05-08 | 2015-07-29 | 中北大学 | Ancient building wall polluted inscription character image edge extraction method |
| CN106960427A (en) * | 2016-01-11 | 2017-07-18 | 中兴通讯股份有限公司 | The method and apparatus of image in 2 D code processing |
Non-Patent Citations (2)
| Title |
|---|
| 王金凤 等: ""基于数学形态学的彩色图像边缘检测"", 《工程图学学报》 * |
| 黄海龙 等: ""一种基于数学形态学的签名真伪鉴别方法"", 《东北大学学报（自然科学版）》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109597592A (en) * | 2018-12-03 | 2019-04-09 | 珠海趣印科技有限公司 | A kind of App text intelligence Enhancement Method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108416377B (en) | Information extraction method and device in histogram | |
| CN110570396B (en) | Industrial product defect detection method based on deep learning | |
| CN103679169B (en) | The extracting method of characteristics of image and device | |
| CN110880000B (en) | Picture character positioning method and device, computer equipment and storage medium | |
| CN112597495B (en) | Malicious code detection method, system, equipment and storage medium | |
| CN111369550A (en) | Image registration and defect detection method, model, training method, device and equipment | |
| CN107169508B (en) | Cheong-sam image emotion semantic recognition method based on fusion features | |
| CN105868797A (en) | Network parameter training method, scene type identification method and devices | |
| CN113436284B (en) | Image processing method, device, computer equipment and storage medium | |
| CN108470350A (en) | Broken line dividing method in line chart and device | |
| CN109920018A (en) | Black-and-white photograph color recovery method, device and storage medium neural network based | |
| CN110400362A (en) | A kind of ABAQUS two dimension crack modeling method, system and computer readable storage medium based on image | |
| CN111489322A (en) | Method and device for adding sky filter to static picture | |
| CN111127360A (en) | Gray level image transfer learning method based on automatic encoder | |
| CN109389607A (en) | Ship Target dividing method, system and medium based on full convolutional neural networks | |
| CN102567969B (en) | Color image edge detection method | |
| CN108681989A (en) | A kind of image data Enhancement Method and device | |
| CN113223098B (en) | Preprocessing optimization method for image color classification | |
| CN107563973A (en) | A kind of image processing method and device | |
| CN117670860A (en) | Photovoltaic glass defect detection method and device | |
| CN112767508B (en) | Pseudo-color display method, system, equipment and medium for respiratory gating image | |
| CN109141457A (en) | Navigate appraisal procedure, device, computer equipment and storage medium | |
| CN115861349A (en) | Color image edge extraction method based on reduction concept structural elements and matrix sequence | |
| CN114387315A (en) | Image processing model training method, image processing device, image processing equipment and image processing medium | |
| CN108268533A (en) | A kind of Image Feature Matching method for image retrieval |
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: 20180109 |
|
| RJ01 | Rejection of invention patent application after publication |