CN102982332A - Retail terminal goods shelf image intelligent analyzing system based on cloud processing method - Google Patents
Retail terminal goods shelf image intelligent analyzing system based on cloud processing method Download PDFInfo
- Publication number
- CN102982332A CN102982332A CN2012103699733A CN201210369973A CN102982332A CN 102982332 A CN102982332 A CN 102982332A CN 2012103699733 A CN2012103699733 A CN 2012103699733A CN 201210369973 A CN201210369973 A CN 201210369973A CN 102982332 A CN102982332 A CN 102982332A
- Authority
- CN
- China
- Prior art keywords
- image
- sample
- retail terminal
- histogram
- shelf
- 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
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to a retail terminal goods shelf image intelligent analyzing system which comprises a sample collecting unit, a retail terminal target image collecting system, and a cloud intelligent image analyzing unit, wherein the sample collecting unit is used for collecting sample image information, the retail terminal target image collecting system is used for collecting image information of goods shelves at a retail terminal, and the cloud intelligent image analyzing unit comprises a feature extraction part, an image segmentation part, a pattern recognition part, and a data analyzing unit. The feature extraction part analyzes target images, obtains color feature data of sample images, and establishes a sample database by the utilization of the data. The feature extraction part further analyzes images collected through a target image collecting system and extracts color feature data of the target images at the retail terminal. The image segmentation part segments the target images to further recognize an outer frame of single goods. The pattern recognizing part recognizes goods of the target images. The data analyzing unit collects and/or analyzes information obtained through other parts in a cloud intelligent image analyzing system. Through the system, the retail terminal goods shelf image intelligent analyzing system based on the cloud processing method is capable of providing display and arrangement information of products quickly, reliably and in a low-cost mode.
Description
Technical field
The invention belongs to image model identification field, by the cloud processing mode image that gathers is carried out pattern-recognition and analysis, be applied to the management of retail terminal.
Background technology
In today that the retail terminal competition increasingly sharpens, consumer goods manufacturer more and more pays close attention to product in the sales situation of each retail terminal, and be well known that, product directly has influence on the consumer to the attention rate of product and brand and then affects the sales volume of product at the retail store conditions of retail terminal shelf, is that vital factor is sold in impact.Yet the collection of existing retail terminal shelf display information is normally finished collection by a large amount of manpowers to retail terminal by hand, cause like this accuracy of information and ageing all very poor, particularly along with the rise of human cost, and competition is fiercer, require in the management more accurate, more have according to, more rapid reaction, need to realize robotization with technological means.
Along with the progress of acquisition technology and network technology, the particularly raising of smart mobile phone shooting capacity, people can adopt mobile phone camera to collect necessary image information, lay the foundation for realizing robotization ground identification shelf display information.Based on this, proposed some in the prior art and obtained the technical scheme of shelf display information based on Digital Image Processing.For example: publication No. is the Chinese invention patent application of CN102184405A, the method that it provides a kind of shelf image acquisition to analyze, in the method scale is placed on the shelf, gather and identify scale epigraph information by the video camera on the mobile phone for example, then above-mentioned image is carried out analysis and calculation and obtain shelf display information.
But the method has larger limitation: at first it must adopt the shelf that add scale through special disposal, does not have a versatility, can't practical operation use, and practical application is very pessimistic; Secondly this algorithm is cut apart and the Direct Recognition monomer without monomer after according to the scale layering, and do not mention supplementary means, in the practical application, commodity on shelf be placed with various complex situations, comprise short supply, overlapping, cover etc., if do not cut apart Direct Recognition, the uncertain result such as probably cause disjunctor (leakage is cut apart), halfbody (over-segmentation), can not identify causes discrimination very low.
Publication No. provides system for the Chinese invention patent application of CN102523758A has then provided a kind of augmented reality, the request image information of tool by sending the content of shooting comprise camera to server unit, that be used for obtaining the merchandise news that is illustrated in the commodity that the shelf that can determine according to image information can display shows overlappingly with the image information that photographs by camera in display device request is replied and the merchandise news that comprises from the answer that server unit sends.Server unit is determined shelf according to the image information that comprises in request, and the size of the free space of definite shelf, from the memory storage of the dimension information of the size of the merchandise news of having stored a plurality of expression commodity and expression commodity, select the merchandise news than the large slight commodity of the free space of determined shelf relatedly, comprise the answer of selected merchandise news to the transmission of A R generator.
Yet this scheme is only for detection of the shelf free space, other commodity display information of None-identified shelf, and only identify the utilization rate that free space only is conducive to promote shelf, and sell not too large help for promoting; Secondly, the merchandise news of this system is to read from database according to the position, therefore the Sample Storehouse that does not have pattern-recognition to use, and merchandise news is constantly to change in the practice, this scheme is difficult to carry out when Sample Storehouse is not stored corresponding information, and uses the also position of None-identified commodity of this scheme; The 3rd, the space for its deployment of this system is limited, normally disposes at head store, and can't cover greatly scope.
Summary of the invention
The present invention proposes for the above-mentioned deficiency that overcomes prior art just.According to an aspect of the present invention, the problem to be solved in the present invention provides a kind of technical scheme that can process based on cloud the shelf display information of obtaining, it can remotely obtain shelf display information on a large scale, is convenient to the rationality promotion that collective analysis and management are put to improve commodity.According to a further aspect in the invention, another technical matters that the present invention will solve is to improve the recognition capability that commodity in the shelf are put information, provides more comprehensively, more accurately commodity display information.
The present invention solves above-mentioned technical matters by following technological means:
A kind of retail terminal shelf image intelligent analysis system comprises: sample collection unit, described sample collection unit be used for the collected specimens image information and with this communication to high in the clouds intelligent image analytic system; Retail terminal target image collecting unit, described retail terminal target image collecting unit is used for gathering at retail terminal the target image information of shelf; And with this communication to high in the clouds intelligent image analytic system; Intelligent image analytic system this high in the clouds intelligent image analytic system in high in the clouds comprises: feature extraction section, the target image that described feature extraction unit analytic sample collecting unit obtains, obtain the color characteristic data of sample image, utilize these data to set up and generate the sample data storehouse; This feature extraction unit is also analyzed the image that the retail terminal image capturing system gathers in addition, extracts the color characteristic data of retail terminal target image; Image segmentation section, the Image Segmentation Using that described image segmentation unit gathers described target image acquisition system, and then the outside framework of identification particular commodity; Pattern-recognition section, described pattern recognition unit utilizes the sample data storehouse to analyze the color characteristic data of described target image, the type of merchandise in the identification retail terminal target image; Data analysis unit, the merchandise news that other parts in the intelligent image analytic unit of high in the clouds are obtained gathers and/or analyzes.By said system, the present invention can realize processing based on cloud the technical scheme of the shelf display information of obtaining, and it can remotely obtain shelf display information on a large scale, is convenient to the rationality promotion that collective analysis and management are put to improve commodity.
According to a further aspect in the invention, the invention provides a kind of digital picture characteristic analysis system based on color, this system comprises: A. color read module, the color numerical value in the reading images; B. the color space quantization modules quantizes color space, and the gray level N of each color component is divided into m part; C. histogram makes up module, and it makes up color histogram after quantification; D. integral image d ties up the histogram construction part module, and it makes up histogram the histogram that makes up in the module again and launches according to the one dimension mode, has m after the expansion
dIndividual data, d represents the color component in the color space.By this system, the present invention carries out feature extraction for original image, is transformed into to have the more colouring information combination of notable feature, so that the content of this image can be accurately identified.
Preferably, this digital picture characteristic analysis system also comprises: pattern-recognition section, described pattern-recognition section utilize integral image d dimension histogram construction part module to analyze the color characteristic data of described target image, the type of merchandise in the identification retail terminal target image.
Preferably: the formula that color space is quantized is:
N
i←[g
i,g
i+1]
Wherein, i[0, m-1], g
iBe image pixel primitive color value, when its scope drops on certain interval [g
i, g
I+1] time, it is mapped as new quantification gradation N
i, original like this color value g
i(0≤i<N) is quantified as N
j(0≤j<m).。
Preferably: make up color histogram, its computing formula is: f (i)=h[N
i]
Preferably: the d dimension histogram calculation formula that makes up integral image is:
f=(f(0),f(1),...,f(m
d-1))
Wherein, p=i*m
D-1+ j*m
D-2+ ...+k, i, j, k ∈ 0,1 ..., m-1}.
Further, the present invention also provides a kind of method that adopts one of above-mentioned digital picture characteristic analysis system, comprising digital picture is gathered; Further be modeled as the supermarket light environment in this collection and eliminate reflective.
According to a third aspect of the present invention, the present invention also provides a kind of shelf image partition method, it comprises: at first make up the vertical projection histogram for the image binaryzation result, choose suitable threshold value T and cut to obtain the position of every layer of shelf in image;
Then, the horizontal projective histogram in the statistical picture under the different colours component of different colours space in order to highlight histogrammic peak value, carries out difference operation to histogram, and its formula is:
Δ
xH
i(b
x)=H
i(x)-H
i(x-1)
Wherein Hi is i the horizontal projective histogram under the color component, and the coordinate of each point is (x, H in the horizontal projective histogram
x), (x ∈ [0, W)), W is the width of shelf;
Then, process above-mentioned histogrammic each row by the logical operation OR operation and obtain final histogram, its computing formula is:
H(x)=Δ
xH
1(x)|Δ
xH
2(x)|...|Δ
xH
n(x);
Then, final histogram data is carried out normalized, so that histogrammic height is limited in certain scope, shows conveniently and analyzing and processing that its computing formula is:
Wherein, the height of IH presentation video, max
0≤i<WThe maximal value of all row of H (i) expression histogram;
By setting threshold, the histogram after the normalized is divided into a plurality of parts at last, marks the position of cut-off rule in the picture position of correspondence, realize that the outside framework of single-piece commodity is cut apart identification.
By the use of the method, the present invention can accurate quick obtains the merchandise news in the shelf image of taking exactly under non-ideal conditions.
Further, the method also comprises, utilizes the engineer's scale that obtains behind the described proportion grading that each commodity that obtains after cutting apart are carried out pattern-recognition in the scope of candidate samples storehouse.
Description of drawings
Fig. 1 is retail terminal shelf image intelligent analysis system process flow diagram
Fig. 2 is retail terminal shelf image intelligent analysis system Organization Chart
Fig. 3 is the synoptic diagram of image rotation during to correct image.
Embodiment
For technical conceive of the present invention is described better, now the preferred embodiment of the present invention is described.Yet need to prove that this embodiment only is to a kind of demonstration of the present invention's design, can not be interpreted as limiting the scope of the invention.
1. specimen sample system
For solving its technical matters, at first comprise the sample collection system in the system of the present invention, sample image is obtained by various feasible modes at retail terminal by the staff, sample in the film studio after preferably buying standard model, light adopts simulation retail termination light light film studio equipment, for example adopts area source, top side projection, adopts rotating base to support standard model; Preferably, it is reflective to adopt polariscope to eliminate.
Except the specimen sample system, system of the present invention also comprises retail terminal sample collection system, retail terminal shelf image passes through picture pick-up device by sale, being preferably smart mobile phone claps to such an extent that be transferred to cloud server shelf image data base behind the shelf image at retail terminal, preferably, transmission mode is via media such as 3G network (can in real time), Internet or storage cards (can in batches).It is reflective that wherein the polariscope elimination is preferably used in the collection of image, and take a sample to increase diversity and the representativeness of light environment at a plurality of retail terminals.
Import cloud server sample image database after the sample collection, preferably, after the sample collection and before importing cloud server sample is carried out necessary pre-service, perhaps after importing cloud server, carry out described pre-service.This pre-service comprises, such as the merging of sectional image and the correction of pattern distortion etc.The sample image that imports to cloud server extracts the feature of sample image through the feature extraction unit of high in the clouds image processing and analyzing system, use these features that the sorter of pattern recognition unit is carried out pattern drill and generate sample knowledge base and sample pattern matching parameter storehouse (pattern drill can adopt mode well known in the art to carry out, and does not launch in the present embodiment to describe).
Although can be by multiple known mode design feature extraction unit, but the present invention preferably, (it is clearly demarcated that this extraction has feature with respect to prior art to the feature extraction in sample image storehouse to adopt following " feature extraction unit ", the characteristics of identifying easily, thereby possessing outstanding substantive distinguishing features and significant progressive, is another innovation of the present invention):
Because the outstanding feature of commodity is colouring information, the different color distribution of commodity external packing that different commodity are corresponding is so the colouring information that can extract commodity is as the features training sorter.For example, the present embodiment utilization be the RGB color space information of original color image, certainly, well known to those skilled in the art is to adopt other color spaces also to be easy to realize.
A. at first color space is quantized, adopt equal interval quantizing here, the gray level N of each color component is divided into m part.The algorithmic formula that this five equilibrium adopts is:
N
i←[g
i,g
i+1]
Wherein, i[0, m-1], g
iBe image pixel primitive color value, when its scope drops on certain interval [g
i, g
I+1] time, it is mapped as new quantification gradation N
i, original like this color value g
i(0≤i<N) is quantified as N
j(0≤j<m).
B. after quantizing, make up color histogram, its computing formula is:
f(i)=h[N
i]
Wherein, f (i) is gray level N in the histogram
iFrequency for correspondence.
C. make up the d dimension histogram of integral image, again histogram is launched according to the one dimension mode, have m after the expansion
dIndividual data, the tool computing formula is:
f=(f(0),f(1),...,f(m
d-1))
Wherein, p=i*m
D-1+ j*m
D-2+ ... + k.i, j, k ∈ 0,1 ..., m-1}.The d here generally gets 3, and expression color space (RGB/HSV/LAR) has 3 color components, and other monochrome or polychrome color component also can be suitable for certainly.
D. extract the above-mentioned characteristic of every width of cloth image in the Sample Storehouse, the training classifier model.
2. retail terminal image capturing system
Except the specimen sample system, system of the present invention also comprises retail terminal shelf image capturing system, this system can realize that the retail terminal shelf display master data and collect in advance for example retail terminal master data tabulation, geographical location information, shelf placing structure, shelf master information extraction or input.This system can also distribute the retail terminal area of responsibility to the terminal data collector.Arrive the retail terminal scene the terminal data collector, obtain the shelf image by the image acquisition equipment in this system, preferably, this equipment can be the smart mobile phone with camera function, further, this mobile phone possesses the module of automatically determining the retail terminal at its place, and the location of place retail terminal also can be selected by hand by the data acquisition personnel certainly.All images that obtain carry out merger by retail terminal and timestamp.
Then, the shelf image is transferred to the original storehouse of shelf view data, high in the clouds by media (can realize in batches) such as 3G express network (can realize in real time) or storage cards.
Preferably, by processing module the shelf image of segmentation is carried out pre-service, for example suitable adjustment and Image Mosaics producing the image of whole horizontal shelf, and import (synthesizing) shelf image data base.
More preferably, when there is defective in the angle of the picture of taking, because each layer of shelf dividing plate is parallel lines, can obtain the information of detected each bar straight line by straight-line detection, determine detected principal direction according to this straight line information, and determine the angle of rotation according to the angle of this principal direction and horizontal direction, according to the image rotation principle to correct image.As shown in Figure 3, rotation formula is:
x
1=cosθx
0-sinθy
0
y
1=sinθx
0+cosθy
0
This correction can realize that at retail terminal also the intelligent image analytic system realizes beyond the clouds.
3. high in the clouds intelligent image Treatment Analysis system
High in the clouds intelligent image Treatment Analysis system comprises:
3.1 image characteristics extraction unit
The image characteristics extraction unit is by the characteristic information in the analysis extraction image of image.Its concrete working method has a detailed description in aforesaid sample analysis system, does not carry out repetition herein.
3.2 cutting unit
Although can design cutting unit by multiple known mode, but the present invention preferably, adopt following " cutting unit " to target image cut apart (this cut apart with respect to prior art have identification accurately, the efficient high, thereby possessing outstanding substantive distinguishing features and significant progressive, is another innovation of the present invention):
This cutting unit comprises:
3.2.1 the level segmenting system of shelf: this module is extracted the shelf image from database, because an original shelf image comprises sandwich construction, so before all identifyings, it need to be divided into the image that several every width of cloth only comprise one deck shelf.Here the method for taking is to make up the vertical projection histogram for the image binaryzation result, chooses suitable threshold value T and obtains the position of every layer of shelf in image and cut.
3.2.2 cutting apart of individual layer shelf: behind the image that has obtained several individual layer shelf, need to carry out multiple goal identification, namely tell the outside framework of particular commodity in the individual layer shelf, before identifying, need to determine the target area size, because the difference of different commodity on color distribution has certain differentiation, so the method that adopts here is:
A) horizontal projective histogram under the different colours component of different colours space in the statistical picture is considered three color space commonly used: RGB, HSV and LAB here.In order to highlight histogrammic peak value, can carry out the differentiate operation to histogram, preferably difference replaces differentiate, and its formula is:
Δ
xH
i(b
x)=H
i(x)-H
i(x-1)
Hi is i the horizontal projective histogram under the color component, and the coordinate of each point is (x, H in the horizontal projective histogram
x), (x ∈ [0, W)), W is the width of shelf.
Then, process histogrammic each row by the logical operation OR operation and obtain final histogram.Its computing formula is:
H(x)=Δ
xH
1(x)|Δ
xH
2(x)|…|Δ
xH
n(x)
Again final histogram data is carried out normalized, so that histogrammic height is limited in certain scope, shows conveniently and analyzing and processing that its computing formula is:
Wherein, the height of IH presentation video, max
0≤i<WThe maximal value of all row of H (i) expression histogram.
B) setting threshold, with the processing of setting threshold, will through step a) histogram after processing be divided into a plurality of parts, mark the position of cut-off rule in the picture position of correspondence, the outside framework of single-piece commodity is cut apart identification and has just been realized like this.Preferably, the parameter informations such as the commodity size that this split position is obtained in conjunction with the proportion grading unit, ratio further improve the accuracy rate of cutting apart.
3.3 pattern recognition unit:
This pattern recognition unit comprises sorter, and the sample image feature that the applicable characteristic extraction unit extracts is carried out pattern drill generation sample knowledge base and sample pattern matching parameter storehouse to the sorter of pattern recognition unit.
Each the commodity zone that will obtain after will cutting apart utilizes feature extraction unit to extract feature, and carries out pattern-recognition in selected candidate's sample database scope, identifies concrete merchandise news, for example, and the kind of information of commodity.Preferably, this pattern-recognition adopts self-adaptation to adjust parameter and algorithm, can improve constantly so intelligent automatic identification level.
3.4 proportion grading unit:
Preferably, high in the clouds intelligent image Treatment Analysis system also comprises the proportion grading unit, and this proportion grading unit preferably adopts object of reference and artificial demarcation mode, and the proportional sizes of commodity is set up foundation in the analysis image, and auxiliary above-mentioned all are cut apart and identified.
3.5 tag recognition unit:
More preferably, high in the clouds intelligent image Treatment Analysis system also comprises the tag recognition unit, and this tag recognition unit obtains relevant merchandise news by the identification of bar code, two-dimension code, commodity packaging literal, price tag.
3.6 other functional unit
High in the clouds intelligent image analytic system selectively possesses following function: after the retail terminal image capturing system is transferred to the original storehouse of shelf view data, high in the clouds, high in the clouds intelligent image Treatment Analysis system detects whether there are corresponding commodity in the available sample storehouse automatically, if there is no, automatic-prompting terminal data collector collecting sample picture.Subsequently by depositing database in behind the image characteristics extraction element analysis the preceding.
In addition by above-mentioned retail terminal shelf image intelligent analysis system, can extract the shelf critical data that the client is concerned about, can analyze with the dimension of zone, commodity, shop, rival, time based on above-mentioned data, launch and collect, and provide corresponding recommendation on improvement.This analysis can adopt known various analysis modes and module to realize.
The beneficial effect that the present invention brings has provided the information that a kind of fast, reliably and cheaply effective means provides product display to arrange to consumer goods manufacturer, fully digitalization, improve first-line staff's work efficiency, can increase the sales region that can cover, it is accurate to improve data, reduction also can be reviewed in the scene, gets rid of interference from human factor.
In sum, retail terminal shelf image intelligent analysis system be one for the important invention that improves retail trade management level and efficient, gather around and have broad application prospects.Just preferred implementation of the present invention under present technical conditions described herein; under the prerequisite that does not break away from know-why of the present invention, step, function, application and implementation framework; also have sizable lifting and development space, these improvement and distortion also should be considered as the protection domain of this patent.
Claims (14)
1. retail terminal shelf image intelligent analysis system comprises:
Sample collection unit, described sample collection unit be used for the collected specimens image information and with this communication to high in the clouds intelligent image analytic system;
Retail terminal target image collecting unit, described retail terminal target image collecting unit is used for gathering at retail terminal the target image information of shelf; And with this communication to high in the clouds intelligent image analytic system;
Intelligent image analytic system this high in the clouds intelligent image analytic system in high in the clouds comprises:
Feature extraction section, the target image that described feature extraction unit analytic sample collecting unit obtains, the color characteristic data of acquisition sample image are utilized this data to set up and are generated the sample data storehouse; This feature extraction unit is also analyzed the image that the retail terminal image capturing system gathers in addition, extracts the color characteristic data of retail terminal target image;
Image segmentation section, the Image Segmentation Using that described image segmentation section gathers described target image acquisition system, and then the outside framework of identification particular commodity;
Pattern-recognition section, described pattern-recognition section utilizes the sample data storehouse to analyze the color characteristic data of described target image, the type of merchandise in the identification retail terminal target image;
Data analysis unit, the merchandise news that other parts in the intelligent image analytic unit of high in the clouds are obtained gathers and/or analyzes.
2. require described system such as right 1, it is characterized in that: the sample collection unit is by film studio sampling and/or retail terminal spot sampling; And the environmental simulation of will photographing is the supermarket light environment and eliminates reflective.
3. require described system such as right 1-2, it is characterized in that: the sample collection unit uses the solid background cancellation to shear sampling.
4. require described system such as right 1-3, it is characterized in that: described sample collection unit and/or target image collecting unit, by internet, mobile network or storage medium sample image data storehouse, high in the clouds and/or the original storehouse of shelf destination image data with described sample image and/or described target image input high in the clouds intelligent image analytic unit.
5. require described system such as right 1-4, it is characterized in that: sample data storehouse image is carried out pattern drill according to the COLOR COMPOSITION THROUGH DISTRIBUTION in the multi-color space by sorter, set up sample knowledge base and sample pattern matching parameter storehouse.
6. require described system such as right 1-5, it is characterized in that: described sample collection unit, retail terminal target image collecting unit adopt mobile phone cam to obtain image.
7. require described system such as right 1-6, it is characterized in that: retail terminal target image collecting unit, in collection, the whole row's shelf of retail terminal are carried out segmentation and take.
8. require described system such as right 1-7, it is characterized in that: when in described system detects Sample Storehouse, not having commodity corresponding to described target image, immediately take a sample.
9. require described system such as right 1-8, it is characterized in that: the method that adopts segmentation to take gathers described target image, described system carries out geometry correction to the target image that collects and synthesizes complete horizontal shelf image, imports shelf destination image data storehouse.
10. require described system such as right 1-9, it is characterized in that: high in the clouds intelligent image analytic unit also comprises proportion grading section, described system, mark the position of cut-off rule in the picture position of correspondence according to the difference of different commodity on color distribution, the engineer's scale that obtains in conjunction with the proportion of utilization analysis portion is again realized cutting apart of single-piece commodity.
11. require described system such as right 1-10, it is characterized in that: high in the clouds intelligent image analytic unit also comprises proportion grading section, and the engineer's scale that utilizes described proportion grading section to obtain carries out pattern-recognition to each commodity that obtains after cutting apart in the scope of candidate samples storehouse.
12. require described system such as right 1-11, it is characterized in that: high in the clouds intelligent image analytic unit also comprises tag recognition section, and this tag recognition section obtains relevant merchandise news by the identification of bar code, two-dimension code, commodity packaging literal, price tag.
13. require described system such as right 1-12, it is characterized in that: described feature extraction section comprises:
A. the color space quantization modules quantizes color space, and the gray level N of each color component is divided into m part, and the algorithmic formula of employing is:
N
i←[g
i,g
i+1]
Wherein, i[0, m-1], g
iBe image pixel primitive color value, when its scope drops on certain interval [g
i, g
I+1] time, it is mapped as new quantification gradation N
i, original like this color value g
i(0≤i<N) is quantified as N
j(0≤j<m);
B. histogram makes up module, and it makes up color histogram after quantification, and its computing formula is:
f(i)=h[N
i]
Wherein, f (i) is gray level N in the histogram
iFrequency for correspondence;
C. integral image d ties up the histogram construction part module, and it makes up histogram the histogram that makes up in the module again and launches according to the one dimension mode, has m after the expansion
dIndividual data, its computing formula is:
f=(f(0),f(1),...,f(m
d-1))
Wherein, p=i*m
D-1+ j*m
D-2+ ... + k, i, j, k ∈ 0,1 ..., m-1}, d represent the color component in the color space.
14. require described system such as right 1-13, it is characterized in that: described image segmentation radicals by which characters are arranged in traditional Chinese dictionaries make up the vertical projection histogram for the image binaryzation result first, choose suitable threshold value T and cut to obtain the position of every layer of shelf in image;
Then add up the horizontal projective histogram under the different colours component of different colours space in the individual layer shelf image, histogram is carried out difference operation, in order to highlight histogrammic peak value, its formula is:
Δ
xH
i(b
x)=H
i(x)-H
i(x-1)
Wherein Hi is i the horizontal projective histogram under the color component, and the coordinate of each point is (x, H in the horizontal projective histogram
x), (x ∈ [0, W)), W is the width of shelf;
Then, process above-mentioned histogrammic each row by the logical operation OR operation and obtain final histogram, its computing formula is:
H(x)=Δ
xH
1(x)|Δ
xH
2(x)|...|Δ
xH
n(x);
Then, final histogram data is carried out normalized, so that histogrammic height is limited in certain scope, shows conveniently and analyzing and processing that its computing formula is:
Wherein, the height of IH presentation video, max
0≤i<wThe maximal value of all row of H (i) expression histogram;
By setting threshold, the histogram after the normalized is divided into a plurality of parts at last, marks the position of cut-off rule in the picture position of correspondence, realize that the outside framework of single-piece commodity is cut apart identification.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2012103699733A CN102982332A (en) | 2012-09-29 | 2012-09-29 | Retail terminal goods shelf image intelligent analyzing system based on cloud processing method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2012103699733A CN102982332A (en) | 2012-09-29 | 2012-09-29 | Retail terminal goods shelf image intelligent analyzing system based on cloud processing method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN102982332A true CN102982332A (en) | 2013-03-20 |
Family
ID=47856315
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN2012103699733A Pending CN102982332A (en) | 2012-09-29 | 2012-09-29 | Retail terminal goods shelf image intelligent analyzing system based on cloud processing method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN102982332A (en) |
Cited By (51)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104331409A (en) * | 2014-09-09 | 2015-02-04 | 吴震 | A method and system for showcasing product information by a handheld electronic device |
CN104616131A (en) * | 2015-02-16 | 2015-05-13 | 北京任我在线电子商务有限公司 | Shelf management system and method |
CN104885111A (en) * | 2013-09-19 | 2015-09-02 | 松下电器(美国)知识产权公司 | Control method for displaying merchandising information on information terminal |
CN105426857A (en) * | 2015-11-25 | 2016-03-23 | 小米科技有限责任公司 | Training method and device of face recognition model |
CN106372638A (en) * | 2016-08-31 | 2017-02-01 | 佛山绿怡信息科技有限公司 | Method and system for identifying brand and model of electronic equipment |
CN107045641A (en) * | 2017-04-26 | 2017-08-15 | 广州图匠数据科技有限公司 | A kind of identification of pallets method based on image recognition technology |
WO2017161665A1 (en) * | 2016-03-23 | 2017-09-28 | 百度在线网络技术(北京)有限公司 | Image recognition method, apparatus and device, and nonvolatile computer storage medium |
CN107403332A (en) * | 2016-05-18 | 2017-11-28 | 中华电信股份有限公司 | Goods shelf fetching detection system and method |
CN108062837A (en) * | 2018-01-26 | 2018-05-22 | 浙江行雨网络科技有限公司 | A kind of unattended supermarket's commodity settlement system based on image identification |
CN108182406A (en) * | 2017-12-28 | 2018-06-19 | 仝西鲁 | The article display recognition methods of retail terminal and system |
CN108230581A (en) * | 2017-12-23 | 2018-06-29 | 浙江行雨网络科技有限公司 | A kind of commodity image dividing method of unmanned supermarket's commodity clearing |
CN108364420A (en) * | 2018-01-08 | 2018-08-03 | 阿里巴巴集团控股有限公司 | Image-recognizing method and device and electronic equipment |
CN108364209A (en) * | 2018-02-01 | 2018-08-03 | 北京京东金融科技控股有限公司 | Methods of exhibiting, device, medium and the electronic equipment of merchandise news |
CN108416403A (en) * | 2018-03-08 | 2018-08-17 | 上海扩博智能技术有限公司 | The automatic correlation method of commodity and label, system, equipment and storage medium |
CN108629318A (en) * | 2018-05-04 | 2018-10-09 | 广州图匠数据科技有限公司 | A kind of shelf display recognition methods based on image recognition technology, device and system |
CN108647553A (en) * | 2018-05-10 | 2018-10-12 | 上海扩博智能技术有限公司 | Rapid expansion method, system, equipment and the storage medium of model training image |
CN108846401A (en) * | 2018-05-30 | 2018-11-20 | 京东方科技集团股份有限公司 | Commodity detect terminal, method, system and computer equipment, readable medium |
CN108898141A (en) * | 2018-06-14 | 2018-11-27 | 中山大学 | Commodity shelf segmentation and layer information statistical method based on image detection |
CN108960202A (en) * | 2018-08-01 | 2018-12-07 | 京东方科技集团股份有限公司 | A kind of intelligent commodity shelf, judges the method that commodity stack at system |
CN109190919A (en) * | 2018-08-10 | 2019-01-11 | 上海扩博智能技术有限公司 | It is sold critical sales index generation method, system, equipment and storage medium |
CN109271920A (en) * | 2018-09-12 | 2019-01-25 | 上海了物网络科技有限公司 | The vision system and method and medium for supporting article to identify |
CN109395405A (en) * | 2018-06-27 | 2019-03-01 | 邓继红 | Parameter extraction platform based on cloud identification |
CN109409213A (en) * | 2018-09-12 | 2019-03-01 | 上海了物网络科技有限公司 | The vision system and method and medium for supporting the article of identification code to identify |
CN109410176A (en) * | 2018-09-28 | 2019-03-01 | 中民筑友科技投资有限公司 | A kind of method, apparatus and storage medium monitoring shelf assembled state |
CN109446883A (en) * | 2018-09-05 | 2019-03-08 | 北京三快在线科技有限公司 | Condition of merchandise recognition methods, device, electronic equipment and readable storage medium storing program for executing |
CN109522933A (en) * | 2018-10-19 | 2019-03-26 | 北京陌上花科技有限公司 | Commodity detection method and device for shelf |
CN109670861A (en) * | 2018-11-30 | 2019-04-23 | 深圳冠思大数据服务有限公司 | Commodity on shelf sale effect method for improving and system based on artificial intelligence |
CN109902636A (en) * | 2019-03-05 | 2019-06-18 | 上海扩博智能技术有限公司 | Commodity identification model training method, system, equipment and storage medium |
CN109948736A (en) * | 2019-04-04 | 2019-06-28 | 上海扩博智能技术有限公司 | Commodity identification model active training method, system, equipment and storage medium |
CN109977886A (en) * | 2019-03-29 | 2019-07-05 | 京东方科技集团股份有限公司 | Shelf vacancy rate calculation method and device, electronic equipment, storage medium |
CN110033452A (en) * | 2019-04-17 | 2019-07-19 | 北京云量互动科技有限公司 | Refrigerator shelf commodity statistics method and system based on image processing techniques |
CN110070299A (en) * | 2019-04-29 | 2019-07-30 | 成都盯盯科技有限公司 | A kind of commodity shelf tidiness methods of marking and system |
CN110322945A (en) * | 2019-06-10 | 2019-10-11 | 广西臻笛生物科技有限公司 | A kind of intelligent screening method of plant compounding |
CN110427985A (en) * | 2019-07-16 | 2019-11-08 | 北京京航安机场工程有限公司 | The method for realizing machine vision training using character |
CN110490418A (en) * | 2019-07-18 | 2019-11-22 | 苏州佳世达电通有限公司 | Article information handling systems utilizing and article method for processing information |
CN110536724A (en) * | 2017-02-21 | 2019-12-03 | 天使游戏纸牌股份有限公司 | The piece number detection system of recreation substitutionary coinage |
CN110533028A (en) * | 2019-07-31 | 2019-12-03 | 北京三快在线科技有限公司 | Detection method, device, electronic equipment and the storage medium of commodity display state |
CN110569789A (en) * | 2019-09-06 | 2019-12-13 | 创新奇智(重庆)科技有限公司 | Commodity combined sku identification method and device |
CN110610149A (en) * | 2019-09-03 | 2019-12-24 | 卓尔智联(武汉)研究院有限公司 | Information processing method and device and computer storage medium |
CN111008655A (en) * | 2019-11-28 | 2020-04-14 | 上海识装信息科技有限公司 | Method and device for assisting in identifying authenticity of physical commodity brand and electronic equipment |
CN111062404A (en) * | 2018-10-17 | 2020-04-24 | 新汉智能系统股份有限公司 | Method for automatically establishing object identification model |
CN111126160A (en) * | 2019-11-28 | 2020-05-08 | 天津瑟威兰斯科技有限公司 | Intelligent Chinese character structure evaluation method and system constructed based on five-stroke input method |
CN111314588A (en) * | 2020-02-21 | 2020-06-19 | 联想(北京)有限公司 | Intelligent retail cabinet and commodity exhibition and sale platform |
CN111736583A (en) * | 2019-03-25 | 2020-10-02 | 东元电机股份有限公司 | Two-stage navigation system |
CN112149450A (en) * | 2019-06-26 | 2020-12-29 | 浙江汉朔电子科技有限公司 | Noodle arrangement inspection system and method |
WO2021081688A1 (en) * | 2019-10-28 | 2021-05-06 | 汉朔科技股份有限公司 | Price tag-based out-of-stock detection method and device |
CN113053024A (en) * | 2021-03-17 | 2021-06-29 | 广东便捷神科技股份有限公司 | Control method of unmanned retail terminal |
CN113392673A (en) * | 2020-03-12 | 2021-09-14 | 顺丰科技有限公司 | Image correction method, device, equipment and computer readable storage medium |
CN113657974A (en) * | 2021-09-02 | 2021-11-16 | 广州微行网络科技有限公司 | Electronic commercial transaction management method and system based on cloud platform |
CN114494892A (en) * | 2022-04-15 | 2022-05-13 | 广州市玄武无线科技股份有限公司 | Goods shelf commodity display information identification method, device, equipment and storage medium |
CN115796899A (en) * | 2022-11-17 | 2023-03-14 | 宋舒涵 | Regional product supply bar code management method and system |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20060029291A1 (en) * | 2004-08-09 | 2006-02-09 | Eastman Kodak Company | Multimodal image registration using compound mutual information |
CN101034442A (en) * | 2006-03-08 | 2007-09-12 | 刘欣融 | System for judging between identical and proximate goods appearance design based on pattern recognition |
CN101436252A (en) * | 2008-12-22 | 2009-05-20 | 北京中星微电子有限公司 | Method and system for recognizing vehicle body color in vehicle video image |
US20100171826A1 (en) * | 2006-04-12 | 2010-07-08 | Store Eyes, Inc. | Method for measuring retail display and compliance |
CN102184405A (en) * | 2011-04-19 | 2011-09-14 | 清华大学 | Image acquisition-analysis method |
CN102375977A (en) * | 2010-08-23 | 2012-03-14 | 东芝泰格有限公司 | Store system and sales registration method |
-
2012
- 2012-09-29 CN CN2012103699733A patent/CN102982332A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20060029291A1 (en) * | 2004-08-09 | 2006-02-09 | Eastman Kodak Company | Multimodal image registration using compound mutual information |
CN101034442A (en) * | 2006-03-08 | 2007-09-12 | 刘欣融 | System for judging between identical and proximate goods appearance design based on pattern recognition |
US20100171826A1 (en) * | 2006-04-12 | 2010-07-08 | Store Eyes, Inc. | Method for measuring retail display and compliance |
CN101436252A (en) * | 2008-12-22 | 2009-05-20 | 北京中星微电子有限公司 | Method and system for recognizing vehicle body color in vehicle video image |
CN102375977A (en) * | 2010-08-23 | 2012-03-14 | 东芝泰格有限公司 | Store system and sales registration method |
CN102184405A (en) * | 2011-04-19 | 2011-09-14 | 清华大学 | Image acquisition-analysis method |
Non-Patent Citations (1)
Title |
---|
YOUFU WU等: "A Monitoring System for Supermarket Based on Trajectory of Palm", 《INTERNATIONAL JOURNAL OF ADVANCEMENTS IN COMPUTING TECHNOLOGY》 * |
Cited By (72)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104885111A (en) * | 2013-09-19 | 2015-09-02 | 松下电器(美国)知识产权公司 | Control method for displaying merchandising information on information terminal |
CN104885111B (en) * | 2013-09-19 | 2020-03-20 | 松下电器(美国)知识产权公司 | Control method for displaying sales information on information terminal |
CN104331409A (en) * | 2014-09-09 | 2015-02-04 | 吴震 | A method and system for showcasing product information by a handheld electronic device |
CN104616131A (en) * | 2015-02-16 | 2015-05-13 | 北京任我在线电子商务有限公司 | Shelf management system and method |
CN104616131B (en) * | 2015-02-16 | 2018-12-21 | 北京任我在线电子商务有限公司 | shelf management system and method |
CN105426857A (en) * | 2015-11-25 | 2016-03-23 | 小米科技有限责任公司 | Training method and device of face recognition model |
CN105426857B (en) * | 2015-11-25 | 2019-04-12 | 小米科技有限责任公司 | Human face recognition model training method and device |
WO2017161665A1 (en) * | 2016-03-23 | 2017-09-28 | 百度在线网络技术(北京)有限公司 | Image recognition method, apparatus and device, and nonvolatile computer storage medium |
US10706314B2 (en) | 2016-03-23 | 2020-07-07 | Baidu Online Network Technology (Beijing) Co., Ltd. | Image recognition method, apparatus and device, and non-volatile computer storage medium |
CN107403332B (en) * | 2016-05-18 | 2020-11-24 | 中华电信股份有限公司 | Goods shelf fetching detection system and method |
CN107403332A (en) * | 2016-05-18 | 2017-11-28 | 中华电信股份有限公司 | Goods shelf fetching detection system and method |
CN106372638A (en) * | 2016-08-31 | 2017-02-01 | 佛山绿怡信息科技有限公司 | Method and system for identifying brand and model of electronic equipment |
CN110536724A (en) * | 2017-02-21 | 2019-12-03 | 天使游戏纸牌股份有限公司 | The piece number detection system of recreation substitutionary coinage |
CN107045641A (en) * | 2017-04-26 | 2017-08-15 | 广州图匠数据科技有限公司 | A kind of identification of pallets method based on image recognition technology |
CN108230581A (en) * | 2017-12-23 | 2018-06-29 | 浙江行雨网络科技有限公司 | A kind of commodity image dividing method of unmanned supermarket's commodity clearing |
CN108182406A (en) * | 2017-12-28 | 2018-06-19 | 仝西鲁 | The article display recognition methods of retail terminal and system |
CN108364420A (en) * | 2018-01-08 | 2018-08-03 | 阿里巴巴集团控股有限公司 | Image-recognizing method and device and electronic equipment |
CN108062837A (en) * | 2018-01-26 | 2018-05-22 | 浙江行雨网络科技有限公司 | A kind of unattended supermarket's commodity settlement system based on image identification |
CN108364209A (en) * | 2018-02-01 | 2018-08-03 | 北京京东金融科技控股有限公司 | Methods of exhibiting, device, medium and the electronic equipment of merchandise news |
CN108416403A (en) * | 2018-03-08 | 2018-08-17 | 上海扩博智能技术有限公司 | The automatic correlation method of commodity and label, system, equipment and storage medium |
CN108629318B (en) * | 2018-05-04 | 2021-03-16 | 广州图匠数据科技有限公司 | Shelf display identification method, device and system based on image identification technology |
CN108629318A (en) * | 2018-05-04 | 2018-10-09 | 广州图匠数据科技有限公司 | A kind of shelf display recognition methods based on image recognition technology, device and system |
CN108647553B (en) * | 2018-05-10 | 2022-01-25 | 上海扩博智能技术有限公司 | Method, system, device and storage medium for rapidly expanding images for model training |
CN108647553A (en) * | 2018-05-10 | 2018-10-12 | 上海扩博智能技术有限公司 | Rapid expansion method, system, equipment and the storage medium of model training image |
CN108846401A (en) * | 2018-05-30 | 2018-11-20 | 京东方科技集团股份有限公司 | Commodity detect terminal, method, system and computer equipment, readable medium |
CN108898141A (en) * | 2018-06-14 | 2018-11-27 | 中山大学 | Commodity shelf segmentation and layer information statistical method based on image detection |
CN109395405A (en) * | 2018-06-27 | 2019-03-01 | 邓继红 | Parameter extraction platform based on cloud identification |
CN108960202A (en) * | 2018-08-01 | 2018-12-07 | 京东方科技集团股份有限公司 | A kind of intelligent commodity shelf, judges the method that commodity stack at system |
US11189044B2 (en) | 2018-08-01 | 2021-11-30 | Boe Technology Group Co., Ltd. | Method and device for detecting object stacking state and intelligent shelf |
CN108960202B (en) * | 2018-08-01 | 2022-05-10 | 京东方科技集团股份有限公司 | Intelligent goods shelf, system and method for judging goods stacking |
CN109190919A (en) * | 2018-08-10 | 2019-01-11 | 上海扩博智能技术有限公司 | It is sold critical sales index generation method, system, equipment and storage medium |
CN109446883B (en) * | 2018-09-05 | 2022-08-23 | 北京三快在线科技有限公司 | Commodity state identification method and device, electronic equipment and readable storage medium |
CN109446883A (en) * | 2018-09-05 | 2019-03-08 | 北京三快在线科技有限公司 | Condition of merchandise recognition methods, device, electronic equipment and readable storage medium storing program for executing |
CN109409213A (en) * | 2018-09-12 | 2019-03-01 | 上海了物网络科技有限公司 | The vision system and method and medium for supporting the article of identification code to identify |
CN109271920A (en) * | 2018-09-12 | 2019-01-25 | 上海了物网络科技有限公司 | The vision system and method and medium for supporting article to identify |
CN109410176A (en) * | 2018-09-28 | 2019-03-01 | 中民筑友科技投资有限公司 | A kind of method, apparatus and storage medium monitoring shelf assembled state |
CN109410176B (en) * | 2018-09-28 | 2022-03-04 | 中民筑友科技投资有限公司 | Method and device for monitoring assembly state of goods shelf and storage medium |
CN111062404A (en) * | 2018-10-17 | 2020-04-24 | 新汉智能系统股份有限公司 | Method for automatically establishing object identification model |
CN109522933A (en) * | 2018-10-19 | 2019-03-26 | 北京陌上花科技有限公司 | Commodity detection method and device for shelf |
CN109670861A (en) * | 2018-11-30 | 2019-04-23 | 深圳冠思大数据服务有限公司 | Commodity on shelf sale effect method for improving and system based on artificial intelligence |
CN109902636A (en) * | 2019-03-05 | 2019-06-18 | 上海扩博智能技术有限公司 | Commodity identification model training method, system, equipment and storage medium |
CN111736583A (en) * | 2019-03-25 | 2020-10-02 | 东元电机股份有限公司 | Two-stage navigation system |
CN109977886A (en) * | 2019-03-29 | 2019-07-05 | 京东方科技集团股份有限公司 | Shelf vacancy rate calculation method and device, electronic equipment, storage medium |
WO2020199776A1 (en) * | 2019-03-29 | 2020-10-08 | 京东方科技集团股份有限公司 | Shelf vacancy rate calculation method and device, and storage medium |
CN109977886B (en) * | 2019-03-29 | 2021-03-09 | 京东方科技集团股份有限公司 | Shelf vacancy rate calculation method and device, electronic equipment and storage medium |
CN109948736A (en) * | 2019-04-04 | 2019-06-28 | 上海扩博智能技术有限公司 | Commodity identification model active training method, system, equipment and storage medium |
CN110033452A (en) * | 2019-04-17 | 2019-07-19 | 北京云量互动科技有限公司 | Refrigerator shelf commodity statistics method and system based on image processing techniques |
CN110033452B (en) * | 2019-04-17 | 2021-06-08 | 北京云量互动科技有限公司 | Refrigerator shelf commodity statistical method and system based on image processing technology |
CN110070299A (en) * | 2019-04-29 | 2019-07-30 | 成都盯盯科技有限公司 | A kind of commodity shelf tidiness methods of marking and system |
CN110322945A (en) * | 2019-06-10 | 2019-10-11 | 广西臻笛生物科技有限公司 | A kind of intelligent screening method of plant compounding |
CN112149450A (en) * | 2019-06-26 | 2020-12-29 | 浙江汉朔电子科技有限公司 | Noodle arrangement inspection system and method |
CN110427985A (en) * | 2019-07-16 | 2019-11-08 | 北京京航安机场工程有限公司 | The method for realizing machine vision training using character |
CN110490418A (en) * | 2019-07-18 | 2019-11-22 | 苏州佳世达电通有限公司 | Article information handling systems utilizing and article method for processing information |
CN110533028B (en) * | 2019-07-31 | 2021-08-13 | 北京三快在线科技有限公司 | Method and device for detecting commodity display state, electronic device and storage medium |
CN110533028A (en) * | 2019-07-31 | 2019-12-03 | 北京三快在线科技有限公司 | Detection method, device, electronic equipment and the storage medium of commodity display state |
CN110610149A (en) * | 2019-09-03 | 2019-12-24 | 卓尔智联(武汉)研究院有限公司 | Information processing method and device and computer storage medium |
CN110610149B (en) * | 2019-09-03 | 2023-01-20 | 卓尔智联(武汉)研究院有限公司 | Information processing method and device and computer storage medium |
CN110569789B (en) * | 2019-09-06 | 2023-05-05 | 创新奇智(重庆)科技有限公司 | Commodity combined sku identification method and device |
CN110569789A (en) * | 2019-09-06 | 2019-12-13 | 创新奇智(重庆)科技有限公司 | Commodity combined sku identification method and device |
WO2021081688A1 (en) * | 2019-10-28 | 2021-05-06 | 汉朔科技股份有限公司 | Price tag-based out-of-stock detection method and device |
CN111126160A (en) * | 2019-11-28 | 2020-05-08 | 天津瑟威兰斯科技有限公司 | Intelligent Chinese character structure evaluation method and system constructed based on five-stroke input method |
CN111126160B (en) * | 2019-11-28 | 2023-04-07 | 天津瑟威兰斯科技有限公司 | Intelligent Chinese character structure evaluation method and system constructed based on five-stroke input method |
CN111008655A (en) * | 2019-11-28 | 2020-04-14 | 上海识装信息科技有限公司 | Method and device for assisting in identifying authenticity of physical commodity brand and electronic equipment |
CN111314588A (en) * | 2020-02-21 | 2020-06-19 | 联想(北京)有限公司 | Intelligent retail cabinet and commodity exhibition and sale platform |
CN113392673A (en) * | 2020-03-12 | 2021-09-14 | 顺丰科技有限公司 | Image correction method, device, equipment and computer readable storage medium |
CN113053024B (en) * | 2021-03-17 | 2022-08-16 | 广东便捷神科技股份有限公司 | Control method of unmanned retail terminal |
CN113053024A (en) * | 2021-03-17 | 2021-06-29 | 广东便捷神科技股份有限公司 | Control method of unmanned retail terminal |
CN113657974A (en) * | 2021-09-02 | 2021-11-16 | 广州微行网络科技有限公司 | Electronic commercial transaction management method and system based on cloud platform |
CN113657974B (en) * | 2021-09-02 | 2024-03-01 | 万全速配网络科技有限公司 | Electronic commercial transaction management method and system based on cloud platform |
CN114494892A (en) * | 2022-04-15 | 2022-05-13 | 广州市玄武无线科技股份有限公司 | Goods shelf commodity display information identification method, device, equipment and storage medium |
CN115796899A (en) * | 2022-11-17 | 2023-03-14 | 宋舒涵 | Regional product supply bar code management method and system |
CN115796899B (en) * | 2022-11-17 | 2024-03-12 | 宋舒涵 | Regional product supply bar code management method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102982332A (en) | Retail terminal goods shelf image intelligent analyzing system based on cloud processing method | |
CN107067003B (en) | Region-of-interest boundary extraction method, device, equipment and computer storage medium | |
Zhu et al. | Concrete column recognition in images and videos | |
CN109447169A (en) | The training method of image processing method and its model, device and electronic system | |
Yang et al. | Detecting and mapping tree crowns based on convolutional neural network and Google Earth images | |
CN108345912A (en) | Commodity rapid settlement system based on RGBD information and deep learning | |
CN104166841A (en) | Rapid detection identification method for specified pedestrian or vehicle in video monitoring network | |
US11403766B2 (en) | Method and device for labeling point of interest | |
CN103617420A (en) | Commodity fast recognition method and system based on image feature matching | |
CN103208008A (en) | Fast adaptation method for traffic video monitoring target detection based on machine vision | |
CN109977983A (en) | Obtain the method and device of training image | |
CN103793717A (en) | Methods for determining image-subject significance and training image-subject significance determining classifier and systems for same | |
CN110517228A (en) | Trunk image rapid detection method based on convolutional neural networks and transfer learning | |
CN108932474B (en) | Remote sensing image cloud judgment method based on full convolution neural network composite characteristics | |
CN110852236A (en) | Target event determination method and device, storage medium and electronic device | |
CN114140665A (en) | Dense small target detection method based on improved YOLOv5 | |
CN106845458A (en) | A kind of rapid transit label detection method of the learning machine that transfinited based on core | |
CN112149348A (en) | Simulation space model training data generation method based on unmanned container scene | |
Recky et al. | Window detection in complex facades | |
CN110705620A (en) | Display image detection method and device based on copying recognition and storage medium | |
CN109492573A (en) | A kind of pointer read method and device | |
CN113378668A (en) | Method, device and equipment for determining accumulated water category and storage medium | |
CN109657728A (en) | Sample production method and model training method | |
Chen et al. | Real-time garbage object detection with data augmentation and feature fusion using SUAV low-altitude remote sensing images | |
CN108109125A (en) | Information extracting method and device based on remote sensing images |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20130320 |