CN108960318A - A kind of commodity recognizer using binocular vision technology for self-service cabinet - Google Patents
A kind of commodity recognizer using binocular vision technology for self-service cabinet Download PDFInfo
- Publication number
- CN108960318A CN108960318A CN201810686517.9A CN201810686517A CN108960318A CN 108960318 A CN108960318 A CN 108960318A CN 201810686517 A CN201810686517 A CN 201810686517A CN 108960318 A CN108960318 A CN 108960318A
- Authority
- CN
- China
- Prior art keywords
- commodity
- self
- service cabinet
- model
- yolo
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G07—CHECKING-DEVICES
- G07F—COIN-FREED OR LIKE APPARATUS
- G07F9/00—Details other than those peculiar to special kinds or types of apparatus
- G07F9/002—Vending machines being part of a centrally controlled network of vending machines
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
A kind of commodity recognizer using binocular vision technology for self-service cabinet, comprising: data preparation；Training pattern is constructed, the training pattern is used based on the yolo v3 model under TensorFlow frame；Above-mentioned model is verified, whether training of judgement result reaches expected；If being not up to expected, adjusting parameter, which optimizes and repeats training process, carries out re -training；If reaching expected, the trained yolo v3 model of arrangement is used for the identification of end article；The image of end article to be identified is obtained using the binocular camera on the self-service cabinet, described image inputs the trained yolo v3 model, identifies end article.The present invention carries out Image Acquisition by binocular camera, reduces the case where commodity are blocked or had an X-rayed；Target detection is carried out by yolo v3 model, reduces the difficulty and calculation amount of recognizer；Also, target detection is carried out by increasing different scales, accuracy of identification is improved, better recognition effect is achieved on Small object commodity.
Description
Technical field
The present invention relates to self-service cabinet and image identification technical fields, and in particular to a kind of adopting for self-service cabinet
With the commodity recognizer of binocular vision technology.
Background technique
Current many intelligence are sold goods or self-service checkout system is all simply to identify to particular commodity, this method
It is very simple, but many inconveniences also are brought to user simultaneously, for example user needs commodity successively sending one by one
Enter camera shooting area, the time for needing to wait is too long.And more commodity identifications will greatly facilitate user, user does not need
Oneself commodity are sequentially sent to fixed area to go to identify, it is only necessary to take commodity, system Automatic-settlement away.However, at present using single
Camera (monocular) more commodity take pictures know method for distinguishing usually all limit put the counters of commodity cannot be too big, if goods
Cabinet is too big, uses a camera that the image taken will be made to generate distortion, such as perspective etc..And there is distortion in image,
Many inconvenience can be brought to recognizer, for example image exists and blocks, or in the case where there is perspective, commodity recognizer
It would become hard to identify these commodity.
The Chinese patent of application number 201710982296.5 discloses the nothing of the self-service cash register of view-based access control model image recognition
People convenience store operation system, obtains the image data of commodity by way of Image Acquisition, then with the commodity image number that prestores
It is compared according to set, finds the commodity to match with the image data, transfer corresponding merchandise news, such as unit price information,
Realize commodity valuation；On-ground weigher is set in the entrance of ware room simultaneously, acquisition customer passes in and out before and after the ware room
Weight realizes that loss prevention is antitheft to control the opening and closing of exit curtain door.The patent is primarily referred to as a kind of solution party of self-help settlement machine
After client chooses commodity, commodity are sent on crawler belt one by one for case, and commodity can be sent into a module that can be taken pictures by crawler belt,
It takes pictures to commodity, then product identification of doing business of going forward side by side counts all commodity of client, then settled accounts again.The patent
Shortcoming is to be taken pictures using single camera to commodity, and commodity accuracy of identification is not high.
Existing self-service cabinet is all lesser counter mostly, and a camera can take all in counter
Commodity, and do not allow situations such as being also easy to produce perspective distortion, still, when counter is larger, a camera will to clap
It photographs the image come and generates serious distortion, this commodity identification after giving brings great difficulty.To solve this problem, existing
Have and mostly use the wider array of camera in the visual field in technology greatly, can solve the problems, such as that counter scene is excessive to a certain degree, but block
It still can not be solved with the problems such as perspective.
In order to solve the problems, such as that monocular cam be easy to cause pattern distortion when shooting commodity in big packing cupboard, the present invention is mentioned
A kind of solution shot using binocular camera and combine identification commodity is gone out.
Summary of the invention
To overcome above-mentioned the deficiencies in the prior art, the present invention provides a kind of use binocular vision skill for self-service cabinet
The commodity recognizer of art carries out Image Acquisition by binocular camera, reduces the case where commodity are blocked or had an X-rayed；Pass through yolo
V3 model carries out target detection, reduces the difficulty and calculation amount of recognizer；Also, mesh is carried out by increasing different scales
Mark detection, improves accuracy of identification, better recognition effect is achieved on Small object commodity.
The present invention is achieved by the following technical programs:
A kind of commodity recognizer using binocular vision technology for self-service cabinet, comprising the following steps:
Step 1: data preparation, a large amount of commodity figure of shooting is artificially manufactured using the binocular camera on self-service cabinet
Piece pre-processes the commodity picture, obtains the training set and verifying collection of commodity picture；
Step 2: building training pattern, the training pattern are used based on the yolo v3 mould under TensorFlow frame
Type is trained including convolution using the training set, in the training process, is carried out by Adam algorithm and SGD algorithm
Optimization, and target detection is carried out by subsequent 5 different scales of the yolo v3 model, utilize non-maxima suppression algorithm
Object detection results are screened, final output is obtained；
Step 3: verifying above-mentioned model using verifying collection, whether training of judgement result reaches expected；If not up to pre-
Phase, then adjusting parameter, which optimizes and repeats training process, carries out re -training, until achieving the desired results；If reaching pre-
Phase, then the trained yolo v3 model of arrangement is used for the identification of end article；
Step 4: obtaining the image of end article to be identified, institute using the binocular camera on the self-service cabinet
It states image and inputs the trained yolo v3 model, identify end article.
Preferably, step 2 further comprises: during being trained to the yolo v3 model, at the beginning of learning rate
Initial value is set as 0.001.
Preferably, step 2 further comprises: 5 different scales subsequent to the yolo v3 model carry out target inspection
It surveys, output result is 600 frames.
Preferably, step 2 further comprises: setting confidence level is screened using non-maxima suppression algorithm, is obtained
40 frames that confidence level sorts forward.
Preferably, step 2 further comprises: given threshold 0.7, and the model output is more than the frame of threshold value 0.7.
Preferably, step 1 further comprises: the self-service cabinet includes multi-layered storage rack, is arranged one on every layer of shelf
The top position being in contact with upper shelf is arranged in binocular camera, the binocular camera.
Preferably, step 1 further comprises: the self-service cabinet includes multi-layered storage rack, is provided with two on every layer of shelf
A monocular cam, described two monocular cams are arranged in the ipsilateral of shelf and are located at the top position being in contact with upper shelf
It sets.
Preferably, step 1 further comprises: described two monocular cams are separately positioned on shelf with the three of lateral roof
The position of Along ent.
Compared with prior art, the beneficial effects of the present invention are 1) it is set on every layer of shelf of large-scale self-service cabinet
Binocular camera is set, the joint identification that commodity image carries out commodity is shot by binocular camera, expands the visual field that can be shot,
Reduce the case where commodity are blocked or had an X-rayed, improves the utilization rate of counter；2) to based on the yolo under TensorFlow frame
V3 model improves, and increases the scale of target detection, for the more situation of Small object commodity on self-service cabinet shelf,
Target detection is carried out on 5 different scales, the recognition effect of Small object commodity is improved, improves accuracy of identification；3) it utilizes
Yolo v3 model carries out target detection, considerably reduces calculation amount and calculates the time, uses field suitable for self-service cabinet
Scape has saved settlement time when user buys commodity, has improved the use feeling of user.
Detailed description of the invention
Fig. 1 is the structural schematic diagram that target detection is carried out according to the improved yolo v3 model of the embodiment of the present invention；
Fig. 2 is the scheme of installation according to the binocular camera of the embodiment of the present invention；
Fig. 3 is to be identified according to a kind of commodity using binocular vision technology for self-service cabinet of the embodiment of the present invention
The flow chart of algorithm.
Specific embodiment
Clear, complete description is carried out below with reference to technical solution of the attached drawing to various embodiments of the present invention, it is clear that is retouched
Stating hair embodiment is only a part of the embodiments of the present invention, instead of all the embodiments.Based on the embodiment of the present invention, originally
Field those of ordinary skill obtained all other embodiment without making creative work, belongs to this hair
Bright protected range.
As shown, the present invention provides a kind of calculating using the commodity identification of binocular vision technology for self-service cabinet
Method, comprising the following steps:
Step 1: data preparation, a large amount of commodity figure of shooting is artificially manufactured using the binocular camera on self-service cabinet
Piece pre-processes the commodity picture, obtains the training set and verifying collection of commodity picture；
Step 2: building training pattern, the training pattern are used based on the yolo v3 mould under TensorFlow frame
Type is trained including convolution using the training set, in the training process, is carried out by Adam algorithm and SGD algorithm
Optimization, and target detection is carried out by subsequent 5 different scales of the yolo v3 model, utilize non-maxima suppression algorithm
Object detection results are screened, final output is obtained；
Specifically, 5 different scales subsequent to the yolo v3 model carry out target detection, and output result is 600
Frame will be more than the frame rejecting of image-region, set confidence level, be screened using non-maxima suppression algorithm, obtain confidence level
40 frames for sorting forward；Given threshold 0.7, the model final output are more than the frame of threshold value 0.7.I.e., it is believed that the threshold of output
Value is more than that 0.7 frame is enough to identify end article；
Step 3: verifying above-mentioned model using verifying collection, whether training of judgement result reaches expected；If not up to pre-
Phase, then adjusting parameter, which optimizes and repeats training process, carries out re -training, until achieving the desired results；If reaching pre-
Phase, then the trained yolo v3 model of arrangement is used for the identification of end article；
Specifically, during being trained to the yolo v3 model, the initial value of learning rate is set as 0.001；Work as instruction
When white silk result is not up to expected, adjustable learning rate re-starts optimization；The desired value of training result can self-setting, for example,
It is set as accuracy of identification and reaches 99% or more, that is, thinks to have reached expected training result；
Step 4: obtaining the image of end article to be identified, institute using the binocular camera on the self-service cabinet
It states image and inputs the trained yolo v3 model, identify end article.
As an implementation, the self-service cabinet includes multi-layered storage rack, and one binocular of setting is taken the photograph on every layer of shelf
As head, the top position being in contact with upper shelf is arranged in the binocular camera.
As an implementation, the self-service cabinet includes multi-layered storage rack, and there are two monoculars for setting on every layer of shelf
Camera, described two monocular cams are arranged in the ipsilateral of shelf and are located at the top position being in contact with upper shelf；Institute
It states two monocular cams and is separately positioned on shelf with the position of the trisection point of lateral roof.
Embodiment
As shown, providing a kind of commodity recognizer using binocular vision technology for self-service cabinet.It is described
Self-service cabinet includes multi-layered storage rack, and two cameras are arranged on every layer of shelf, i.e. a camera left side and camera is right, described two
Leaning on door side and being located at the top position contacted by door side with upper shelf for shelf is arranged in camera, described two cameras
Shooting the visual field cover flood shelf, can by the commodity in locating shelf shoot more completely, and reduce commodity block and
Have an X-rayed the appearance of phenomenon.It is provided in every layer of shelf commodity maximum value line (that is, commodity Max line), the article height in shelf is not
It can exceed that the commodity maximum value line.The self-service cabinet further includes controller and communication module, and the communication module is used
Communication connection between the self-service cabinet and remote server, the remote server include memory module and processing mould
Block prestores merchandise news, pricing information etc. in self-service cabinet in the memory module, in each user's purchasing process knot
Shu Hou, the processing module update the storage content of the memory module according to new merchandise news.
When user has purchasing demand, the cabinet door of self-service cabinet is opened, described two cameras is triggered and starts to clap for the first time
It takes the photograph, the commodity image of shooting inputs trained yolo v3 model for the first time, the commodity recognition result before obtaining user's purchase
And the commodity number of statistics；After user takes commodity away and shuts cabinet door, triggers described two cameras and start second of shooting, the
The commodity image of secondary shooting inputs trained yolo v3 model, the commodity recognition result and system after obtaining user's purchase
The commodity number of meter；User buys the commodity recognition result of front and back and the commodity number of statistics is sent to controller, the control
The commodity number of commodity recognition result and statistics is sent to remote server by communication module by device, and remote server is according to connecing
The commodity price information that the information and calling of receipts prestore obtains the price of the commodity of user's purchase and the total amount of cost, and will
Last total amount is sent to payment terminals, and payment terminals are deducted fees automatically, and the purchasing process of user terminates.
Finally, it should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although
Present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that: it still may be used
To modify to technical solution documented by previous embodiment, or some or all of the technical features are equal
Replacement；And these are modified or replaceed, technical solution of the embodiment of the present invention that it does not separate the essence of the corresponding technical solution.
Claims (8)
1. a kind of commodity recognizer using binocular vision technology for self-service cabinet, which is characterized in that including following
Step:
Step 1: data preparation, a large amount of commodity picture of shooting is artificially manufactured using the binocular camera on self-service cabinet, it is right
The commodity picture is pre-processed, and the training set and verifying collection of commodity picture are obtained；
Step 2: building training pattern, the training pattern is used based on the yolo v3 model under TensorFlow frame,
In include convolution, be trained using the training set, in the training process, optimized by Adam algorithm and SGD algorithm,
And target detection is carried out by subsequent 5 different scales of the yolo v3 model, using non-maxima suppression algorithm to target
Testing result is screened, and final output is obtained；
Step 3: verifying above-mentioned model using verifying collection, whether training of judgement result reaches expected；If being not up to expected,
Then adjusting parameter, which optimizes and repeats training process, carries out re -training, until achieving the desired results；If reaching expected,
Then the trained yolo v3 model of arrangement is used for the identification of end article；
Step 4: obtaining the image of end article to be identified, the figure using the binocular camera on the self-service cabinet
As inputting the trained yolo v3 model, end article is identified.
2. a kind of commodity recognizer using binocular vision technology for self-service cabinet as described in claim 1,
Be characterized in that, step 2 further comprises: during being trained to the yolo v3 model, the initial value of learning rate is set
It is 0.001.
3. a kind of commodity recognizer using binocular vision technology for self-service cabinet as described in claim 1,
Be characterized in that, step 2 further comprises: 5 different scales subsequent to the yolo v3 model carry out target detection, output
It as a result is 600 frames.
4. a kind of commodity recognizer using binocular vision technology for self-service cabinet as claimed in claim 3,
Be characterized in that, step 2 further comprises: setting confidence level is screened using non-maxima suppression algorithm, obtains confidence level
40 frames for sorting forward.
5. a kind of commodity recognizer using binocular vision technology for self-service cabinet as claimed in claim 4,
Be characterized in that, step 2 further comprises: given threshold 0.7, the model output are more than the frame of threshold value 0.7.
6. a kind of commodity recognizer using binocular vision technology for self-service cabinet as described in claim 1,
Be characterized in that, step 1 further comprises: the self-service cabinet includes multi-layered storage rack, and one binocular of setting is taken the photograph on every layer of shelf
As head, the top position being in contact with upper shelf is arranged in the binocular camera.
7. a kind of commodity recognizer using binocular vision technology for self-service cabinet as described in claim 1,
Be characterized in that, step 1 further comprises: the self-service cabinet includes multi-layered storage rack, and there are two monoculars for setting on every layer of shelf
Camera, described two monocular cams are arranged in the ipsilateral of shelf and are located at the top position being in contact with upper shelf.
8. a kind of commodity recognizer using binocular vision technology for self-service cabinet as claimed in claim 7,
Be characterized in that, step 1 further comprises: described two monocular cams are separately positioned on trisection point of the shelf with lateral roof
Position.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810686517.9A CN108960318A (en) | 2018-06-28 | 2018-06-28 | A kind of commodity recognizer using binocular vision technology for self-service cabinet |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810686517.9A CN108960318A (en) | 2018-06-28 | 2018-06-28 | A kind of commodity recognizer using binocular vision technology for self-service cabinet |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108960318A true CN108960318A (en) | 2018-12-07 |
Family
ID=64487418
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810686517.9A Pending CN108960318A (en) | 2018-06-28 | 2018-06-28 | A kind of commodity recognizer using binocular vision technology for self-service cabinet |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108960318A (en) |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109815896A (en) * | 2019-01-23 | 2019-05-28 | 武汉金盒子智能科技有限公司 | Parameter-configurable virtual goods channel dividing method for intelligent sales counter |
| CN109982141A (en) * | 2019-03-22 | 2019-07-05 | 李宗明 | Utilize the method for the video image region analysis and product placement of AI technology |
| CN110020511A (en) * | 2019-03-26 | 2019-07-16 | 深兰科技（上海）有限公司 | A kind of management method and device of article |
| CN110705371A (en) * | 2019-09-09 | 2020-01-17 | 上海零眸智能科技有限公司 | Refrigerator structure and commodity arrangement detection method thereof |
| CN111160248A (en) * | 2019-12-30 | 2020-05-15 | 北京每日优鲜电子商务有限公司 | Method and device for tracking articles, computer equipment and storage medium |
| CN111242094A (en) * | 2020-02-25 | 2020-06-05 | 深圳前海达闼云端智能科技有限公司 | Commodity identification method, intelligent container and intelligent container system |
| CN111292261A (en) * | 2020-01-17 | 2020-06-16 | 杭州电子科技大学 | Container detection and locking method based on multi-sensor fusion |
| CN111339887A (en) * | 2020-02-20 | 2020-06-26 | 深圳前海达闼云端智能科技有限公司 | Commodity identification method and intelligent container system |
| CN111626150A (en) * | 2020-05-11 | 2020-09-04 | 广东顺德工业设计研究院(广东顺德创新设计研究院) | Commodity identification method |
| CN112270624A (en) * | 2020-10-26 | 2021-01-26 | 链盟智能科技(广州)有限公司 | Artificial intelligence bill discernment generation system based on fast food trade |
| CN112270562A (en) * | 2020-10-15 | 2021-01-26 | 郭英 | Automatic agricultural product price adjustment and settlement method and system for agricultural product business surpassing |
| CN112906759A (en) * | 2021-01-29 | 2021-06-04 | 哈尔滨工业大学 | Pure vision-based entrance-guard-free unmanned store checkout method |
| CN113012345A (en) * | 2019-12-20 | 2021-06-22 | 深圳汇通云信息系统有限公司 | Automatic sell quick-witted shipment system |
| WO2021139237A1 (en) * | 2020-05-14 | 2021-07-15 | 平安科技（深圳）有限公司 | Method and apparatus for loss function optimization, device, and storage medium |
| CN113627393A (en) * | 2021-09-09 | 2021-11-09 | 河北工业大学 | Commodity identification method based on dual neural network and intelligent vending system |
| CN113727029A (en) * | 2021-11-03 | 2021-11-30 | 武汉星巡智能科技有限公司 | Intelligent order generation method for combining collected images at multiple visual angles and intelligent vending machine |
| CN113780441A (en) * | 2021-09-16 | 2021-12-10 | 广东佩服科技有限公司 | Method for constructing commodity identification model |
| CN115412652A (en) * | 2021-05-26 | 2022-11-29 | 哈尔滨工业大学（深圳） | Image data acquisition device and method for unmanned vending machine |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140052674A1 (en) * | 2012-08-17 | 2014-02-20 | International Business Machines Corporation | System, method and computer program product for classification of social streams |
| CN205644646U (en) * | 2016-01-11 | 2016-10-12 | 湖北师范学院 | Down unmanned intelligent purchase system of on -line line |
| CN106952402A (en) * | 2017-03-22 | 2017-07-14 | 帮团成都电子商务有限责任公司 | A kind of data processing method and device |
| CN107067510A (en) * | 2017-03-27 | 2017-08-18 | 杭州赛狐科技有限公司 | A kind of unattended Supermarket shopping system |
| CN107729923A (en) * | 2017-09-19 | 2018-02-23 | 浙江师范大学 | A kind of commodity Computer Vision Recognition method based on TensorFlow |
| CN107844790A (en) * | 2017-11-15 | 2018-03-27 | 上海捷售智能科技有限公司 | A kind of vegetable identification and POS and method based on image recognition |
| CN108154403A (en) * | 2018-01-22 | 2018-06-12 | 合肥美的智能科技有限公司 | Self-service system and its good selling method |
| CN108492464A (en) * | 2018-02-08 | 2018-09-04 | 中山简单点网络技术有限公司 | A kind of self-service cabinet and its identification commodity method |
-
2018
- 2018-06-28 CN CN201810686517.9A patent/CN108960318A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140052674A1 (en) * | 2012-08-17 | 2014-02-20 | International Business Machines Corporation | System, method and computer program product for classification of social streams |
| CN205644646U (en) * | 2016-01-11 | 2016-10-12 | 湖北师范学院 | Down unmanned intelligent purchase system of on -line line |
| CN106952402A (en) * | 2017-03-22 | 2017-07-14 | 帮团成都电子商务有限责任公司 | A kind of data processing method and device |
| CN107067510A (en) * | 2017-03-27 | 2017-08-18 | 杭州赛狐科技有限公司 | A kind of unattended Supermarket shopping system |
| CN107729923A (en) * | 2017-09-19 | 2018-02-23 | 浙江师范大学 | A kind of commodity Computer Vision Recognition method based on TensorFlow |
| CN107844790A (en) * | 2017-11-15 | 2018-03-27 | 上海捷售智能科技有限公司 | A kind of vegetable identification and POS and method based on image recognition |
| CN108154403A (en) * | 2018-01-22 | 2018-06-12 | 合肥美的智能科技有限公司 | Self-service system and its good selling method |
| CN108492464A (en) * | 2018-02-08 | 2018-09-04 | 中山简单点网络技术有限公司 | A kind of self-service cabinet and its identification commodity method |
Non-Patent Citations (3)
| Title |
|---|
| JOSEPH REDMON ET AL: "YOLOv3: An Incremental Improvement", 《COMPUTER VISION AND PATTERN RECOGNITION (CS.CV)》 * |
| 机器视觉: "和"RFID"说分手，帮团"双目立体视觉"重新定义无人店", 《搜狐》 * |
| 网友: "从零开始PyTorch项目：YOLO v3目标检测实现", 《机器之心》 * |
Cited By (29)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109815896A (en) * | 2019-01-23 | 2019-05-28 | 武汉金盒子智能科技有限公司 | Parameter-configurable virtual goods channel dividing method for intelligent sales counter |
| CN109815896B (en) * | 2019-01-23 | 2022-12-27 | 武汉嗨柜技术有限公司 | Parameter-configurable virtual goods channel dividing method for intelligent sales counter |
| CN109982141A (en) * | 2019-03-22 | 2019-07-05 | 李宗明 | Utilize the method for the video image region analysis and product placement of AI technology |
| CN109982141B (en) * | 2019-03-22 | 2021-04-23 | 李宗明 | Method for analyzing video image area and implanting advertisement by using AI technology |
| CN110020511A (en) * | 2019-03-26 | 2019-07-16 | 深兰科技（上海）有限公司 | A kind of management method and device of article |
| CN110705371A (en) * | 2019-09-09 | 2020-01-17 | 上海零眸智能科技有限公司 | Refrigerator structure and commodity arrangement detection method thereof |
| CN110705371B (en) * | 2019-09-09 | 2023-06-06 | 上海零眸智能科技有限公司 | Refrigerator structure and commodity surface arrangement detection method thereof |
| CN113012345A (en) * | 2019-12-20 | 2021-06-22 | 深圳汇通云信息系统有限公司 | Automatic sell quick-witted shipment system |
| CN113012345B (en) * | 2019-12-20 | 2022-08-05 | 深圳汇通云信息系统有限公司 | Automatic sell quick-witted shipment system |
| CN111160248A (en) * | 2019-12-30 | 2020-05-15 | 北京每日优鲜电子商务有限公司 | Method and device for tracking articles, computer equipment and storage medium |
| CN111292261A (en) * | 2020-01-17 | 2020-06-16 | 杭州电子科技大学 | Container detection and locking method based on multi-sensor fusion |
| CN111292261B (en) * | 2020-01-17 | 2023-04-18 | 杭州电子科技大学 | Container detection and locking method based on multi-sensor fusion |
| CN111339887A (en) * | 2020-02-20 | 2020-06-26 | 深圳前海达闼云端智能科技有限公司 | Commodity identification method and intelligent container system |
| CN111339887B (en) * | 2020-02-20 | 2023-07-21 | 达闼机器人股份有限公司 | Commodity identification method and intelligent container system |
| CN111242094B (en) * | 2020-02-25 | 2023-05-02 | 达闼机器人股份有限公司 | Commodity identification method, intelligent container and intelligent container system |
| CN111242094A (en) * | 2020-02-25 | 2020-06-05 | 深圳前海达闼云端智能科技有限公司 | Commodity identification method, intelligent container and intelligent container system |
| CN111626150B (en) * | 2020-05-11 | 2023-08-18 | 广东顺德工业设计研究院(广东顺德创新设计研究院) | Commodity identification method |
| CN111626150A (en) * | 2020-05-11 | 2020-09-04 | 广东顺德工业设计研究院(广东顺德创新设计研究院) | Commodity identification method |
| WO2021139237A1 (en) * | 2020-05-14 | 2021-07-15 | 平安科技（深圳）有限公司 | Method and apparatus for loss function optimization, device, and storage medium |
| CN112270562A (en) * | 2020-10-15 | 2021-01-26 | 郭英 | Automatic agricultural product price adjustment and settlement method and system for agricultural product business surpassing |
| CN112270562B (en) * | 2020-10-15 | 2023-12-26 | 郭英 | Automatic price adjustment settlement method and system for agricultural products with commercial and super agricultural products |
| CN112270624A (en) * | 2020-10-26 | 2021-01-26 | 链盟智能科技(广州)有限公司 | Artificial intelligence bill discernment generation system based on fast food trade |
| CN112906759A (en) * | 2021-01-29 | 2021-06-04 | 哈尔滨工业大学 | Pure vision-based entrance-guard-free unmanned store checkout method |
| CN115412652A (en) * | 2021-05-26 | 2022-11-29 | 哈尔滨工业大学（深圳） | Image data acquisition device and method for unmanned vending machine |
| CN115412652B (en) * | 2021-05-26 | 2024-03-22 | 哈尔滨工业大学（深圳） | Image data acquisition device and method for vending machine |
| CN113627393A (en) * | 2021-09-09 | 2021-11-09 | 河北工业大学 | Commodity identification method based on dual neural network and intelligent vending system |
| CN113627393B (en) * | 2021-09-09 | 2024-03-29 | 河北工业大学 | Commodity identification method based on dual neural network and intelligent vending system |
| CN113780441A (en) * | 2021-09-16 | 2021-12-10 | 广东佩服科技有限公司 | Method for constructing commodity identification model |
| CN113727029A (en) * | 2021-11-03 | 2021-11-30 | 武汉星巡智能科技有限公司 | Intelligent order generation method for combining collected images at multiple visual angles and intelligent vending machine |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108960318A (en) | A kind of commodity recognizer using binocular vision technology for self-service cabinet | |
| AU2018102235A4 (en) | Goods sensing system and method for goods sensing based on image monitoring | |
| US11468697B2 (en) | Pedestrian re-identification method based on spatio-temporal joint model of residual attention mechanism and device thereof | |
| CN109409175B (en) | Settlement method, device and system | |
| Liu et al. | A smart unstaffed retail shop based on artificial intelligence and IoT | |
| CN105069472B (en) | A kind of vehicle checking method adaptive based on convolutional neural networks | |
| CN109559458A (en) | Cash method and self-service cashier based on neural network recognization commodity | |
| CN105518709B (en) | The method, system and computer program product of face for identification | |
| CN108875669B (en) | Commodity identification technology based on visible light and infrared image fusion | |
| CN108198052A (en) | User's free choice of goods recognition methods, device and intelligent commodity shelf system | |
| CN108320379A (en) | Good selling method, device and the self-service machine compared based on image | |
| CN109559453A (en) | Human-computer interaction device and its application for Automatic-settlement | |
| CN109522790A (en) | Human body attribute recognition approach, device, storage medium and electronic equipment | |
| CN111222870B (en) | Settlement method, device and system | |
| CN109358922A (en) | A kind of personalized menu methods of exhibiting, device, intelligent terminal and storage medium | |
| CN109360358A (en) | Payment devices, payment system and self-service accounts settling method | |
| CN108230553A (en) | A kind of mirror image visual identity Self-help vending machine system | |
| CN109635634A (en) | A kind of pedestrian based on stochastic linear interpolation identifies data enhancement methods again | |
| CN111325051A (en) | Face recognition method and device based on face image ROI selection | |
| CN110298268B (en) | Method and device for identifying bidirectional passenger flow through single lens, storage medium and camera | |
| CN109146906A (en) | Image processing method and device, electronic equipment, computer readable storage medium | |
| CN106657937A (en) | Virtual clothes color changing device, and implementation method thereof | |
| CN110503022A (en) | A kind of personal identification method, apparatus and system | |
| CN109784230A (en) | A kind of facial video image quality optimization method, system and equipment | |
| CN109977251A (en) | A method of building identifies commodity based on RGB histogram feature |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181207 |