WO2021081688A1 - Procédé et dispositif de détection de rupture de stock sur la base d'étiquettes de prix - Google Patents

Procédé et dispositif de détection de rupture de stock sur la base d'étiquettes de prix Download PDF

Info

Publication number
WO2021081688A1
WO2021081688A1 PCT/CN2019/113556 CN2019113556W WO2021081688A1 WO 2021081688 A1 WO2021081688 A1 WO 2021081688A1 CN 2019113556 W CN2019113556 W CN 2019113556W WO 2021081688 A1 WO2021081688 A1 WO 2021081688A1
Authority
WO
WIPO (PCT)
Prior art keywords
shelf
stock detection
stock
grid
image
Prior art date
Application number
PCT/CN2019/113556
Other languages
English (en)
Chinese (zh)
Inventor
庄艺唐
苏汛沅
鲜霞
Original Assignee
汉朔科技股份有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 汉朔科技股份有限公司 filed Critical 汉朔科技股份有限公司
Priority to PCT/CN2019/113556 priority Critical patent/WO2021081688A1/fr
Publication of WO2021081688A1 publication Critical patent/WO2021081688A1/fr

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features

Definitions

  • the present invention relates to the technical field of data detection and processing, in particular to a method and device for out-of-stock detection based on price tags.
  • Commodities are the most important part of the supermarket.
  • the supermarket In order to replenish the goods in time, the supermarket will usually send tally staff to inspect the goods on the shelves.
  • Real-time monitoring of the goods can improve the efficiency of the supermarket’s replenishment, which not only improves the beauty of the shelves, but also increases the supermarket’s Sales volume, which brings huge profits to the supermarket.
  • the second type of gravity sensor method uses shelf gravity to detect out-of-stock products.
  • the embodiment of the present invention provides a method for out-of-stock detection based on price tags to improve the efficiency and accuracy of out-of-stock detection and reduce the cost of out-of-stock detection.
  • the method includes:
  • the shelf grid coordinate information predetermined based on the position information of the price tag, and the out-of-stock detection model generated by pre-training the out-of-stock detection result is obtained;
  • the out-of-stock detection model is based on Multiple out-of-stock detection samples of preset scenes are pre-trained and generated.
  • the embodiment of the present invention also provides an out-of-stock detection device based on price tags to improve the efficiency and accuracy of out-of-stock detection and reduce the cost of out-of-stock detection.
  • the device includes:
  • the collection unit is used to collect images of the shelf to be tested
  • the occlusion detection unit is used to input the image of the shelf to be detected into the obstacle detection model generated by pre-training to obtain the shelf grid image that is not blocked by the obstacle; the obstacle detection model is based on a plurality of preset scenes Obstacle detection samples are pre-trained and generated;
  • the out-of-stock detection unit is used to obtain the out-of-stock detection result according to the shelf grid image that is not blocked by obstacles, the shelf grid coordinate information predetermined based on the position information of the price tag, and the pre-trained out-of-stock detection model;
  • the out-of-stock detection model is pre-trained and generated according to a plurality of out-of-stock detection samples in a preset scene.
  • the embodiment of the present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor.
  • a computer device including a memory, a processor, and a computer program stored in the memory and capable of running on the processor.
  • the processor executes the computer program, the above-mentioned defect based on the price tag is realized. Goods inspection method.
  • the embodiment of the present invention also provides a computer-readable storage medium, the computer-readable storage medium stores a computer program that executes the above-mentioned price tag-based out-of-stock detection method.
  • the out-of-stock detection solution provided by the embodiment of the present invention passes: collection to be detected The image of the shelf; the image of the shelf to be inspected is input into the obstacle detection model generated by pre-training to obtain the shelf grid image that is not blocked by the obstacle; the obstacle detection model is based on multiple obstacle detection samples in the preset scene.
  • Training generation according to the shelf grid image that is not blocked by obstacles, the shelf grid coordinate information predetermined based on the position information of the price tag, and the pre-trained out-of-stock detection model to obtain the out-of-stock detection result; the out-of-stock detection
  • the model is pre-trained and generated based on multiple out-of-stock detection samples in preset scenarios, which realizes out-of-stock detection based on price tags, improves the accuracy of out-of-stock detection, and does not need to make a lot of improvements to supermarket shelves, and reduces out-of-stock The cost of testing.
  • the price tag-based out-of-stock detection scheme provided by the embodiment of the present invention improves the efficiency and accuracy of out-of-stock detection, and reduces the cost of out-of-stock detection.
  • FIG. 1 is a schematic flow chart of a method for out-of-stock detection based on price tags in an embodiment of the present invention
  • FIG. 2 is a schematic diagram of the principle of a method for detecting out of stock based on a price tag in an embodiment of the present invention
  • FIG. 3 is a schematic diagram of the principle of generating an obstacle detection model in an embodiment of the present invention.
  • FIG. 4 is a schematic diagram of the principle of pre-determining the coordinate information of the shelf grid in an embodiment of the present invention.
  • FIG. 5 is a schematic diagram of the principle of generating a stock-out detection model in an embodiment of the present invention.
  • FIG. 6 is a schematic diagram of the labeling of out-of-stock detection sample data when the out-of-stock detection model is generated in an embodiment of the present invention
  • FIG. 7 is a schematic diagram of a grid area type map predicted by a stock-out detection model in an embodiment of the present invention.
  • Fig. 8 is a schematic structural diagram of an out-of-stock detection device based on a price tag in an embodiment of the present invention.
  • the first is the manual inspection method: a dedicated tally is used to continuously inspect the supermarket and count the lack of stock on the shelves.
  • this type of method has high detection accuracy, supermarkets need to hire special tally staff to conduct continuous inspections.
  • special tally staff to conduct continuous inspections.
  • the workload and labor intensive not only is the workload and labor intensive, but also the tally staff cannot guarantee timely inspections. It takes a long time from shortage to statistics and actual replenishment, which brings a certain economic loss to the supermarket.
  • the second method is the gravity sensor method.
  • Gravity sensors are installed on the supermarket shelves, and the shortage detection is realized according to the change of the gravity of the shelves.
  • This type of method can provide early warning of shortages in time, but this type of method needs to update the detection shelf, and a large number of gravity sensors must be installed on the commodity shelf, which increases the deployment cost.
  • the shelf changes are relatively large, and the investment
  • the gravity sensor method needs to set the gravity warning threshold for it. It is difficult to set the threshold every time you change the placement of the goods on the shelf, which limits the adjustment of the goods on the shelf at any time, so it has huge limitations.
  • Deep learning technology is more and more frequently used in image detection and recognition.
  • Target detection and recognition based on deep learning is a very popular research direction in the field of image processing in recent years. Deep learning is widely used in various fields. industry.
  • the box grid map is generated by signing the position, and then the box grid image is input into the deep learning neural network for out-of-stock detection, and finally the out-of-stock is judged for early warning. It solves the shortcomings of the current manual detection and gravity sensor out-of-stock detection methods, such as too much manpower, huge financial resources, low out-of-stock detection accuracy, especially for small objects that are not well detected, and the missed detection rate is high, small The product cannot be identified and so on. Therefore, the price tag-based out-of-stock detection scheme provided by the embodiment of the present invention not only speeds up the supermarket replenishment rate, improves the accuracy of out-of-stock detection and recognition, but also greatly saves manpower, material and financial resources, and has strong market applicability. . The following is a detailed introduction to the out-of-stock detection program based on price tags.
  • Fig. 1 is a schematic flow chart of a method for out-of-stock detection based on a price tag in an embodiment of the present invention. As shown in Fig. 1, the method includes the following steps:
  • Step 101 Collect an image of the shelf to be inspected
  • Step 102 Input the image of the shelf to be detected into an obstacle detection model generated by pre-training to obtain a shelf grid image that is not blocked by obstacles; the obstacle detection model detects multiple obstacles in a preset scene Sample pre-training generation;
  • Step 103 According to the shelf grid image that is not blocked by obstacles, the shelf grid coordinate information predetermined based on the position information of the price tag, and the pre-trained out-of-stock detection model to obtain the out-of-stock detection result;
  • the detection model is pre-trained and generated based on multiple out-of-stock detection samples in a preset scene.
  • the out-of-stock detection solution provided by the embodiment of the present invention passes: collect shelves to be inspected The image; input the image of the shelf to be detected into the obstacle detection model generated by pre-training to obtain the shelf grid image that is not blocked by obstacles; the obstacle detection model is pre-trained according to multiple obstacle detection samples in the preset scene Generated; According to the shelf grid image that is not blocked by obstacles, the shelf grid coordinate information predetermined based on the location information of the price tag, and the pre-trained out-of-stock detection model to obtain the out-of-stock detection result; the out-of-stock detection model Pre-training and generating multiple out-of-stock detection samples based on preset scenarios realizes out-of-stock detection based on price tags, improves the accuracy of out-of-stock detection, and does not need to make a lot of improvements to supermarket shelves, and
  • the price tag-based out-of-stock detection scheme provided by the embodiment of the present invention improves the efficiency and accuracy of out-of-stock detection, and reduces the cost of out-of-stock detection.
  • the aforementioned method for detecting out of stock based on price tags may further include: pre-training the generated out of stock detection model according to the following method:
  • the out-of-stock detection sample data includes a plurality of shelf grid diagrams and corresponding grid area type diagrams;
  • the tested out-of-stock detection model is tested by using the test set to obtain the out-of-stock detection model.
  • the aforementioned method for detecting out of stock based on price tags may further include: pre-training the generated obstacle detection model according to the following method:
  • the verified obstacle detection model is tested by using the test set to obtain the obstacle detection model.
  • the detailed process of the obstacle detection model, the out-of-stock detection model, and the pre-determining the coordinate information of the shelf grid includes:
  • the obstacle data labeling rules are: human occlusion, object occlusion (labeling the sample data for training obstacle detection models);
  • Collecting price tag data labeling rules the categories are two types, namely, price tags and non- Price tags (labeling the sample data of the training price tag recognition model);
  • the out-of-stock data annotations are divided into 8 categories (product status types), and the tags correspond to the out-of-stock situation of the product, ( Label the sample data for training the out-of-stock detection model).
  • the labeled data for model training obstacle detection model training, price tag recognition model training, and out-of-stock detection model training.
  • This step (1) The obstacle data described in "1" above (obstacles) Detection sample data) Enter the obstacle detection deep learning network (deep learning neural network), and train the obstacle detection model to obtain the obstacle detection model; (2) the price tag data set (price tag recognition sample data) enter the price tag recognition depth In the learning network, the price tag recognition model is obtained by training; (3) the out-of-stock data set (out-of-stock detection sample data) is input into the out-of-stock detection deep learning network, and the out-of-stock detection model is trained.
  • the step of pre-determining the coordinate information of the shelf grid take photos of all the shelves that need to be detected and then enter the price tag identification model to complete the price tag positioning. According to the identified price tag and location information, generate the shelf grid in the shelf. The generated grid coordinate information is temporarily stored, which is convenient for multiple use in subsequent out-of-stock detection.
  • the above three steps can be regarded as a one-time pre-preparation work, because the adjustment range and frequency of the shelf position and the product placement position are small, so one model training and grid generation can be used multiple times.
  • the following describes the steps of applying the aforementioned obstacle detection model, out-of-stock detection model, and pre-determined shelf grid coordinate information to perform out-of-stock detection with reference to Figure 2.
  • collecting the image of the shelf to be inspected may include: collecting the image of the shelf to be inspected by using a supermarket monitoring device.
  • supermarkets generally install monitoring equipment to obtain detailed information about the shelves, and each commodity corresponds to its own price tag.
  • the original monitoring equipment of the supermarket obtains the image of the shelf to be tested, and then performs subsequent out-of-stocks. For detection, there is no need to carry out a large number of transformations to supermarkets as in the prior art using the gravity sensor method for out-of-stock detection schemes, which saves costs.
  • inputting the image of the shelf to be detected into an obstacle detection model generated by pre-training to obtain a shelf grid image that is not blocked by obstacles may include:
  • the step of collecting the image of the shelf to be detected is re-executed.
  • obstacles can include people and objects.
  • the shelf grid coordinate information predetermined based on the position information of the price tag, and the pre-trained out-of-stock detection model the out-of-stock detection result can be obtained.
  • the predetermined coordinate information of the shelf grid determine the position information corresponding to the shelf grid that is not blocked by obstacles
  • the out-of-stock detection result is obtained.
  • obtaining the out-of-stock detection result according to the position information corresponding to the shelf grid that is not blocked by obstacles and the out-of-stock detection model generated by pre-training may include:
  • the grid area type map identify the type of commodity status in the grid (whether it is out of stock or in stock, etc.);
  • the shelf grid coordinates corresponding to the detection image read the shelf grid coordinates corresponding to the detection image, generate the shelf grid graph and input it into the pre-trained out-of-stock detection model to perform out-of-stock detection.
  • the out-of-stock detection model as shown in Figure 7 will be obtained.
  • the predicted result of the model is the type map of the trellis area.
  • the prediction result of the model is processed logically.
  • the main task of the logic processing algorithm is: according to the target category identified in the shelf (see the meaning corresponding to the label in Figure 6, that is, the product corresponding to each shelf area Status type, out of stock or not out of stock, etc.), and calculate the out-of-stock rate based on its area, where: in stock, the product in the box, and the shade indicate not out of stock; out of stock, side, and empty box indicate out of stock, But its proportion of out of stock is different; exclusions and advertising exclusions are not included in the calculation. Finally, output the out-of-stock detection result.
  • the detection accuracy of the out-of-stock detection scheme based on the price tag provided by the implementation of the present invention is as high as over 99%. It can monitor the shortage of goods on the shelves in real time, remind the replenishers to replenish the goods in time when they are out of stock, increase the beauty of the shelves, reduce a lot of manpower and financial costs, and bring huge profits to the supermarket
  • the embodiment of the present invention also provides an out-of-stock detection device based on a price tag, as described in the following embodiment. Because the principle of the price tag-based out-of-stock detection device to solve the problem is similar to the price-tag-based out-of-stock detection method, the implementation of the price tag-based out-of-stock detection device can refer to the implementation of the price tag-based out-of-stock detection method, repeat the same I won't repeat it here.
  • the term "unit” or "module” can be a combination of software and/or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented by software, implementation by hardware or a combination of software and hardware is also possible and conceived.
  • Fig. 8 is a schematic structural diagram of an out-of-stock detection device based on a price tag in an embodiment of the present invention. As shown in Fig. 8, the device includes:
  • the collection unit 01 is used to collect images of the shelf to be tested
  • the occlusion detection unit 02 is used to input the image of the shelf to be detected into an obstacle detection model generated by pre-training to obtain a shelf grid image that is not blocked by obstacles; the obstacle detection model is based on the number of preset scenes. Obstacle detection samples are pre-trained and generated;
  • the out-of-stock detection unit 03 is used to obtain the out-of-stock detection result based on the shelf grid image that is not blocked by obstacles, the shelf grid coordinate information predetermined based on the position information of the price tag, and the pre-trained out-of-stock detection model
  • the out-of-stock detection model is pre-trained and generated according to multiple out-of-stock detection samples in a preset scene.
  • the collection unit may be a supermarket monitoring device.
  • the out-of-stock detection unit may be specifically used for:
  • the supermarket shelf grid image into the pre-trained and generated price tag identification model to obtain the price tag corresponding to the supermarket shelf grid and the position information of the price tag, and obtain the predetermined shelf grid according to the position information of the price tag Coordinate information;
  • the price tag recognition model is pre-trained and generated according to multiple price tag recognition samples in a preset scene;
  • the predetermined coordinate information of the shelf grid determine the position information corresponding to the shelf grid that is not blocked by obstacles
  • the out-of-stock detection result is obtained.
  • the out-of-stock detection result is obtained according to the position information corresponding to the shelf grid that is not blocked by obstacles and the out-of-stock detection model generated by pre-training, including:
  • the grid area type map identify the type of commodity status in the grid
  • the occlusion detection unit may be specifically used for:
  • the step of collecting the image of the shelf to be detected is re-executed.
  • the aforementioned price tag-based out-of-stock detection device may further include: a training unit for pre-training the generated out-of-stock detection model according to the following method:
  • the out-of-stock detection sample data includes a plurality of shelf grid diagrams and corresponding grid area type diagrams;
  • the tested out-of-stock detection model is tested by using the test set to obtain the out-of-stock detection model.
  • the aforementioned price tag-based out-of-stock detection device may further include: an obstacle detection model training unit for pre-training the generated obstacle detection model according to the following method:
  • the verified obstacle detection model is tested by using the test set to obtain the obstacle detection model.
  • the embodiment of the present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor.
  • a computer device including a memory, a processor, and a computer program stored in the memory and capable of running on the processor.
  • the processor executes the computer program, the above-mentioned defect based on the price tag is realized. Goods inspection method.
  • the embodiment of the present invention also provides a computer-readable storage medium, the computer-readable storage medium stores a computer program that executes the above-mentioned price tag-based out-of-stock detection method.
  • the beneficial technical effects achieved by the price tag-based out-of-stock detection scheme provided by the implementation of the present invention are: the price tag-based out-of-stock detection is realized, the accuracy of the out-of-stock detection is improved, and there is no need to make a lot of improvements to the supermarket shelves. Reduce the cost of out-of-stock detection.
  • the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.
  • computer-usable storage media including but not limited to disk storage, CD-ROM, optical storage, etc.
  • These computer program instructions can also be stored in a computer-readable memory that can guide a computer or other programmable data processing equipment to work in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including the instruction device.
  • the device implements the functions specified in one process or multiple processes in the flowchart and/or one block or multiple blocks in the block diagram.
  • These computer program instructions can also be loaded on a computer or other programmable data processing equipment, so that a series of operation steps are executed on the computer or other programmable equipment to produce computer-implemented processing, so as to execute on the computer or other programmable equipment.
  • the instructions provide steps for implementing the functions specified in one process or multiple processes in the flowchart and/or one block or multiple blocks in the block diagram.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Image Analysis (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

La présente invention concerne un procédé et un dispositif de détection de rupture de stock sur la base d'étiquettes de prix. Le procédé consiste : à collecter une image d'une étagère à détecter ; à entrer l'image de l'étagère dans un modèle de détection d'obstacle généré par un pré-apprentissage afin d'obtenir une image quadrillée de l'étagère qui n'est pas bloquée par des obstacles, le modèle de détection d'obstacle étant généré par pré-apprentissage en fonction d'une pluralité d'échantillons de détection d'obstacle dans une scène prédéfinie ; et en fonction de l'image quadrillée de l'étagère qui n'est pas bloquée par des obstacles, d'informations de coordonnées quadrillées de l'étagère prédéterminées sur la base d'informations de position d'étiquettes de prix et d'un modèle de détection de rupture de stock généré par pré-apprentissage, à obtenir un résultat de détection de rupture de stock, le modèle de détection de rupture de stock étant généré par pré-apprentissage en fonction d'une pluralité d'échantillons de détection de rupture de stock dans la scène prédéfinie. La solution technique qui précède améliore l'efficacité et la précision de détection de rupture de stock, et réduit les coûts de détection de rupture de stock.
PCT/CN2019/113556 2019-10-28 2019-10-28 Procédé et dispositif de détection de rupture de stock sur la base d'étiquettes de prix WO2021081688A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2019/113556 WO2021081688A1 (fr) 2019-10-28 2019-10-28 Procédé et dispositif de détection de rupture de stock sur la base d'étiquettes de prix

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2019/113556 WO2021081688A1 (fr) 2019-10-28 2019-10-28 Procédé et dispositif de détection de rupture de stock sur la base d'étiquettes de prix

Publications (1)

Publication Number Publication Date
WO2021081688A1 true WO2021081688A1 (fr) 2021-05-06

Family

ID=75714474

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/113556 WO2021081688A1 (fr) 2019-10-28 2019-10-28 Procédé et dispositif de détection de rupture de stock sur la base d'étiquettes de prix

Country Status (1)

Country Link
WO (1) WO2021081688A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117409261A (zh) * 2023-12-14 2024-01-16 成都数之联科技股份有限公司 一种基于分类模型的元件角度分类方法及系统

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102982332A (zh) * 2012-09-29 2013-03-20 顾坚敏 基于云处理方式的零售终端货架影像智能分析系统
CN108364005A (zh) * 2018-03-07 2018-08-03 上海扩博智能技术有限公司 价格标签的自动识别方法、系统、设备及存储介质
CN109308434A (zh) * 2017-07-27 2019-02-05 浙江汉朔电子科技有限公司 电子价签、服务器及系统

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102982332A (zh) * 2012-09-29 2013-03-20 顾坚敏 基于云处理方式的零售终端货架影像智能分析系统
CN109308434A (zh) * 2017-07-27 2019-02-05 浙江汉朔电子科技有限公司 电子价签、服务器及系统
CN108364005A (zh) * 2018-03-07 2018-08-03 上海扩博智能技术有限公司 价格标签的自动识别方法、系统、设备及存储介质

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117409261A (zh) * 2023-12-14 2024-01-16 成都数之联科技股份有限公司 一种基于分类模型的元件角度分类方法及系统
CN117409261B (zh) * 2023-12-14 2024-02-20 成都数之联科技股份有限公司 一种基于分类模型的元件角度分类方法及系统

Similar Documents

Publication Publication Date Title
US11093886B2 (en) Methods for real-time skill assessment of multi-step tasks performed by hand movements using a video camera
CN102376061B (zh) 基于全方位视觉的消费者购买行为分析装置
Zhang et al. Toward new retail: A benchmark dataset for smart unmanned vending machines
CN107895160A (zh) 人脸检测与识别装置及方法
WO2020134102A1 (fr) Procédé et dispositif de reconnaissance d'article, distributeur automatique et support de stockage
CN108898109A (zh) 物品关注度的确定方法、装置和系统
CN114066900A (zh) 图像分割方法、装置、电子设备和存储介质
CN110175590A (zh) 一种商品识别方法及装置
RU2695056C1 (ru) Система и способ выявления факта потенциального мошенничества со стороны кассира, а также способ формирования выборки изображений товаров для обучения искусственной нейронной сети
Dhulekar et al. Development of bottle recycling machine using machine learning algorithm
CN109840503B (zh) 一种确定种类信息的方法及装置
CN111881894A (zh) 采集货柜商品售卖信息的方法、系统、设备和存储介质
WO2021179138A1 (fr) Procédé et système permettant d'analyser des produits sur une étagère de supermarché
Pham et al. A YOLO-based real-time packaging defect detection system
CN108629318A (zh) 一种基于图像识别技术的货架陈列识别方法、装置和系统
WO2021081688A1 (fr) Procédé et dispositif de détection de rupture de stock sur la base d'étiquettes de prix
CN110826481A (zh) 数据处理方法、商品识别方法、服务器和存储介质
Santra et al. Graph-based non-maximal suppression for detecting products on the rack
CN108898067A (zh) 确定人和物关联度的方法、装置及计算机可读存储介质
CN113065447A (zh) 一种图像集中自动识别商品的方法和设备
CN112800804A (zh) 基于价签的缺货检测方法及装置
CN111626981A (zh) 识别待检测货物类别的方法和装置
Allegra et al. Exploiting egocentric vision on shopping cart for out-of-stock detection in retail environments
Kini Real time moving vehicle congestion detection and tracking using OpenCV
WO2021030924A1 (fr) Procédé, appareil et système permettant d'identifier un article sur une étagère

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19950412

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19950412

Country of ref document: EP

Kind code of ref document: A1