CN118135483A - Unmanned retail commodity identification system - Google Patents
Unmanned retail commodity identification system Download PDFInfo
- Publication number
- CN118135483A CN118135483A CN202410235621.1A CN202410235621A CN118135483A CN 118135483 A CN118135483 A CN 118135483A CN 202410235621 A CN202410235621 A CN 202410235621A CN 118135483 A CN118135483 A CN 118135483A
- Authority
- CN
- China
- Prior art keywords
- data
- commodity
- image
- unit
- commodities
- 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
- 238000007499 fusion processing Methods 0.000 claims abstract description 40
- 238000001514 detection method Methods 0.000 claims abstract description 24
- 230000003993 interaction Effects 0.000 claims abstract description 24
- 238000012545 processing Methods 0.000 claims abstract description 20
- 238000012544 monitoring process Methods 0.000 claims abstract description 10
- 238000010801 machine learning Methods 0.000 claims abstract description 8
- 230000000007 visual effect Effects 0.000 claims description 39
- 238000000034 method Methods 0.000 claims description 21
- 238000013135 deep learning Methods 0.000 claims description 18
- 230000004927 fusion Effects 0.000 claims description 9
- 238000013136 deep learning model Methods 0.000 claims description 7
- 238000012549 training Methods 0.000 claims description 7
- 230000007613 environmental effect Effects 0.000 claims description 5
- 238000003909 pattern recognition Methods 0.000 claims description 5
- 230000008859 change Effects 0.000 claims description 4
- 238000005516 engineering process Methods 0.000 description 13
- 238000007726 management method Methods 0.000 description 10
- 230000008569 process Effects 0.000 description 4
- 230000004044 response Effects 0.000 description 4
- 238000013473 artificial intelligence Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000013523 data management Methods 0.000 description 3
- 230000007246 mechanism Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
- 230000026676 system process Effects 0.000 description 1
- 230000001960 triggered effect Effects 0.000 description 1
Landscapes
- Image Analysis (AREA)
Abstract
The invention provides an unmanned retail commodity identification system, which comprises an object detection unit, an image acquisition unit, an image identification unit, a sensor array, a data fusion processing unit, an identification algorithm module and a commodity interaction database, wherein data interaction is carried out between the object detection unit, the image acquisition unit, the image identification unit, the sensor array and the commodity interaction database; the object detection unit: monitoring an unmanned retail cabinet in real time, detecting commodities taken by customers, determining the positions and moments of the commodities through a laser scanner and a motion detection sensor, and sending trigger signals to an image recognition unit when the positions and moments reach a preset area to trigger an image acquisition unit to capture commodity images; the image recognition unit is internally provided with a plurality of high-resolution cameras, and when the object detection unit detects that the commodity moves, the cameras are activated and capture images of the commodity, and the acquired images are uploaded to the image recognition unit; the image recognition unit is used for processing the image data captured by the image acquisition unit and analyzing the image by using an image processing and machine learning algorithm.
Description
Technical Field
The invention belongs to the field of unmanned retail, and particularly relates to an unmanned retail commodity identification system.
Background
The unmanned retail commodity identification system is an intelligent system based on computer vision and artificial intelligence technology, and aims to realize automatic identification and management of commodities in an unmanned retail scene. The system utilizes advanced image processing, object detection and machine learning algorithms, combines sensor technology and database management, and can realize functions of automatic identification, pricing, inventory management, user interaction and the like of commodities, thereby improving the efficiency and the intelligent level of retail business. The unmanned retail commodity identification system is a novel retail solution based on artificial intelligence and the Internet of things technology, and realizes automatic identification and metering of commodities by using computer vision and a deep learning algorithm, so that unmanned retail experience is realized. The system not only improves the efficiency and convenience of retail industry, but also brings brand new shopping experience for consumers.
The technical background of unmanned retail item identification systems has mainly resulted from the development of computer vision and deep learning techniques. Computer vision is a technology for simulating and realizing human vision functions by using devices such as a computer, a camera and the like, and can convert information in images and videos into digital signals so as to realize identification and analysis of objects. The deep learning is a machine learning method based on an artificial neural network, and can realize the recognition and analysis of complex modes through the training of a large amount of data, thereby realizing the automatic processing of complex tasks.
In unmanned retail item identification systems, computer vision and deep learning techniques are applied to the identification and metering of items. Firstly, the system shoots the commodity through equipment such as a camera, then processes and analyzes the image by utilizing a computer vision technology, and extracts characteristic information of the commodity. The system then learns and trains the feature information by using a deep learning algorithm, thereby realizing automatic identification and metering of commodities. Finally, the system can match the identified commodity information with inventory information, so that automatic selling and management of commodities are realized.
Besides computer vision and deep learning technologies, the unmanned retail commodity identification system also relates to application of the technology of the Internet of things. The internet of things is a technology for connecting and communicating objects and equipment through the internet, and can realize real-time monitoring and control of the objects. In the unmanned retail commodity identification system, the internet of things technology can connect and cooperatively work with devices such as cameras, sensors and data processing equipment, so that real-time monitoring and management of commodities are realized.
The unmanned retail commodity identification system is a novel retail solution based on computer vision, deep learning and the Internet of things technology, and realizes automatic identification and metering of commodities by using advanced technical means, so that the efficiency and convenience of retail industry are improved. With the continuous development of artificial intelligence and internet of things, it is believed that unmanned retail commodity identification systems will find wider application in the future.
Thus, there is a need for an unmanned retail product identification system.
Disclosure of Invention
The invention provides an unmanned retail commodity identification system, which solves the problem that the existing unmanned retail commodities cannot be identified aiming at some non-warehouse-in commodities in the prior art, and can cause low commodity loading efficiency and need to be accessed according to a set area when unmanned retail is carried out.
The technical scheme of the invention is realized as follows: the unmanned retail commodity identification system comprises an object detection unit, an image acquisition unit, an image identification unit, a sensor array, a data fusion processing unit, an identification algorithm module and a commodity interaction database for data interaction;
the object detection unit: monitoring an unmanned retail cabinet in real time, detecting commodities taken by customers, determining the positions and moments of the commodities through a laser scanner and a motion detection sensor, and sending trigger signals to an image recognition unit when the positions and moments reach a preset area to trigger an image acquisition unit to capture commodity images;
the image recognition unit is internally provided with a plurality of high-resolution cameras, and when the object detection unit detects that the commodity moves, the cameras are activated and capture images of the commodity, and the acquired images are uploaded to the image recognition unit;
the image recognition unit is used for processing the image data captured by the image acquisition unit, analyzing the image by using an image processing and machine learning algorithm, extracting the characteristics of the commodity from the image data, and sending the extracted commodity characteristics to the data fusion processing unit;
the sensor array comprises a weight sensor, a temperature sensor and a pressure sensor, auxiliary data acquisition is carried out on the environmental temperature change data in the commodity moving and unmanned container by collecting non-visual properties of the commodity, and the acquired data are sent to the data fusion processing unit;
The data fusion processing unit is responsible for integrating data from the image recognition unit and the sensor array, integrating visual and non-visual data through a fusion algorithm to form integrated data and sending the integrated data into the recognition algorithm module;
the recognition algorithm module performs commodity recognition on the integrated data uploaded by the data fusion processing unit through the deep learning and pattern recognition algorithm, compares the integrated data with commodity information in the commodity interaction database to determine the identity of the commodity, and sends commodity data with the identity confirmed to the commodity interaction database for data update;
the commodity interaction database is a centralized storage library of commodity information which can be identified, visual characteristics, non-visual attributes, prices and inventory information of the commodities, and the commodity characteristics are identified and stored after the commodities newly identified by the identification algorithm module are received.
The current commodity identification system generally uses bar codes, two-dimensional codes or RFID and other technologies to realize commodity identification. These techniques require that the merchandise be marked prior to being put on shelf and that information and location of the merchandise be entered into the system. This approach has several drawbacks: the merchandise must be entered into the system prior to placement in the designated location of the unmanned cabinet. This means that if new goods are put on shelf or the goods location is changed, the system needs to be manually updated, which increases the management cost and the workload. The merchandise must be placed in a designated location to ensure that the system is able to properly identify the merchandise. This limits the manner in which the merchandise is displayed and the layout of the cabinet, affecting the flexibility and efficiency of the retail space. If the merchandise is not properly marked or entered into the system, or placed in the wrong location, the system cannot accurately identify the merchandise, resulting in a misidentification or misrecognition problem. This requires manual intervention and management as the goods need to be entered into the system in advance. This increases the complexity of the work and possible errors. For newly released commodities, the system is required to be firstly recorded and then placed at a designated position, so that a certain difficulty and delay are brought to the marketing of new commodities. The existing commodity identification system needs to be input into the system in advance and placed at a designated position, and has certain limitations and disadvantages, so that a more intelligent unmanned retail commodity identification system is needed to solve the problems. The novel unmanned retail commodity identification system is also used for improving the accuracy and instantaneity of commodity identification by adopting technologies such as real-time monitoring, multi-camera image identification, data fusion processing and deep learning algorithm, so that the problems existing in the conventional system are solved.
The unmanned retail commodity identification system in the document of the present application has the following differences from the prior art in that the object detection unit monitors the position and timing of the commodity in real time through the laser scanner and the motion detection sensor, and transmits a trigger signal to the image recognition unit when reaching a preset area. The real-time triggering mechanism can improve the response speed and accuracy of the system, and compared with the traditional timing acquisition, the real-time triggering mechanism can capture images of commodities more quickly. The image recognition unit is internally provided with a plurality of high-resolution cameras, so that a plurality of angles and details of the commodity can be captured, and the accuracy and stability of recognition of the commodity are improved. Compared with a single camera system, the multi-camera system has stronger information acquisition capability. The sensor array collects non-visual attribute data of the commodity, such as weight, temperature, pressure and the like, and the data fusion processing unit is responsible for integrating the data from the image recognition unit and the sensor array, and integrates visual and non-visual data through a fusion algorithm. The comprehensive processing can provide more comprehensive and accurate commodity information, and is helpful for improving the accuracy and reliability of commodity identification. The recognition algorithm module adopts a deep learning and pattern recognition algorithm to carry out commodity recognition on the integrated data, and compared with the traditional image processing algorithm, the deep learning algorithm has stronger learning and recognition capability, and can improve the accuracy and adaptability of commodity recognition. The commodity interaction database is a centralized storage library of commodity information for identification, and comprises visual features, non-visual attributes, prices, inventory information and the like of commodities, so that the commodity information can be comprehensively managed and updated. The centralized storage and management mode can improve the data management efficiency and accuracy of the system. The unmanned retail commodity identification system has obvious advantages and innovations in the aspects of real-time monitoring, multi-camera image identification, data fusion processing, deep learning algorithm, commodity interaction database and the like, can improve the identification accuracy, response speed and reliability of the system, and brings more advanced technical solutions to the unmanned retail field.
As a preferred implementation manner, in the data fusion module, when visual and non-visual data are combined, the data are subjected to weighted fusion processing, and the accuracy and the robustness of recognition are improved through the weighted fusion processing.
In a preferred embodiment, during the weighted fusion process, the collected data is first subjected to comprehensive weighted average by a weighted average method, and then subjected to real-time data processing by a kalman filter.
In a preferred embodiment, when the weighted average is performed by a weighted average method, the data of different sensors or sources are weighted-averaged, the accuracy of the data is low, the weighting coefficient is obtained according to the accuracy and the reliability of the sensor, and the data of different sources are weighted-averaged by the weighting coefficient.
As a preferred implementation mode, the recognition algorithm module is used for carrying out algorithm model establishment by training a deep learning model and adopting any one of ResNet, VGG, inception, recognizing objects in the image through the algorithm model, extracting features in the image, and carrying out matching correspondence on image data and commodities by utilizing feature matching.
After the technical scheme is adopted, the invention has the beneficial effects that: through the multi-camera image recognition and the deep learning algorithm, the system can more accurately recognize commodities, and the conditions of false recognition and missing recognition are reduced, so that the reliability and the accuracy of the whole system are improved. The system can monitor the position and the moment of the commodity in real time and trigger image acquisition when reaching a preset area, which is helpful for capturing the image of the commodity in time and improves the response speed and the instantaneity of the system. By integrating data from the image recognition unit and the sensor array, the system can provide more comprehensive and accurate merchandise information, which helps to improve the accuracy and reliability of merchandise recognition. Because the system can rapidly and accurately identify the commodity, the user can more conveniently and rapidly complete the shopping and payment of the commodity in the shopping process, and the shopping experience of the user is improved. The commodity interaction database can intensively store and manage commodity information, realizes comprehensive management and update of commodity information, and is beneficial to improving the data management efficiency and accuracy of the system. The unmanned retail commodity identification system can bring more efficient, convenient and reliable solutions to the unmanned retail industry by improving the identification accuracy, real-time monitoring and triggering mechanism, data fusion processing, user experience, data management and updating and other optimization, and is beneficial to improving the competitiveness and user satisfaction of the retail industry.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions of the prior art, the drawings which are used in the description of the embodiments or the prior art will be briefly described, it being obvious that the drawings in the description below are only some embodiments of the invention, and that other drawings can be obtained according to these drawings without inventive faculty for a person skilled in the art.
FIG. 1 is a block diagram of a system of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Examples:
As shown in fig. 1, the unmanned retail commodity identification system comprises an object detection unit, an image acquisition unit, an image identification unit, a sensor array, a data fusion processing unit, an identification algorithm module and a commodity interaction database for data interaction;
the object detection unit: monitoring an unmanned retail cabinet in real time, detecting commodities taken by customers, determining the positions and moments of the commodities through a laser scanner and a motion detection sensor, and sending trigger signals to an image recognition unit when the positions and moments reach a preset area to trigger an image acquisition unit to capture commodity images;
the image recognition unit is internally provided with a plurality of high-resolution cameras, and when the object detection unit detects that the commodity moves, the cameras are activated and capture images of the commodity, and the acquired images are uploaded to the image recognition unit;
the image recognition unit is used for processing the image data captured by the image acquisition unit, analyzing the image by using an image processing and machine learning algorithm, extracting the characteristics of the commodity from the image data, and sending the extracted commodity characteristics to the data fusion processing unit;
the sensor array comprises a weight sensor, a temperature sensor and a pressure sensor, auxiliary data acquisition is carried out on the environmental temperature change data in the commodity moving and unmanned container by collecting non-visual properties of the commodity, and the acquired data are sent to the data fusion processing unit;
The data fusion processing unit is responsible for integrating data from the image recognition unit and the sensor array, integrating visual and non-visual data through a fusion algorithm to form integrated data and sending the integrated data into the recognition algorithm module;
the recognition algorithm module performs commodity recognition on the integrated data uploaded by the data fusion processing unit through the deep learning and pattern recognition algorithm, compares the integrated data with commodity information in the commodity interaction database to determine the identity of the commodity, and sends commodity data with the identity confirmed to the commodity interaction database for data update;
the commodity interaction database is a centralized storage library of commodity information which can be identified, visual characteristics, non-visual attributes, prices and inventory information of the commodities, and the commodity characteristics are identified and stored after the commodities newly identified by the identification algorithm module are received.
The working principle and the working flow of the unmanned retail commodity identification system are as follows: the positions of the goods in the unmanned retail cabinets and the moment when the customer takes the goods are monitored in real time through the laser scanners and the motion detection sensors. Once the commodity moves to the preset area, the object detection unit triggers the image acquisition unit to capture the commodity image. When the object detection unit is triggered, the high-resolution camera built in the image recognition unit is activated and captures images of commodities, and then the acquired images are uploaded to the image recognition unit. The image recognition unit analyzes the image using image processing and machine learning algorithms, extracts features of the commodity, and transmits the extracted commodity features to the data fusion processing unit. The weight sensor, the temperature sensor and the pressure sensor collect non-visual properties of the commodity and environmental temperature change data in the unmanned container, and then the collected data are sent to the data fusion processing unit. And integrating the visual and non-visual data through a fusion algorithm to form integrated data and sending the integrated data into the recognition algorithm module. And carrying out commodity identification on the integrated data uploaded by the data fusion processing unit through deep learning and a pattern recognition algorithm, comparing the integrated data with commodity information in a commodity interaction database to determine the identity of the commodity, and sending commodity data with the identity confirmed to the commodity interaction database for data updating. And the commodity information which can be identified is stored in a centralized way, wherein the commodity information comprises visual characteristics, non-visual attributes, prices, inventory information and the like of the commodity, and meanwhile, the commodity characteristics which are newly identified by the identification algorithm module are received and stored.
The system of the document monitors the commodity position in real time through the laser scanner and the motion detection sensor, and triggers image acquisition when reaching a preset area, so that the response speed and the instantaneity of the system are improved, and the defect that commodity information needs to be input in advance in the traditional system is overcome. The data of the image recognition unit and the sensor array are integrated, and visual and non-visual data are integrated through a fusion algorithm, so that more comprehensive and accurate commodity information is provided, and the accuracy and reliability of commodity recognition are improved. The commodity identification is carried out by adopting the deep learning and pattern identification algorithm, so that the method has stronger learning and identification capability, and the accuracy and adaptability of commodity identification are improved. The unmanned retail commodity identification system solves the defect that commodity information and positions need to be input in advance in the traditional system through innovative application of technologies such as real-time monitoring, data fusion processing and deep learning algorithm, improves the accuracy and instantaneity of commodity identification, and has higher intelligent and automatic levels.
When combining visual and non-visual data, the data fusion module performs weighted fusion processing on the data, and the accuracy and the robustness of recognition are improved through the weighted fusion processing. The accuracy and the robustness of commodity identification are improved by carrying out weighted fusion processing on visual and non-visual data. The weighted fusion process herein refers to assigning different weights to different types of data and combining them to achieve more reliable results. In this case, the visual data may include image features of the merchandise, while the non-visual data may include environmental and merchandise attribute data collected by sensors such as weight, temperature, and pressure. Through the weighted fusion processing, the system can carry out weighted processing on different data according to the importance and the reliability of different data types, so as to ensure that the actual condition of the commodity is reflected more accurately when various data are comprehensively considered. The image features of the merchandise may be more intuitive and important and thus may be given a higher weight; the data such as the ambient temperature and the weight may be used as auxiliary information, and an appropriate weight may be given to the commodity according to the degree of contribution to commodity identification. Through the weighted fusion processing, the system can more comprehensively consider the contributions of different data, thereby improving the accuracy and the robustness of commodity identification. By carrying out weighted fusion processing on different types of data, the system can better comprehensively utilize various information, and accuracy and reliability of commodity identification are improved, so that performance and efficiency of the whole unmanned retail commodity identification system are improved.
When the weighted fusion processing is carried out, firstly, the weighted average method is adopted to carry out comprehensive weighted average on the acquired data, and then the real-time data processing is carried out on the data through the Kalman filter. And adopting a weighted average method to comprehensively weighted average the collected data of different types. This means that the system assigns a weight to each data type, then multiplies each data by its corresponding weight, and then adds them to obtain a comprehensive weighted average. This method can ensure that different types of data are properly valued in the synthesis, and weighting is performed according to the importance and contribution degree.
And carrying out real-time data processing on the data through a Kalman filter. The kalman filter is a mathematical tool for extracting state information from incomplete and noisy data. The method can eliminate noise in the data and improve the accuracy of the data by dynamically estimating and predicting the data. In this case, the kalman filter can help the system process the data in real time, and improve the stability and robustness of the data, so as to better support the accuracy of the commodity identification system. The weighted average method is adopted to carry out comprehensive weighted average, so that different types of data can be ensured to be properly valued, and the stability and accuracy of the data can be improved by carrying out real-time data processing through a Kalman filter. The method can better comprehensively utilize various information in the weighted fusion processing process, thereby improving the accuracy and the robustness of the commodity identification system.
When the weighted average is carried out by the weighted average method, the data of different sensors or sources are weighted-averaged, the noise is low, the accuracy of the data is improved, the weighting coefficient is obtained according to the accuracy and the reliability of the sensors, and the data of different sources are weighted-averaged by the weighting coefficient. Based on the accuracy and reliability of the sensors, the system assigns a weighting factor to each sensor or data source. Sensors with high accuracy and reliability will be given higher weights, while sensors with low accuracy and reliability will be given lower weights. This is based on the assumption that the data generated by a highly accurate and reliable sensor is closer to the true value and should therefore play a greater role in the weighted average.
The data from different sources are weighted averaged by these weighting coefficients. Data with higher weights will occupy more weight in the weighted average, while data with lower weights will have less impact on the results. This ensures that data from different sensors or sources is properly valued when combined, thereby improving the accuracy and reliability of the data.
According to the method, the weighted average is carried out on the data from different sources, the weighted coefficient is obtained according to the accuracy and the reliability of the sensor, the noise can be reduced, the accuracy of the data is improved, and therefore the system is ensured to obtain more accurate and reliable results. This strategy is very important in data fusion processing, and can effectively improve the performance of the system and the accuracy of the results, especially in the case where multiple sensors or data sources are required to be comprehensively utilized.
The recognition algorithm module is used for carrying out algorithm model establishment by training a deep learning model and adopting any one of ResNet, VGG, inception, recognizing objects in the image through the algorithm model, extracting features in the image, and matching and corresponding image data with commodities by utilizing feature matching. The deep learning model (such as ResNet, VGG, inception) is a machine learning model based on an artificial neural network, and features and modes in an image can be learned by training a large amount of image data. In this case, by training the deep learning model, the system can learn the visual features and patterns of various commodities, so that the system can more accurately recognize objects in the image. Once the deep learning model is trained, the system can use the model to identify objects in the image. By extracting features from the image, the system can compare and match these features with pre-trained merchandise features to determine the identity of the object in the image. The characteristic matching process can help the system to accurately match the image data with the corresponding commodity, thereby realizing the identification and management of the commodity. By training a deep learning model and adopting models such as ResNet, VGG, inception for algorithm model establishment, the system can learn visual features and modes of commodities, so that objects in images can be more accurately identified, and image data and the commodities are correspondingly matched through feature matching. The method can improve the accuracy and the robustness of the system for identifying the commodities, and provides stronger identification capability for an unmanned retail commodity identification system.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, alternatives, and improvements that fall within the spirit and scope of the invention.
Claims (5)
1. The unmanned retail commodity identification system is characterized by comprising an object detection unit, an image acquisition unit, an image identification unit, a sensor array, a data fusion processing unit, an identification algorithm module and a commodity interaction database for data interaction;
the object detection unit: monitoring an unmanned retail cabinet in real time, detecting commodities taken by customers, determining the positions and moments of the commodities through a laser scanner and a motion detection sensor, and sending trigger signals to an image recognition unit when the positions and moments reach a preset area to trigger an image acquisition unit to capture commodity images;
the image recognition unit is internally provided with a plurality of high-resolution cameras, and when the object detection unit detects that the commodity moves, the cameras are activated and capture images of the commodity, and the acquired images are uploaded to the image recognition unit;
the image recognition unit is used for processing the image data captured by the image acquisition unit, analyzing the image by using an image processing and machine learning algorithm, extracting the characteristics of the commodity from the image data, and sending the extracted commodity characteristics to the data fusion processing unit;
the sensor array comprises a weight sensor, a temperature sensor and a pressure sensor, auxiliary data acquisition is carried out on the environmental temperature change data in the commodity moving and unmanned container by collecting non-visual properties of the commodity, and the acquired data are sent to the data fusion processing unit;
The data fusion processing unit is responsible for integrating data from the image recognition unit and the sensor array, integrating visual and non-visual data through a fusion algorithm to form integrated data and sending the integrated data into the recognition algorithm module;
the recognition algorithm module performs commodity recognition on the integrated data uploaded by the data fusion processing unit through the deep learning and pattern recognition algorithm, compares the integrated data with commodity information in the commodity interaction database to determine the identity of the commodity, and sends commodity data with the identity confirmed to the commodity interaction database for data update;
the commodity interaction database is a centralized storage library of commodity information which can be identified, visual characteristics, non-visual attributes, prices and inventory information of the commodities, and the commodity characteristics are identified and stored after the commodities newly identified by the identification algorithm module are received.
2. An unmanned retail item identification system as recited in claim 1, wherein: when combining visual and non-visual data, the data fusion module performs weighted fusion processing on the data, and the accuracy and the robustness of recognition are improved through the weighted fusion processing.
3. An unmanned retail item identification system as recited in claim 2, wherein: when the weighted fusion processing is carried out, firstly, the weighted average method is adopted to carry out comprehensive weighted average on the acquired data, and then the real-time data processing is carried out on the data through the Kalman filter.
4. An unmanned retail item identification system as recited in claim 3, wherein: when the weighted average is carried out by the weighted average method, the data of different sensors or sources are weighted-averaged, the noise is low, the accuracy of the data is improved, the weighting coefficient is obtained according to the accuracy and the reliability of the sensors, and the data of different sources are weighted-averaged by the weighting coefficient.
5. An unmanned retail item identification system as recited in claim 3, wherein: the recognition algorithm module is used for carrying out algorithm model establishment by training a deep learning model and adopting any one of ResNet, VGG, inception, recognizing objects in the image through the algorithm model, extracting features in the image, and matching and corresponding image data with commodities by utilizing feature matching.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410235621.1A CN118135483A (en) | 2024-03-01 | 2024-03-01 | Unmanned retail commodity identification system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410235621.1A CN118135483A (en) | 2024-03-01 | 2024-03-01 | Unmanned retail commodity identification system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN118135483A true CN118135483A (en) | 2024-06-04 |
Family
ID=91245416
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410235621.1A Pending CN118135483A (en) | 2024-03-01 | 2024-03-01 | Unmanned retail commodity identification system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118135483A (en) |
-
2024
- 2024-03-01 CN CN202410235621.1A patent/CN118135483A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111415461B (en) | Article identification method and system and electronic equipment | |
| US20210158053A1 (en) | Constructing shopper carts using video surveillance | |
| CN101965576B (en) | Object matching for tracking, indexing, and search | |
| CN111626201A (en) | Commodity detection method and device and readable storage medium | |
| CN111061890A (en) | Method for verifying labeling information, method and device for determining category | |
| CA3014365C (en) | System and method for gathering data related to quality of service in a customer service environment | |
| CN109712324A (en) | A kind of automatic vending machine image-recognizing method, good selling method and vending equipment | |
| KR20100041177A (en) | The automatic physical distribution system and control method | |
| Liu et al. | Grab: Fast and accurate sensor processing for cashier-free shopping | |
| CN110689389A (en) | Computer vision-based shopping list automatic maintenance method and device, storage medium and terminal | |
| CN113468914A (en) | Method, device and equipment for determining purity of commodities | |
| EP3629276A1 (en) | Context-aided machine vision item differentiation | |
| US11825240B2 (en) | Video based monitoring system using multiple cameras, method, computer program and storage medium | |
| CN114255377A (en) | Differential commodity detection and classification method for intelligent container | |
| WO2019077559A1 (en) | System for tracking products and users in a store | |
| JP2023504871A (en) | Fraud detection system and method | |
| CN118135483A (en) | Unmanned retail commodity identification system | |
| CN110096946A (en) | A kind of self-service system merged based on pressure sensitivity and vision | |
| CN110866746A (en) | Order matching method based on face recognition and computer equipment | |
| KR102493331B1 (en) | Method and System for Predicting Customer Tracking and Shopping Time in Stores | |
| CN115294325A (en) | Dynamic commodity identification system, method, medium, equipment and terminal for sales counter | |
| CN110443946A (en) | Vending machine, the recognition methods of type of goods and device | |
| CN111461104B (en) | Visual recognition method, device, equipment and storage medium | |
| CN114821729B (en) | Commodity shopping guide method and device, cloud server and storage medium | |
| CN117253194B (en) | Commodity damage detection method, commodity damage detection device and storage medium |
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 |