CN116403197B - Intelligent weighing method and system based on AI image recognition - Google Patents

Intelligent weighing method and system based on AI image recognition Download PDF

Info

Publication number
CN116403197B
CN116403197B CN202310313233.6A CN202310313233A CN116403197B CN 116403197 B CN116403197 B CN 116403197B CN 202310313233 A CN202310313233 A CN 202310313233A CN 116403197 B CN116403197 B CN 116403197B
Authority
CN
China
Prior art keywords
weighed
image
information
article
intelligent
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.)
Active
Application number
CN202310313233.6A
Other languages
Chinese (zh)
Other versions
CN116403197A (en
Inventor
石再亮
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guangzhou Zonescale Weighing Apparatus Co ltd
Original Assignee
Guangzhou Zonescale Weighing Apparatus Co ltd
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 Guangzhou Zonescale Weighing Apparatus Co ltd filed Critical Guangzhou Zonescale Weighing Apparatus Co ltd
Priority to CN202310313233.6A priority Critical patent/CN116403197B/en
Publication of CN116403197A publication Critical patent/CN116403197A/en
Application granted granted Critical
Publication of CN116403197B publication Critical patent/CN116403197B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/60—Type of objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/51—Indexing; Data structures therefor; Storage structures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/53—Querying
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/54—Browsing; Visualisation therefor
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/10—Image acquisition
    • G06V10/12—Details of acquisition arrangements; Constructional details thereof
    • G06V10/14—Optical characteristics of the device performing the acquisition or on the illumination arrangements
    • G06V10/141—Control of illumination
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • 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
    • G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
    • G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Abstract

The invention provides an intelligent weighing method and system based on AI image recognition, wherein the method comprises the following steps: s1, placing an object to be weighed by a weighing platform module, acquiring weight information of the object to be weighed, and transmitting the acquired weight information to an intelligent analysis module; s2, an image acquisition module acquires an image of an object to be weighed placed on the weighing platform, and transmits the acquired image of the object to be weighed to an intelligent analysis module; s3, the intelligent analysis module performs article identification analysis according to the acquired image of the article to be weighed, obtains identification information of the article to be weighed, and acquires associated article information according to the identification information of the article to be weighed; obtaining corresponding pricing information according to the obtained article information and weight information; and S4, the interaction module displays the article information, the weight information and the pricing information of the article to be weighed, and generates corresponding payment information according to the pricing information for display. The intelligent weighing system is beneficial to improving the intelligent level of weighing the articles and reducing the labor cost in the weighing process of the articles.

Description

Intelligent weighing method and system based on AI image recognition
Technical Field
The invention relates to the technical field of electronic scales, in particular to an intelligent weighing method and system based on AI image recognition.
Background
Currently, in each supermarket or shopping mall, an electronic scale is generally provided to measure the weight of an article sold in bulk, such as fruit, candy, etc., which is counted by weight, and to obtain the price of the article purchased by the user.
In the process of weighing and settling the articles by adopting the existing electronic scale, the staff is usually required to memorize the numbers of the articles, and then the electronic scale can settle the articles after inputting the corresponding article numbers manually; however, in a manner of operating and weighing by the staff, on one hand, the corresponding staff is required to be specially arranged for each electronic scale to operate, and a great deal of manpower is required to be consumed; on the other hand, in the process of operation, the situation that misoperation or deviation occurs due to the influence of human factors is easy to occur. The requirements of modern intelligent shopping mall construction cannot be met.
Disclosure of Invention
In view of the above problems, the present invention aims to provide an intelligent weighing method based on AI image recognition and a system thereof.
The aim of the invention is realized by adopting the following technical scheme:
in a first aspect, the present invention shows an intelligent weighing method based on AI image recognition, comprising:
S1, placing an object to be weighed by a weighing platform module, acquiring weight information of the object to be weighed, and transmitting the acquired weight information to an intelligent analysis module;
s2, an image acquisition module acquires an image of an object to be weighed placed on the weighing platform, and transmits the acquired image of the object to be weighed to an intelligent analysis module;
s3, the intelligent analysis module performs article identification analysis according to the acquired image of the article to be weighed, obtains identification information of the article to be weighed, and acquires associated article information according to the identification information of the article to be weighed; obtaining corresponding pricing information according to the obtained article information and weight information;
and S4, the interaction module displays the article information, the weight information and the pricing information of the article to be weighed, and generates corresponding payment information according to the pricing information for display.
In one embodiment, step S1 includes:
s11, placing an object to be weighed to a weighing platform module;
s12, the weighing platform module acquires weight information of the object to be weighed placed on the object placing unit and transmits the acquired weight information to the intelligent analysis module;
and S13, when the object placing unit is detected to prevent the object to be weighed, the weighing platform module sends a driving signal to the image acquisition module.
In one embodiment, step S2 includes: the image acquisition module acquires an image of the object to be weighed placed on the weighing platform according to the received driving signal, and transmits the acquired image of the object to be weighed to the intelligent analysis module.
In one embodiment, step S3 includes:
s31, the intelligent analysis module performs pretreatment on the acquired image of the object to be weighed to obtain a pretreated image of the object to be weighed;
s32, the intelligent analysis module identifies the object to be weighed based on the trained intelligent object identification model according to the preprocessed object image to be weighed, and identification information of the object to be weighed is obtained;
s33, the intelligent analysis module matches article information associated with the article from a database according to the obtained identification information of the article to be weighed, wherein the article information comprises an article name, a type, an article ID, a production place and a unit price;
s34, the intelligent analysis module calculates pricing information of the to-be-weighed object according to the weight information and the object information of the to-be-weighed object.
In one embodiment, a trained intelligent object identification model is built based on a CNN neural convolution network and comprises an input layer, a first convolution layer, a second convolution layer, a pooling layer, a first full-connection layer, a second full-connection layer, an activation layer and an output layer which are sequentially arranged;
The method comprises the steps of inputting a preprocessed image of an object to be weighed into an input layer, wherein a first convolution layer and a second convolution layer respectively comprise 16 convolution kernels and 32 convolution kernels, and the sizes of the convolution kernels are 3 multiplied by 3; the pooling layer is set as the largest pooling layer, and the size of the pooling core is 3 multiplied by 3; the first full-connection layer is provided with 64 neurons, the second full-connection layer is provided with 16 neurons, the activation layer adopts a softmax function, the characteristic vector of the reaction object is output through the full-connection layer, and the identification probability of the object to be weighed is obtained through the softmax function according to the characteristic vector; and the output layer outputs the identification result of the object to be weighed according to the obtained identification probability.
In one embodiment, step S3 further includes:
and S30, training the intelligent article identification model by the intelligent analysis module according to the constructed training set and the test set to obtain a trained intelligent article identification model.
In one embodiment, step S4 includes:
s41, the interaction module displays weight information, article information and pricing information of the articles to be weighed;
s42, the interaction module counts price information of all the objects to be weighed of the user, generates payment information according to the obtained payment total price, and further displays the payment information by the display unit;
S43, the interaction module prints a receipt or label of the paid article after the user finishes paying.
In one embodiment, the method further comprises the step S5 of managing the article information and the intelligent article identification model by the database module;
the step S5 specifically includes:
s51, the database module stores article information of articles and a trained intelligent article identification model;
s52, inputting the article information into a database module and managing the article information in the database.
In a second aspect, the present invention shows an intelligent weighing system based on AI image recognition, comprising: the system comprises a weighing platform module, an image acquisition module, an intelligent analysis module and an interaction module; wherein,
the weighing platform module is used for placing an object to be weighed, acquiring weight information of the object to be weighed and transmitting the acquired weight information to the intelligent analysis module;
the image acquisition module is used for acquiring an image of the object to be weighed placed on the weighing platform and transmitting the acquired image of the object to be weighed to the intelligent analysis module;
the intelligent analysis module is used for carrying out article identification analysis according to the acquired image of the article to be weighed, obtaining identification information of the article to be weighed, and acquiring associated article information according to the identification information of the article to be weighed; obtaining corresponding pricing information according to the obtained article information and weight information;
The interaction module is used for displaying article information, weight information and pricing information of the articles to be weighed, and generating corresponding payment information according to the pricing information for displaying.
In one embodiment, the weighing platform module comprises a storage unit, an electronic scale unit and a driving unit;
the object placing unit is used for placing objects to be weighed;
the electronic scale unit is used for acquiring weight information of the object to be weighed placed on the object placing unit and transmitting the acquired weight information to the intelligent analysis module;
the driving unit is used for sending a driving signal to the image acquisition module when detecting that the object to be weighed is prevented from being placed on the object placing unit.
In one embodiment, the image acquisition module comprises a camera unit;
the camera unit is used for collecting an image of the object to be weighed placed on the weighing platform according to the received driving signal and transmitting the collected image of the object to be weighed to the intelligent analysis module.
In one embodiment, the image acquisition module further comprises an illumination unit;
the illumination unit is used for providing a light source for a shooting area of the camera unit.
In one embodiment, the intelligent analysis module comprises an image preprocessing unit, a pre-detection unit, an intelligent analysis unit, a matching unit and a calculation unit; wherein,
The image preprocessing unit is used for preprocessing the acquired image of the object to be weighed to obtain a preprocessed image of the object to be weighed;
the intelligent analysis unit is used for identifying the object to be weighed based on the trained intelligent object identification model according to the preprocessed object image to be weighed, and obtaining identification information of the object to be weighed;
the matching unit is used for matching the article information associated with the article from the database according to the obtained identification information of the article to be weighed, wherein the article information comprises the name, the type, the ID, the place of production, the unit price and the like of the article;
the calculating unit is used for calculating pricing information of the object to be weighed according to the weight information and the object information of the object to be weighed.
In one embodiment, the intelligent analysis module further comprises a model training unit;
the model training unit is used for training the intelligent article identification model according to the constructed training set and the test set to obtain a trained intelligent article identification model.
In one embodiment, the interaction module comprises a display unit, a payment unit and a printing unit; wherein,
the display unit is used for displaying weight information, article information and pricing information of the articles to be weighed;
The payment unit is used for counting the price information of all the objects to be weighed of the user, generating payment information according to the obtained total payment price, and further displaying the payment information by the display unit;
the printing unit is used for printing the receipt or the label of the paid article after the user finishes paying.
In one embodiment, the system further comprises a database module for managing the item information and the intelligent item identification model; the database module comprises a database unit and a data management unit;
the database unit is used for storing article information of articles and a trained intelligent article identification model;
the data management unit is used for inputting the article information and managing the article information in the database.
The beneficial effects of the invention are as follows: image information of the article is acquired when the article is weighed, AI image identification is carried out according to the acquired image, an identification result of the article is obtained, pricing information of the article is intelligently calculated, the intelligent level of weighing the article is improved, labor cost in the weighing process of the article is reduced, and accuracy and reliability of intelligent weighing of the article are improved.
Drawings
The invention will be further described with reference to the accompanying drawings, in which embodiments do not constitute any limitation of the invention, and other drawings can be obtained by one of ordinary skill in the art without inventive effort from the following drawings.
FIG. 1 is a schematic diagram of an intelligent weighing method based on AI image recognition according to an embodiment of the invention;
fig. 2 is a schematic diagram of an intelligent weighing system based on AI image recognition according to an embodiment of the present invention.
Detailed Description
The invention is further described in connection with the following application scenario.
Referring to the embodiment of fig. 1, an intelligent weighing method based on AI image recognition is shown, comprising:
s1, placing an object to be weighed by a weighing platform module, acquiring weight information of the object to be weighed, and transmitting the acquired weight information to an intelligent analysis module;
in one embodiment, step S1 includes:
s11, placing an object to be weighed to a weighing platform module;
s12, the weighing platform module acquires weight information of the object to be weighed placed on the object placing unit and transmits the acquired weight information to the intelligent analysis module;
and S13, when the object placing unit is detected to prevent the object to be weighed, the weighing platform module sends a driving signal to the image acquisition module.
In the above embodiment, the weighing platform module can be set based on the intelligent electronic scale, when the weighing is needed, firstly, the object to be weighed is placed on the weighing platform, the electronic scale measures the weight of the object to be weighed on the bearing platform, so as to obtain the weight information of the object to be weighed, and the weighing platform module can complete the basic weight measurement function of the object to be weighed; meanwhile, through linkage with the weighing platform, when the weighing platform detects that an object is placed on the weighing platform, a driving signal is sent to the image acquisition module, so that the image acquisition module can start image acquisition of the object to be weighed on the weighing platform, and further intelligent object identification processing is performed based on the obtained image.
In one scene, a pressure sensor is arranged on the weighing platform, and when the value of the pressure sensor is changed from 0, a driving signal is sent to the image acquisition module.
S2, an image acquisition module acquires an image of an object to be weighed placed on the weighing platform, and transmits the acquired image of the object to be weighed to an intelligent analysis module;
in one embodiment, step S2 includes: the image acquisition module acquires an image of the object to be weighed placed on the weighing platform according to the received driving signal, and transmits the acquired image of the object to be weighed to the intelligent analysis module.
According to the embodiment of the invention, the image acquisition module can be arranged based on the high-definition camera, and after receiving the driving signal, the image acquisition module automatically captures the image information of the object to be weighed on the bearing platform and transmits the acquired image information to the intelligent analysis module for further object identification processing.
In a scene, a camera of the image acquisition module is arranged above the weighing platform and is 50cm away from the weighing platform, and the camera is inclined downwards by 40% to shoot.
S3, the intelligent analysis module performs article identification analysis according to the acquired image of the article to be weighed, obtains identification information of the article to be weighed, and acquires associated article information according to the identification information of the article to be weighed; and obtaining corresponding pricing information according to the obtained article information and weight information.
In one embodiment, step S3 includes:
s31, the intelligent analysis module performs pretreatment on the acquired image of the object to be weighed to obtain a pretreated image of the object to be weighed;
s32, the intelligent analysis module identifies the object to be weighed based on the trained intelligent object identification model according to the preprocessed object image to be weighed, and identification information of the object to be weighed is obtained;
s33, the intelligent analysis module matches article information associated with the article from a database according to the obtained identification information of the article to be weighed, wherein the article information comprises an article name, a type, an article ID, a production place and a unit price;
s34, the intelligent analysis module calculates pricing information of the to-be-weighed object according to the weight information and the object information of the to-be-weighed object.
According to the embodiment of the invention, the intelligent analysis module can be built based on a local intelligent terminal or a cloud platform arranged in a shopping mall, wherein the intelligent analysis module is respectively connected with the weighing platform module and the image acquisition module through wireless communication, and receives weight information and an image of an object to be weighed, which are transmitted by the weighing platform module and the image acquisition module, in real time. The method comprises the steps of firstly preprocessing a received image of an object to be weighed, such as size adjustment or image enhancement, obtaining a preprocessed image, inputting the preprocessed image into an intelligent object recognition model, and recognizing the input image based on an AI image recognition technology, so that a recognition result of the object to be weighed is obtained. According to the identification result of the object to be weighed, the object information of the object is further matched from the database, the pricing information of the object is calculated by combining the obtained weight information, and the data such as the object information, the weight information and the pricing information are transmitted to the interaction module for further display.
Based on the acquired image of the object to be weighed, the object to be weighed is intelligently identified and corresponding object information is extracted through an AI image identification technology, so that the intelligent identification and automatic pricing functions of the object to be weighed can be realized, the workload of staff in the object weighing process is greatly reduced, and support is provided for unmanned weighing and pricing construction of supermarket commodities. Meanwhile, the article is identified based on the AI image identification technology, so that the accuracy of article identification can be ensured, and the reliability of intelligent weighing is improved.
The weighing platform is easy to receive environmental influence of the weighing platform and conditions such as surface reflection (such as reflection of a transparent shopping bag or reflection of material on the surface of an article) of the article in the process of acquiring the image of the article to be weighed through the image acquisition module, so that the image of the article to be weighed is unclear, and the problem of influence on the accuracy of article identification according to the image of the article to be weighed is solved. In the embodiment of the invention, a technical scheme of preprocessing the acquired image to be weighed firstly through the intelligent analysis module is also provided, so that the definition of the image is improved.
In one embodiment, in step S3, the intelligent analysis module performs preprocessing on the acquired image of the object to be weighed, and specifically includes:
According to the obtained object image X to be weighed, carrying out edge detection processing on the object image X to be weighed to obtain edge characteristic information of the object image X to be weighed, and according to comparison of the obtained edge characteristic information and the edge characteristic information of the standard weighing table image B, extracting a foreground area sigma of the object image to be weighed a And a background area sigma b The method comprises the steps of carrying out a first treatment on the surface of the The standard weighing platform image B is obtained by shooting the image acquisition module at the same position/angle symmetrical platform module;
wavelet decomposition is carried out on the image X of the object to be weighed to obtain a low-frequency component image f of the image X of the object to be weighed d And a high-frequency component image f h Wherein the decomposition scale adopted by wavelet decomposition is 1 layer, and the wavelet base is db2;
from the obtained low-frequency component image f d Further the low frequency component image f d Conversion from RGB color space to LAB color space to obtain low frequency component image f d A luminance component eL, a color component eA, and a color component eB;
local brightness adjustment is carried out on the foreground area according to the obtained brightness component eL to obtain an adjusted brightness componentThe local brightness adjustment function used is:
in the method, in the process of the invention,representing luminance values of pixel points (x, y) after local luminance adjustment, wherein the pixelsPoint (x, y) e sigma a ，Represented in background area sigma b The nearest pixel point to the pixel point (x, y); />Representing pixel dot +.>Luminance value of>Representing background area sigma b Average luminance value of L G Represents a set standard luminance value, wherein L G ∈[60,65]；Representing foreground region sigma a L (x, y) represents the luminance value of the pixel point (x, y),represents a judgment function in which when->And L G When the positive and negative signs of L (x, y) are identical (e.g.)>And L is G -L (x, y) > 0, or +.>And L is G -L (x, y) < 0), thenOtherwise->ω 1 Local adjustment of presentation settingsFactor, ω 1 ∈[0.3,0.4]，ω 2 Representing a foreground modulator, wherein omega 2 ∈[0.2,0.4]，ω 3 Representing an abnormality-compensating regulatory factor, wherein ω 3 ∈[0.2,0.3]，ω 1 ≥ω 2 ≥ω 3 ；
Based on the adjusted luminance componentReconstructing the color component eA and the color component eB to obtain an adjusted low-frequency component image +.>
From the obtained high-frequency component image f h For the high frequency component image f h Performing detail adjustment processing to obtain a high-frequency component image after the detail adjustment processingThe detail adjusting function adopted is as follows:
in the method, in the process of the invention,representing gray values of pixel points (x, y) after detail adjustment processing, and H (x, y) represents the high-frequency component image f h Gray values of the middle pixel points (x, y); (i, j) represents a pixel point in a 3×3 neighborhood region θ (x, y) centered on the pixel point (x, y), max (|h (x, y) -H (i, j) |) represents a maximum value of differences between gray values of the pixel point (x, y) and respective pixel points in its neighborhood region, and min (|h (x, y) -H (i, j) |) represents a minimum value of differences between gray values of the pixel point (x, y) and respective pixel points in its neighborhood region; t1 represents a set first gray threshold, T1 ε [150,200 ] ]；H θ(x,y) Represents the average gray value of a 3 x 3 neighborhood region centered on pixel (x, y), T2 represents a set second gray threshold, where T2 ε [50,70 ]]The method comprises the steps of carrying out a first treatment on the surface of the N tableShows the number of pixels, H, in the neighborhood region θ (x, y) k Representing the set compensation gray value, where H k ∈[45,55]；ω 4 Represents a set adjustment factor, wherein ω 4 ∈[0.8,0.9]，ω 5 Representing a set compensation factor, wherein ω 5 ∈[0.2,0.3]；
From the adjusted low-frequency component imageAnd a high-frequency component image after detail adjustment processing +.>And reconstructing to obtain the enhanced image of the object to be weighed.
In one embodiment, according to the obtained enhanced image of the object to be weighed, the background area is further removed, and only the foreground area is reserved, so that the preprocessed image of the object to be weighed is obtained.
The embodiment of the invention provides a technical scheme for carrying out self-adaptive pretreatment on an acquired image of an object to be weighed, wherein the acquired image of the object to be weighed is subjected to wavelet decomposition to respectively obtain a high-frequency component image and a low-frequency component image of the image; for the obtained low-frequency component image, a brightness adjustment processing function based on an LAB color space is provided for adaptively adjusting the brightness component value of the pixel point, wherein in the brightness adjustment function, the possible light emission interference of a packaging bag or an object existing in the image is considered, so that the brightness characteristic of each position in the foreground area is reflected by particularly taking the average brightness of the foreground area as a basis and combining the brightness of the surrounding area of the foreground area as an adjustment parameter, the condition of reflection interference is effectively restrained, meanwhile, the brightness compensation is carried out on the foreground area part according to the brightness characteristic of the whole image, and the brightness of the foreground area is adjusted to a proper level, so that the definition of an object part in the image is improved. For the obtained high-frequency component image, carrying out self-adaptive adjustment processing on detail information in the image, wherein the adjustment is carried out for possible noise interference, so that the suppression of the interference information in the image is facilitated, and the representation level of detail characteristics is improved; and reconstructing according to the high-frequency component image and the low-frequency component image after the adjustment processing to obtain an enhanced object image to be weighed, extracting according to a foreground part in the image to obtain an object image, and further inputting the object image into an intelligent object recognition model to perform object recognition, so that the accuracy and the adaptability of object recognition are improved, and meanwhile, the reliability of intelligent weighing is improved.
In one embodiment, a trained intelligent object identification model is built based on a CNN neural convolution network and comprises an input layer, a first convolution layer, a second convolution layer, a pooling layer, a first full-connection layer, a second full-connection layer, an activation layer and an output layer which are sequentially arranged;
the method comprises the steps of inputting a preprocessed image of an object to be weighed into an input layer, wherein a first convolution layer and a second convolution layer respectively comprise 16 convolution kernels and 32 convolution kernels, and the sizes of the convolution kernels are 3 multiplied by 3; the pooling layer is set as the largest pooling layer, and the size of the pooling core is 3 multiplied by 3; the first full-connection layer is provided with 64 neurons, the second full-connection layer is provided with 16 neurons, the activation layer adopts a softmax function, the characteristic vector of the reaction object is output through the full-connection layer, and the identification probability of the object to be weighed is obtained through the softmax function according to the characteristic vector; and the output layer outputs the identification result of the object to be weighed according to the obtained identification probability.
The intelligent article identification model is constructed based on a CNN neural convolution network, automatic feature extraction is carried out on articles in the image according to the input image, intelligent article identification is carried out according to the obtained feature vector, and corresponding article identification results are obtained, wherein the softmax layer calculates the probability that the articles are different identification results according to the obtained feature vector, and the output layer is set to be corresponding article names according to the obtained probability when the probability that the articles belong to a certain identification result exceeds a set standard threshold value; when the softmax layer calculates that the probability difference of the two identification results is smaller than a set threshold value (if the articles to be identified are confused with different types of articles), the output identification result is an abnormal prompt message; when the softmax layer calculates that the probability of each recognition result is smaller than the standard threshold, the recognition result is output as unrecognizable. The intelligent article identification model constructed based on the CNN neural convolution network can ensure the accuracy of article identification, and simultaneously control the identification time within a proper range, thereby being beneficial to improving the reliability and adaptability of article identification.
In one embodiment, step S3 further includes:
and S30, training the intelligent article identification model by the intelligent analysis module according to the constructed training set and the test set to obtain a trained intelligent article identification model.
When training the model, firstly constructing a training set, wherein the training set comprises images of different article types and corresponding article marks; the images of the training set select the physical images of the articles (such as single fruit images, single candy images and the like) or the images of the articles with transparent packages (such as fruit images of apples, bananas, pineapples and the like which are packaged in transparent packaging bags or rice images which are packaged in transparent packaging bags) according to actual conditions, so that the model accuracy is improved due to the fact that the images of the training set are suitable for different presentation forms of weighing articles in supermarket scenes. Meanwhile, based on the constructed training set, 20% of the training set is used as a test set, the trained model is tested, and when the test rate exceeds a standard threshold (for example, 97%), the training of the intelligent object identification model is finished.
In one embodiment, after the training set is constructed, the images in the training set are also first preprocessed in step S3, so as to obtain a preprocessed image (such as a picture image with only the foreground area reserved) as the training set.
And S4, the interaction module displays the article information, the weight information and the pricing information of the article to be weighed, and generates corresponding payment information according to the pricing information for display.
In one embodiment, step S4 includes:
s41, the interaction module displays weight information, article information and pricing information of the articles to be weighed;
s42, the interaction module counts price information of all the objects to be weighed of the user, generates payment information according to the obtained payment total price, and further displays the payment information by the display unit;
s43, the interaction module prints a receipt or label of the paid article after the user finishes paying.
The interaction module can be set based on an intelligent display screen (such as a settlement terminal) or an electronic scale terminal arranged in a supermarket or a shopping place, and after weighing and intelligent identification of the object to be weighed are completed, the weight information, the object information and the pricing information of the object are displayed for a customer/staff to confirm the purchased/sold object. Meanwhile, according to actual design requirements, payment information (such as a two-dimensional code) of the article is generated according to pricing information of the article, and a receipt is printed after payment; or directly printing labels and the like according to the pricing information of the articles so as to meet the requirements under different shopping scenes and be beneficial to improving the adaptability of the intelligent supermarket/shopping mall setting.
In one embodiment, the method further comprises:
s5, the database module manages the article information and the intelligent article identification model;
in one embodiment, step S5 includes:
s51, the database module stores article information of articles and a trained intelligent article identification model;
s52, inputting the article information into a database module and managing the article information in the database.
Meanwhile, the database module is built based on a background terminal or a cloud platform, and a manager can input and set and adjust article information of commodities in real time through the database module and synchronously update the database to meet the requirements of commodity information management.
Corresponding to the intelligent weighing method shown in the embodiment of fig. 1, the intelligent weighing system based on AI image recognition shown in the embodiment of fig. 2 includes: the system comprises a weighing platform module, an image acquisition module, an intelligent analysis module and an interaction module; wherein,
the weighing platform module is used for placing an object to be weighed, acquiring weight information of the object to be weighed and transmitting the acquired weight information to the intelligent analysis module;
the image acquisition module is used for acquiring an image of the object to be weighed placed on the weighing platform and transmitting the acquired image of the object to be weighed to the intelligent analysis module;
The intelligent analysis module is used for carrying out article identification analysis according to the acquired image of the article to be weighed, obtaining identification information of the article to be weighed, and acquiring associated article information according to the identification information of the article to be weighed; obtaining corresponding pricing information according to the obtained article information and weight information;
the interaction module is used for displaying article information, weight information and pricing information of the articles to be weighed, and generating corresponding payment information according to the pricing information for displaying.
In one embodiment, the weighing platform module comprises a storage unit, an electronic scale unit and a driving unit;
the object placing unit is used for placing objects to be weighed;
the electronic scale unit is used for acquiring weight information of the object to be weighed placed on the object placing unit and transmitting the acquired weight information to the intelligent analysis module;
the driving unit is used for sending a driving signal to the image acquisition module when detecting that the object to be weighed is prevented from being placed on the object placing unit.
In one embodiment, the image acquisition module comprises a camera unit;
the camera unit is used for collecting an image of the object to be weighed placed on the weighing platform according to the received driving signal and transmitting the collected image of the object to be weighed to the intelligent analysis module.
In one embodiment, the image acquisition module further comprises an illumination unit;
the illumination unit is used for providing a light source for a shooting area of the camera unit.
In one embodiment, the intelligent analysis module comprises an image preprocessing unit, a pre-detection unit, an intelligent analysis unit, a matching unit and a calculation unit; wherein,
the image preprocessing unit is used for preprocessing the acquired image of the object to be weighed to obtain a preprocessed image of the object to be weighed;
the intelligent analysis unit is used for identifying the object to be weighed based on the trained intelligent object identification model according to the preprocessed object image to be weighed, and obtaining identification information of the object to be weighed;
the matching unit is used for matching the article information associated with the article from the database according to the obtained identification information of the article to be weighed, wherein the article information comprises the name, the type, the ID, the place of production, the unit price and the like of the article;
the calculating unit is used for calculating pricing information of the object to be weighed according to the weight information and the object information of the object to be weighed.
In one embodiment, the intelligent analysis module further comprises a model training unit;
the model training unit is used for training the intelligent article identification model according to the constructed training set and the test set to obtain a trained intelligent article identification model.
In one embodiment, the interaction module comprises a display unit, a payment unit and a printing unit; wherein,
the display unit is used for displaying weight information, article information and pricing information of the articles to be weighed;
the payment unit is used for counting the price information of all the objects to be weighed of the user, generating payment information according to the obtained total payment price, and further displaying the payment information by the display unit;
the printing unit is used for printing the receipt or the label of the paid article after the user finishes paying.
In one embodiment, the system further comprises a database module for managing the item information and the intelligent item identification model; the database module comprises a database unit and a data management unit;
the database unit is used for storing article information of articles and a trained intelligent article identification model;
the data management unit is used for inputting the article information and managing the article information in the database.
It should be noted that, each functional module/unit in the intelligent weighing system based on AI image recognition shown in the foregoing embodiment is further used to implement the embodiment corresponding to each method step in the intelligent weighing method based on AI image recognition shown in fig. 1, and the description of the present invention is not repeated again.
According to the intelligent weighing method and system based on the AI image recognition, the image information of the article is collected when the article is weighed, the AI image recognition is carried out according to the collected image, the article recognition result is obtained, the pricing information of the article is calculated intelligently, the intelligent level of the article weighing is improved, the labor cost in the article weighing process is reduced, and the accuracy and reliability of the intelligent weighing of the article are improved.
It should be noted that, in each embodiment of the present invention, each functional unit/module may be integrated in one processing unit/module, or each unit/module may exist alone physically, or two or more units/modules may be integrated in one unit/module. The integrated units/modules described above may be implemented either in hardware or in software functional units/modules.
From the above description of embodiments, it will be apparent to those skilled in the art that the embodiments described herein may be implemented in hardware, software, firmware, middleware, code, or any suitable combination thereof. For a hardware implementation, the processor may be implemented in one or more of the following units: an Application Specific Integrated Circuit (ASIC), a Digital Signal Processor (DSP), a Digital Signal Processing Device (DSPD), a Programmable Logic Device (PLD), a Field Programmable Gate Array (FPGA), a processor, a controller, a microcontroller, a microprocessor, other electronic units designed to perform the functions described herein, or a combination thereof. For a software implementation, some or all of the flow of an embodiment may be accomplished by a computer program to instruct the associated hardware. When implemented, the above-described programs may be stored in or transmitted as one or more instructions or code on a computer-readable medium. Computer-readable media includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another. A storage media may be any available media that can be accessed by a computer. The computer readable media can include, but is not limited to, RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer.
Finally, it should be noted that the above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the scope of the present invention, and although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solution of the present invention without departing from the spirit and scope of the technical solution of the present invention.

Claims (8)

1. The intelligent weighing method based on AI image recognition is characterized by comprising the following steps:
s1, placing an object to be weighed by a weighing platform module, acquiring weight information of the object to be weighed, and transmitting the acquired weight information to an intelligent analysis module;
s2, an image acquisition module acquires an image of an object to be weighed placed on the weighing platform, and transmits the acquired image of the object to be weighed to an intelligent analysis module;
s3, the intelligent analysis module performs article identification analysis according to the acquired image of the article to be weighed, obtains identification information of the article to be weighed, and acquires associated article information according to the identification information of the article to be weighed; and obtaining corresponding pricing information according to the obtained article information and weight information, including:
s31, the intelligent analysis module performs pretreatment on the acquired image of the object to be weighed to obtain a pretreated image of the object to be weighed; the method specifically comprises the following steps:
According to the obtained image X of the object to be weighed, performing edge detection on the image X of the object to be weighedMeasuring to obtain edge characteristic information of the image X of the object to be weighed, comparing the obtained edge characteristic information with the edge characteristic information of the standard weighing platform image B, and extracting a foreground area sigma of the image of the object to be weighed a And a background area sigma b The method comprises the steps of carrying out a first treatment on the surface of the The standard weighing platform image B is obtained by shooting the image acquisition module at the same position or the angle symmetry platform module;
wavelet decomposition is carried out on the image X of the object to be weighed to obtain a low-frequency component image f of the image X of the object to be weighed d And a high-frequency component image f h Wherein the decomposition scale adopted by wavelet decomposition is 1 layer, and the wavelet base is db2;
from the obtained low-frequency component image f d Further the low frequency component image f d Conversion from RGB color space to LAB color space to obtain low frequency component image f d A luminance component eL, a color component eA, and a color component eB;
local brightness adjustment is carried out on the foreground area according to the obtained brightness component eL to obtain an adjusted brightness componentThe local brightness adjustment function used is:
in the method, in the process of the invention,representing luminance values of pixel points (x, y) after local luminance adjustment, wherein the pixel points (x, y) e sigma a ，Represented in background area sigma b The nearest pixel point to the pixel point (x, y); />Representing pixel dot +.>Luminance value of>Representing background area sigma b Average luminance value of L G Represents a set standard luminance value, wherein +.> Representing a foreground region v a L (x, y) represents the luminance value of the pixel point (x, y),represents a judgment function in which when->And L G When the positive and negative signs of L (x, y) are the same, e.g.And L is G -L (x, y) > 0, or +.> And->ThenOtherwise-> ω 1 Representation ofA set local regulatory factor, wherein ω 1 ∈[0.3,0.4]，ω 2 Representing a foreground modulator, wherein omega 2 ∈[0.2,0.4]，ω 3 Representing an abnormality-compensating regulatory factor, wherein ω 3 ∈[0.2,0.3]，ω 1 ≥ω 2 ≥ω 3 ；
Based on the adjusted luminance componentReconstructing the color component eA and the color component eB to obtain an adjusted low-frequency component image +.>
From the obtained high-frequency component image f h For the high frequency component image f h Performing detail adjustment processing to obtain a high-frequency component image after the detail adjustment processingThe detail adjusting function adopted is as follows:
in the method, in the process of the invention,representing gray values of pixel points (x, y) after detail adjustment processing, and H (x, y) represents the high-frequency component image f h Gray values of the middle pixel points (x, y); (i, j) represents a pixel point in a 3×3 neighborhood region θ (x, y) centered on the pixel point (x, y), max (|h (x, y) -H (i, j) |) represents a maximum value of differences between gray values of the pixel point (x, y) and respective pixel points in its neighborhood region, and min (|h (x, y) -H (i, j) |) represents a minimum value of differences between gray values of the pixel point (x, y) and respective pixel points in its neighborhood region; t1 represents a set first gray threshold, T1 ε [150,200 ] ]；H θ(x,y) Represents the average gray value of a 3×3 neighborhood region centered on the pixel point (x, y), T2 represents the set second gray levelThreshold, where T2 ε [50,70]The method comprises the steps of carrying out a first treatment on the surface of the N represents the number of pixel points in the neighborhood region theta (x, y), H k Representing the set compensation gray value, where H k ∈[45,55]；ω 4 Represents a set adjustment factor, wherein ω 4 ∈[0.8,0.9]，ω 5 Representing a set compensation factor, wherein ω 5 ∈[0.2,0.3]；
From the adjusted low-frequency component imageAnd a high-frequency component image after detail adjustment processing +.>Reconstructing to obtain an enhanced image of the object to be weighed;
according to the obtained enhanced image of the object to be weighed, the background area is further removed, and only the foreground area is reserved, so that the preprocessed image of the object to be weighed is obtained;
s32, the intelligent analysis module identifies the object to be weighed based on the trained intelligent object identification model according to the preprocessed object image to be weighed, and identification information of the object to be weighed is obtained;
s33, the intelligent analysis module matches article information associated with the article from a database according to the obtained identification information of the article to be weighed, wherein the article information comprises an article name, a type, an article ID, a production place and a unit price;
s34, calculating pricing information of the object to be weighed according to the weight information and the object information of the object to be weighed by the intelligent analysis module; and S4, the interaction module displays the article information, the weight information and the pricing information of the article to be weighed, and generates corresponding payment information according to the pricing information for display.
2. The intelligent weighing method based on AI image recognition of claim 1, wherein step S1 includes:
s11, placing an object to be weighed to a weighing platform module;
s12, the weighing platform module acquires weight information of the object to be weighed placed on the object placing unit and transmits the acquired weight information to the intelligent analysis module;
and S13, when the object placing unit is detected to prevent the object to be weighed, the weighing platform module sends a driving signal to the image acquisition module.
3. The intelligent weighing method based on AI image recognition of claim 1, wherein step S2 includes: the image acquisition module acquires an image of the object to be weighed placed on the weighing platform according to the received driving signal, and transmits the acquired image of the object to be weighed to the intelligent analysis module.
4. The intelligent weighing method based on AI image recognition according to claim 1, wherein the trained intelligent object recognition model is built based on a CNN neural convolution network, and comprises an input layer, a first convolution layer, a second convolution layer, a pooling layer, a first full-connection layer, a second full-connection layer, an activation layer and an output layer which are sequentially arranged;
the method comprises the steps of inputting a preprocessed image of an object to be weighed into an input layer, wherein a first convolution layer and a second convolution layer respectively comprise 16 convolution kernels and 32 convolution kernels, and the sizes of the convolution kernels are 3 multiplied by 3; the pooling layer is set as the largest pooling layer, and the size of the pooling core is 3 multiplied by 3; the first full-connection layer is provided with 64 neurons, the second full-connection layer is provided with 16 neurons, the activation layer adopts a softmax function, the characteristic vector of the reaction object is output through the full-connection layer, and the identification probability of the object to be weighed is obtained through the softmax function according to the characteristic vector; and the output layer outputs the identification result of the object to be weighed according to the obtained identification probability.
5. The AI-image-recognition-based intelligent weighing method of claim 4, wherein step S3 further comprises:
and S30, training the intelligent article identification model by the intelligent analysis module according to the constructed training set and the test set to obtain a trained intelligent article identification model.
6. The intelligent weighing method based on AI image recognition of claim 1, wherein step S4 includes:
s41, the interaction module displays weight information, article information and pricing information of the articles to be weighed;
s42, the interaction module counts price information of all the objects to be weighed of the user, generates payment information according to the obtained payment total price, and further displays the payment information by the display unit;
s43, the interaction module prints a receipt or label of the paid article after the user finishes paying.
7. The AI-image-recognition-based intelligent weighing method according to claim 1, further comprising the step of managing the item information and the intelligent item recognition model by a database module S5;
the step S5 specifically includes:
s51, the database module stores article information of articles and a trained intelligent article identification model;
s52, inputting the article information into a database module and managing the article information in the database.
8. Intelligent weighing system based on AI image recognition, characterized by comprising: the system comprises a weighing platform module, an image acquisition module, an intelligent analysis module and an interaction module; wherein,
the weighing platform module is used for placing an object to be weighed, acquiring weight information of the object to be weighed and transmitting the acquired weight information to the intelligent analysis module;
the image acquisition module is used for acquiring an image of the object to be weighed placed on the weighing platform and transmitting the acquired image of the object to be weighed to the intelligent analysis module;
the intelligent analysis module is used for carrying out article identification analysis according to the acquired image of the article to be weighed, obtaining identification information of the article to be weighed, and acquiring associated article information according to the identification information of the article to be weighed; obtaining corresponding pricing information according to the obtained article information and weight information;
the interaction module is used for displaying article information, weight information and pricing information of the articles to be weighed, and generating corresponding payment information according to the pricing information for display;
the intelligent analysis module comprises an image preprocessing unit, a pre-detection unit, an intelligent analysis unit, a matching unit and a calculation unit; wherein,
the image preprocessing unit is used for preprocessing the acquired image of the object to be weighed to obtain a preprocessed image of the object to be weighed;
The intelligent analysis unit is used for identifying the object to be weighed based on the trained intelligent object identification model according to the preprocessed object image to be weighed, and obtaining identification information of the object to be weighed;
the matching unit is used for matching the article information associated with the article from the database according to the obtained identification information of the article to be weighed, wherein the article information comprises the name, the type, the ID, the place of production and the unit price of the article;
the calculating unit is used for calculating pricing information of the object to be weighed according to the weight information and the object information of the object to be weighed;
the method for preprocessing the acquired image of the object to be weighed specifically comprises the following steps:
according to the obtained object image X to be weighed, carrying out edge detection processing on the object image X to be weighed to obtain edge characteristic information of the object image X to be weighed, and according to comparison of the obtained edge characteristic information and the edge characteristic information of the standard weighing table image B, extracting a foreground area sigma of the object image to be weighed a And a background area sigma b The method comprises the steps of carrying out a first treatment on the surface of the The standard weighing platform image B is obtained by shooting the image acquisition module at the same position or the angle symmetry platform module;
wavelet decomposition is carried out on the image X of the object to be weighed to obtain a low-frequency component image f of the image X of the object to be weighed d And a high-frequency component image f h Wherein the decomposition scale adopted by wavelet decomposition is 1 layer, and the wavelet base is db2;
from the obtained low-frequency component image f d Further the low frequency component image f d Conversion from RGB color spaceTo LAB color space, obtain low-frequency component image f d A luminance component eL, a color component eA, and a color component eB;
local brightness adjustment is carried out on the foreground area according to the obtained brightness component eL to obtain an adjusted brightness componentThe local brightness adjustment function used is:
in the method, in the process of the invention,representing luminance values of pixel points (x, y) after local luminance adjustment, wherein the pixel points (x, y) e sigma a ，Represented in background area sigma b The nearest pixel point to the pixel point (x, y); />Representing pixel dot +.>Luminance value of>Representing background area sigma b Average luminance value of L G Represents a set standard luminance value, wherein L G ∈[60,65]；/>Representing foreground region sigma a L (x, y) represents the luminance value of the pixel point (x, y),representing judgmentBreak function, wherein->And L G When the positive and negative signs of L (x, y) are the same, e.g.And L is G -L (x, y) > 0, or +.> And L is G L (x, y) < 0, thenOtherwise-> ω 1 Representing the local adjustment factor of the setting, wherein ω 1 ∈[0.3,0.4]，ω 2 Representing a foreground modulator, wherein omega 2 ∈[0.2,0.4]，ω 3 Representing an abnormality-compensating regulatory factor, wherein ω 3 ∈[0.2,0.3]，ω 1 ≥ω 2 ≥ω 3 ；
Based on the adjusted luminance componentReconstructing the color component eA and the color component eB to obtain an adjusted low-frequency component image +.>
From the obtained high-frequency component image f h For the high frequency component image f h Performing detail adjustment processing to obtain a high-frequency component image after the detail adjustment processingThe detail adjusting function adopted is as follows:
in the method, in the process of the invention,representing gray values of pixel points (x, y) after detail adjustment processing, and H (x, y) represents the high-frequency component image f h Gray values of the middle pixel points (x, y); (i, j) represents a pixel point in a 3×3 neighborhood region θ (x, y) centered on the pixel point (x, y), max (|h (x, y) -H (i, j) |) represents a maximum value of differences between gray values of the pixel point (x, y) and respective pixel points in its neighborhood region, and min (|h (x, y) -H (i, j) |) represents a minimum value of differences between gray values of the pixel point (x, y) and respective pixel points in its neighborhood region; t1 represents a set first gray threshold, T1 ε [150,200 ]]；H θ(x,y) Represents the average gray value of a 3 x 3 neighborhood region centered on pixel (x, y), T2 represents a set second gray threshold, where T2 ε [50,70 ]]The method comprises the steps of carrying out a first treatment on the surface of the N represents the number of pixel points in the neighborhood region theta (x, y), H k Representing the set compensation gray value, where H k ∈[45,55]；ω 4 Represents a set adjustment factor, wherein ω 4 ∈[0.8,0.9]，ω 5 Representing a set compensation factor, wherein ω 5 ∈[0.2,0.3]；
From the adjusted low-frequency component imageAnd a high-frequency component image after detail adjustment processing +.>Reconstructing to obtain an enhanced image of the object to be weighed;
and further removing the background area according to the obtained enhanced image of the object to be weighed, and reserving only the foreground area to obtain the preprocessed image of the object to be weighed.
CN202310313233.6A 2023-03-27 2023-03-27 Intelligent weighing method and system based on AI image recognition Active CN116403197B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310313233.6A CN116403197B (en) 2023-03-27 2023-03-27 Intelligent weighing method and system based on AI image recognition

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310313233.6A CN116403197B (en) 2023-03-27 2023-03-27 Intelligent weighing method and system based on AI image recognition

Publications (2)

Publication Number Publication Date
CN116403197A CN116403197A (en) 2023-07-07
CN116403197B true CN116403197B (en) 2024-01-26

Family

ID=87015390

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310313233.6A Active CN116403197B (en) 2023-03-27 2023-03-27 Intelligent weighing method and system based on AI image recognition

Country Status (1)

Country Link
CN (1) CN116403197B (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104361698A (en) * 2014-11-25 2015-02-18 湖南大学 Self-service intelligent electronic weighing settlement method and system
CN111082517A (en) * 2019-10-14 2020-04-28 上海大学 Intelligent electrical equipment management device based on cloud computing
CN111814614A (en) * 2020-06-28 2020-10-23 袁精侠 Intelligent object-identifying electronic scale weighing method and system
CN113744557A (en) * 2021-09-03 2021-12-03 武汉理工大学 Intelligent parking system based on Internet of things technology
CN114235122A (en) * 2021-12-16 2022-03-25 广州市超赢信息科技有限公司 Weighing settlement method and system of electronic scale based on AI image recognition

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104361698A (en) * 2014-11-25 2015-02-18 湖南大学 Self-service intelligent electronic weighing settlement method and system
CN111082517A (en) * 2019-10-14 2020-04-28 上海大学 Intelligent electrical equipment management device based on cloud computing
CN111814614A (en) * 2020-06-28 2020-10-23 袁精侠 Intelligent object-identifying electronic scale weighing method and system
CN113744557A (en) * 2021-09-03 2021-12-03 武汉理工大学 Intelligent parking system based on Internet of things technology
CN114235122A (en) * 2021-12-16 2022-03-25 广州市超赢信息科技有限公司 Weighing settlement method and system of electronic scale based on AI image recognition

Also Published As

Publication number Publication date
CN116403197A (en) 2023-07-07

Similar Documents

Publication Publication Date Title
CN108922026B (en) Replenishment management method and device for vending machine and user terminal
US11151427B2 (en) Method and apparatus for checkout based on image identification technique of convolutional neural network
CN208044756U (en) Self-service system
CN108921645B (en) Commodity purchase judgment method and device and user terminal
WO2019184646A1 (en) Method and device for identifying merchandise, merchandise container
CN108320404A (en) Commodity recognition method, device, self-service cashier based on neural network
CN107767590A (en) Automatic identification commercialization bar code electronic scale and Automatic identification method
CN108596137A (en) A kind of commodity scanning record method based on image recognition algorithm
CN109682451A (en) One self-help intelligent identification valuation and label printing system
US11669948B2 (en) Learned model generating method, learned model generating device, product identifying method, product identifying device, product identifying system, and measuring device
CN110836717A (en) Financial service-oriented intelligent fruit and vegetable identification and pricing system
CN116403197B (en) Intelligent weighing method and system based on AI image recognition
WO2021048813A1 (en) Scale and method for the automatic recognition of a product
CN115761457A (en) Commodity identification method and device, electronic equipment and computer readable medium
US20230104061A1 (en) Methods and apparatuses for determining statuses of positions on shelves, shelves and non-transitory computer-readable storage media
US11748702B2 (en) Methods of determining status of position in shelf, shelves and non-transitory computer-readable storage media
CN111639928B (en) Bulk commodity metering device and method based on convolutional neural network
CN114612897A (en) Intelligent fruit and vegetable weighing and ticketing method and device, electronic equipment and storage medium
CN114235122A (en) Weighing settlement method and system of electronic scale based on AI image recognition
CN109374042B (en) Quality inspection system and method for intelligent customized furniture assembly parts
CN112763041A (en) Fresh food identification weighing electronic scale and identification method
CN109857880B (en) Model-based data processing method and device and electronic equipment
WO2023138447A1 (en) Ai weighing system, and method for improving precision of ai model by using various types of data sets
CN113077308B (en) Intelligent visual commercial transaction platform
CN107944860A (en) A kind of bakery identification and cash register system and method based on neutral net

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
GR01 Patent grant
GR01 Patent grant