CN115272446B - Method and system for calculating head-stacking occupied area - Google Patents

Method and system for calculating head-stacking occupied area Download PDF

Info

Publication number
CN115272446B
CN115272446B CN202211178396.XA CN202211178396A CN115272446B CN 115272446 B CN115272446 B CN 115272446B CN 202211178396 A CN202211178396 A CN 202211178396A CN 115272446 B CN115272446 B CN 115272446B
Authority
CN
China
Prior art keywords
pile head
commodity
head
pile
length
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
CN202211178396.XA
Other languages
Chinese (zh)
Other versions
CN115272446A (en
Inventor
刘国俭
许允杰
王炳璇
刘昭
李曼曼
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.)
Nanjing Zhangkong Network Science & Technology Co ltd
Original Assignee
Nanjing Zhangkong Network Science & Technology 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 Nanjing Zhangkong Network Science & Technology Co ltd filed Critical Nanjing Zhangkong Network Science & Technology Co ltd
Priority to CN202211178396.XA priority Critical patent/CN115272446B/en
Publication of CN115272446A publication Critical patent/CN115272446A/en
Application granted granted Critical
Publication of CN115272446B publication Critical patent/CN115272446B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/60—Analysis of geometric attributes
    • G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
    • G01B11/00—Measuring arrangements characterised by the use of optical techniques
    • G01B11/28—Measuring arrangements characterised by the use of optical techniques for measuring areas
    • 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
    • G06V10/30—Noise filtering
    • 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
    • G06V10/762—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using clustering, e.g. of similar faces in social networks
    • 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
    • G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
    • G06V10/765—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects using rules for classification or partitioning the feature space
    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V2201/00—Indexing scheme relating to image or video recognition or understanding
    • G06V2201/07—Target detection
    • 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 a method for calculating the floor area of a pile head, which is characterized in that a pile head front photo and a side photo meeting the requirements are limited and submitted through repeated judgment and positive shooting judgment, the front length and the side width of the pile head are automatically calculated according to the leftmost and rightmost coordinates of a commodity identified in the pile head photo after the pile head photo is identified, and errors caused by missing identification or error identification or shooting inclination and the like are reduced, so that the floor area of the pile head is quickly calculated.

Description

Method and system for calculating head-stacking occupied area
Technical Field
The invention relates to the technical field of computer vision, in particular to a method and a system for calculating the occupied area of a pile head.
Background
The ground heap, also known as a heap or stack, is a popular form of display promotion by merchants, and serves to display the image of the goods and promote sales. The excess area of the supplier is limited, and the heap head generally needs a certain fee paid by the supplier to apply. Therefore, the floor area of the pile head is an important item for display and check in the visiting process of the service staff, the service staff needs to estimate the floor area of the pile head and ensure the full utilization of the pile head, and the commodity supplier needs to be capable of quickly, efficiently and accurately calculating the floor area.
The Chinese patent application CN109784172A provides a method, a system, equipment and a medium for estimating the floor area of ground heap commodities in a market, wherein the method inputs a plurality of collected images into a preset scene classification model, and screens the plurality of ground heap commodity images from the plurality of images through the scene classification model; through a preset commodity identification model, identifying each ground-push commodity area in each ground-pile commodity image and generating a polygonal marking frame marked out of a plurality of ground-pile commodity areas through lines; and estimating the occupied area of the ground-pushed commodities in the ground-piled commodity image according to the polygonal bodies of the labeling frames corresponding to the ground-pushed commodity areas in each ground-piled commodity image.
Application CN113034574A provides a commodity ground heap area calculation method and system based on target detection, the method comprises extracting ground heap scenes of picture recognition results, and calculating projection edges of ground heap bottom layers; judging the type of the ground heap scene according to the projection edge, and matching different methods for different types to calculate to obtain a layered projection line; and calculating the ground pile area according to the longest projection line in the layered projection lines.
The existing technical scheme for calculating the ground heap area depends on the recognition result of a commodity recognition model, and the floor area of each commodity is calculated through the recognition result of the commodity recognition model, but the problem of false recognition and missing recognition inevitably exists in the commodity recognition model, so that the calculation error of the ground heap area is increased, and when a salesman takes a picture, the picture is taken irregularly, such as inclined picture taking, meanwhile, the shot ground pushes the front and the side, and the like, so that the error of the prior technical scheme is increased.
Disclosure of Invention
The invention aims to provide a method and a system for calculating the floor area of a pile head, which aim to solve the problem of calculation errors in the prior art. According to the scheme, the front photo and the side photo of the pile head meeting the requirements are submitted in a limiting mode, the front length and the side width of the pile head are automatically calculated after the pile head photo is recognized, and errors caused by missing recognition, error recognition or photographing inclination and the like are reduced, so that the occupied area of the pile head is rapidly calculated, the ground area saving verification efficiency of an enterprise is facilitated, and the enterprise management cost is reduced.
In order to achieve the purpose, the invention provides the following technical scheme:
a method of pile head footprint calculation, the method comprising the steps of:
1) Acquiring images, establishing a commodity identification model training set of commodities appearing in a pile head, labeling the commodities by using a rectangular labeling frame, training a commodity identification model by using a deep learning technology, and simultaneously recording length, width and height information of the commodities in the commodity image acquisition process;
2) Respectively acquiring a front photograph and a side photograph of a pile head to be calculated, respectively identifying commodities in a pile head image by using a trained commodity identification model, and acquiring commodity rectangular frames of the front photograph and the side photograph of the pile head;
3) Denoising recognition results of the pile-head front side illumination and the side illumination respectively, and removing non-pile-head commodities in the pile-head front side illumination and the side illumination;
4) Judging whether the front photograph and the side photograph are the same face, if so, not calculating the area of the pile head, and if not, performing the next step;
5) Judging whether the front side photograph and the side photograph are side photographed images, namely whether the front side photograph and the side photograph contain images of two faces of the pile head, if so, not calculating the area of the pile head, and if not, performing the next step;
6) Respectively calculating the length and the width of the pile head;
7) And calculating the area of the pile head, and calculating to obtain the area of the pile head according to the calculated length of the front surface and the width of the side surface of the pile head.
Preferably, in step 1), the deep learning techniques utilized include, but are not limited to, the Faster R-CNN or SSD destination detection algorithms.
In the above method, in step 3), the process of denoising the recognition result of the front illumination and the side illumination of the pile head includes: calculating the coordinate of the center point of each commodity rectangular frame according to the coordinates of the commodity rectangular frames, acquiring the coordinate sets of the center points of the front side illumination commodity and the side illumination commodity of the pile head, clustering the coordinate set of the center point of each graph by using a density clustering algorithm, acquiring the cluster of the center point of the commodity coordinate and the point far away from the cluster, regarding the commodity represented by the point far away from the cluster as a non-pile head commodity, and removing the commodity far away from the cluster in the recognition results of the front side illumination commodity and the side illumination of the pile head.
In the above method, the process of determining whether the front photograph and the side photograph are the same face in step 4) includes: after denoising the head front photo and the side photo, calculating the center point coordinate of each commodity rectangular frame according to the commodity rectangular frame coordinates, sorting commodities on the front photo and the side photo from left to right and from top to bottom according to the center point coordinates respectively, comparing the commodity sequences after the front photo and the side photo are sorted, and judging the commodities to be the image of the same surface if the two sequences are the same.
In the above method, the process of determining whether the front photograph and the side photograph are side-shot images in step 5) includes:
5.1 Intercepting the area where the pile-head commodity is located according to the denoised commodity coordinate to be regarded as a pile-head image, detecting straight lines in the pile-head image by using a straight line detection algorithm, and obtaining a straight line set on each pile-head image, wherein each straight line is represented as y = Ax + b, A is the slope of each straight line, and b is the offset of each straight line;
5.2 Constructing a side-shot image and a forward-shot image training set, acquiring a linear slope on each collected pile-head image as the training set, wherein the label of the forward-shot image is 1, the label of the side-shot image is 0, and training a forward-shot side-shot two-classification model;
5.3 Whether the pile head image to be calculated is a side shot image is judged by using the trained two classification models, if the pile head image to be calculated is the side shot image, the pile head image is directly returned, the calculation of the ground pile area is not carried out, and the calculation deviation caused by the side shot is avoided.
Preferably, the line detection algorithm is a hough line detection method, and the binary model is a support vector machine or a random forest.
In the above method, in step 6), the process of calculating the length and width of the stack head respectively includes:
6.1 According to the target detection result, calculating the length of the pile head in the picture for the picture on the front side of the pile head, calculating the pixel length X with the maximum interval in the horizontal direction in all the identified commodities in the identification result, taking the pixel length X1 of the commodity with the identification position in the middle of the picture, and obtaining the length attribute L of the commodity, and then calculating the length of the pile head as the length of the pile head by multiplying the length of the pile head by X1 after dividing the length of the pile head by X1;
6.2 The width of the pile head in the picture is calculated according to the target detection result, the pixel length X with the maximum interval among the left commodity and the right commodity in the identification result is calculated, the pixel length X1 of the commodity with the identification position in the middle in the picture is taken, the width attribute D of the commodity is obtained, and then the width of the pile head is calculated as the width of the pile head which is obtained by multiplying the X by the X1.
Preferably, in the calculation of the length of the pile head, acquiring a plurality of pictures of the front of the pile head and calculating the length, and taking the average value of the calculated lengths of the plurality of pictures as the length of the pile head; meanwhile, in the calculation of the width of the pile head, a plurality of pictures of the side face of the pile head are collected, the width is calculated, and the average value of the calculated widths of the plurality of pictures is taken as the width of the pile head.
The invention also provides a system for head stack floor space calculation, which comprises a storage medium and a calculation module, wherein the storage module is used for storing a program for executing the method, and the calculation module is used for executing the program stored in the storage module.
The invention also provides a device for pile head floor area calculation, comprising a shooting device, a memory and a processor, wherein the shooting device is used for acquiring images, the memory is used for storing programs for executing the method, and the processor is used for executing the programs stored in the memory.
Compared with the prior art, the invention has the beneficial effects that:
1. calculating the area of the pile head by using the front image and the side image of the pile head, and repeatedly judging whether the two images contain two surfaces or not and judging whether the images contain two surfaces or not by oblique shooting, so that the calculation accuracy of the area of the pile head is improved;
2. the method has the advantages that the linear characteristics of the pile-head image region are extracted to establish a two-classification model, so that the accuracy of forward shooting and side shooting classification directly by using images is improved, and the submission of the images which do not meet the requirements can be limited;
3. and calculating the pixel length of the commodity area by utilizing the leftmost and rightmost coordinates of the commodity area of the image, and avoiding the error caused by target detection and missed identification.
Drawings
FIG. 1 is a schematic overall flow diagram of the process of the present invention;
fig. 2 is a schematic flow chart of the determination of whether the submitted image is a side-shot image.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In the description of the present invention, it should be noted that the terms "upper", "lower", "inner", "outer", "front", "rear", "both ends", "one end", "the other end", and the like indicate orientations or positional relationships based on orientations or positional relationships shown in the drawings, and are only for convenience of description and simplification of description, but do not indicate or imply that the device or element referred to must have a specific orientation, be configured in a specific orientation, and operate, and thus, should not be construed as limiting the present invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.
Referring to fig. 1, the method for calculating the pile head area of the present invention includes the steps of collecting an image training model, collecting front and side illumination of the pile head, denoising, judging whether the image is a side shot, calculating the length and width of the pile head, calculating the occupied area of the pile head, and the like. The method comprises the steps of training by utilizing collected commodity data to obtain a commodity identification model, marking the data by adopting a rectangular frame mark of the existing marking method, and adopting an existing deep learning model training method such as fast rcnn by adopting the training method. And identifying and judging the collected pile head frontal illumination and the collected side illumination by using the trained model, eliminating errors, and obtaining accurate length and width so as to calculate accurate pile head area. These steps will be described in more detail below.
As shown in the figure, firstly, a commodity identification model training set of commodities appearing in a pile head is established, the commodities are marked by using a rectangular marking box, and a target detection model is trained by using the acquired image by using a fast R-CNN or SSD target detection algorithm included in a deep learning technology but not limited thereto. And simultaneously recording the information of the length L, the width D, the height H and the like of the commodity in the commodity acquisition process.
After the model is trained, the front and side views of the pile head are processed by the model. Specifically, the front photo and the side photo of the pile head to be calculated are respectively collected, the trained commodity identification model is used for respectively identifying commodities in the pile head image, and the commodity rectangular frames of the front photo and the side photo of the pile head are obtained.
And denoising the recognition results of the front illumination and the side illumination of the pile head respectively. Calculating the central point coordinate of each commodity rectangular frame according to the commodity rectangular frame coordinates, obtaining the central point coordinate sets of the front face and side face of the head, clustering the central point coordinate sets of each graph by using a Density-Based Spatial Clustering of Applications with Noise, obtaining the central point cluster of the commodity coordinates and points far away from the cluster, regarding the commodity represented by the points far away from the cluster as a non-head commodity, and removing the commodity far away from the cluster in the recognition results of the front face and side face of the head.
And after denoising the head front side illumination and the side illumination, judging whether the front side illumination and the side illumination are the same surface. And respectively sequencing the commodities on the front photo and the side photo from left to right and from top to bottom according to the coordinates of the central points of the commodities, comparing the commodity sequences after sequencing the front photo and the side photo, judging that the commodities are images of the same plane if the two sequences are the same, and not calculating the area of the pile head.
Then, it is further determined whether the front and side photographs are side-shot images, that is, images including both faces of the pile head. As shown in fig. 2, the method of determining whether the front photograph and the side photograph are side-shot images is as follows:
1. intercepting a region where a pile-head commodity is located according to the denoised commodity coordinate to serve as a pile-head image, detecting straight lines in the pile-head image by using a straight line detection algorithm (such as a Hough straight line detection method), and obtaining a straight line set on each pile-head image, wherein each straight line can be represented as y = Ax + b, A is the slope of each straight line, and b is the offset of each straight line;
2. constructing a side-shot image and a forward-shot image training set, acquiring a straight slope on each collected pile head image as the training set, wherein the label of the forward-shot image is 1, the label of the side-shot image is 0, and training a forward-shot side-shot two-classification model, including but not limited to a support vector machine, a random forest and the like;
3. and judging whether the pile head image to be calculated is a side shot image or not by using the trained two-classification model, and if the pile head image to be calculated is the side shot image, directly returning without calculating the area of the ground pile, thereby avoiding the calculation deviation caused by the side shot.
And after determining that the front photograph and the side photograph are not side-shot images according to the steps, respectively calculating the length and the width of the pile head.
The calculation steps of the length and width of the pile head are as follows:
1. and calculating the length of the pile head in the picture according to the target detection result. And calculating the pixel length X with the maximum interval in the horizontal direction in all the identified commodities in the identification result, taking the pixel length X1 of the commodity with the identification position in the middle of the picture, obtaining the length attribute L of the commodity, and calculating the length of the pile head as X divided by X1 and multiplied by L. The front pictures of the front of the pile head can be collected and the length can be calculated, and the average value of the calculated lengths of the pictures is taken as the length of the pile head.
2. And calculating the width of the pile head in the picture according to the target detection result. And calculating the pixel length X with the maximum interval in the left and right commodities in the recognition result, taking the pixel length X1 of the commodity with the recognition position in the middle of the picture, acquiring the width attribute D of the commodity, and calculating the width of the pile head as X divided by X1 and multiplied by D. The method can collect photos of the side surfaces of a plurality of pile heads and calculate the width, and the average value of the calculated widths of the plurality of pictures is taken as the width of the pile head.
After the length and width of the pile head are calculated, the area of the pile head can be calculated. And (4) multiplying the calculated front length and the side width of the pile head to calculate the area of the pile head.
The method of the invention limits and submits the pile head front photo and the side photo which meet the requirements through repeated judgment and positive shooting judgment, automatically calculates the front length and the side width of the pile head according to the leftmost coordinate and the rightmost coordinate in the pile head photo after the pile head photo is identified, reduces errors caused by missing identification or error identification or shooting inclination and the like, and further quickly calculates the occupied area of the pile head.
The invention is not described in detail, but is well known to those skilled in the art.
Finally, it is to be noted that: although the present invention has been described in detail with reference to examples, it will be understood by those skilled in the art that various changes may be made and equivalents may be substituted without departing from the spirit and scope of the invention as defined in the appended claims.

Claims (9)

1. A method of pile head footprint calculation, said method comprising the steps of:
1) Acquiring images, establishing a commodity identification model training set of commodities appearing in a pile head, labeling the commodities by using a rectangular labeling frame, training a commodity identification model by using a deep learning technology, and simultaneously recording length, width and height information of the commodities in the commodity image acquisition process;
2) Respectively acquiring a front photograph and a side photograph of a pile head to be calculated, respectively identifying commodities in a pile head image by using a trained commodity identification model, and acquiring commodity rectangular frames of the front photograph and the side photograph of the pile head;
3) Denoising recognition results of the pile-head front side illumination and the side illumination respectively, and removing non-pile-head commodities in the pile-head front side illumination and the side illumination;
4) Judging whether the front photograph and the side photograph are the same face, if so, not calculating the area of the pile head, and if not, performing the next step;
5) Judging whether the front photograph and the side photograph are side photographed images, namely whether the front photograph and the side photograph contain images of two faces of the pile head, if so, not calculating the area of the pile head, and if not, performing the next step;
6) Respectively calculating the length and the width of the pile head;
7) Calculating the area of the pile head, calculating the area of the pile head according to the calculated length of the front surface and the width of the side surface of the pile head,
wherein in step 6), the process of calculating the length and width of the pile head respectively comprises:
6.1 According to the target detection result, calculating the length of the pile head in the picture for the picture on the front side of the pile head, calculating the pixel length X with the maximum interval in the horizontal direction in all the identified commodities in the identification result, taking the pixel length X1 of the commodity with the identification position in the middle of the picture, and obtaining the length attribute L of the commodity, and then calculating the length of the pile head as the length of the pile head by multiplying the length of the pile head by X1 after dividing the length of the pile head by X1;
6.2 The width of the pile head in the picture is calculated according to the target detection result, the pixel length X with the maximum interval between the left commodity and the right commodity in the identification result is calculated, the pixel length X1 of the commodity with the identification position in the middle in the picture is taken, the width attribute D of the commodity is obtained, and then the width of the pile head is calculated as the width of the pile head which is obtained by multiplying the width of the pile head by X1.
2. The method of heap head footprint computation of claim 1 wherein in step 1) the deep learning techniques utilized include, but are not limited to, the Faster R-CNN or SSD destination detection algorithms.
3. The method for calculating the head footprint of claim 1, wherein in step 3), the denoising process for the recognition result of the head front and side illumination comprises: calculating the coordinate of the center point of each commodity rectangular frame according to the coordinates of the commodity rectangular frames, acquiring the coordinate sets of the center points of the front side illumination commodity and the side illumination commodity of the pile head, clustering the coordinate set of the center point of each graph by using a density clustering algorithm, acquiring the cluster of the center point of the commodity coordinate and the point far away from the cluster, regarding the commodity represented by the point far away from the cluster as a non-pile head commodity, and removing the commodity far away from the cluster in the recognition results of the front side illumination commodity and the side illumination of the pile head.
4. The method of calculating a pile-up footprint of claim 1, in which the process of determining whether the front-view and the side-view are the same face in step 4) comprises: after denoising the head front photo and the side photo, calculating the center point coordinate of each commodity rectangular frame according to the commodity rectangular frame coordinates, sorting commodities on the front photo and the side photo from left to right and from top to bottom according to the center point coordinates respectively, comparing the commodity sequences after the front photo and the side photo are sorted, and judging the commodities to be the image of the same surface if the two sequences are the same.
5. The method of pile-up head footprint calculation according to claim 1, in step 5), the process of determining whether the front and side shots are side shots comprises:
5.1 Intercepting the area where the pile-head commodity is located according to the denoised commodity coordinate to be regarded as a pile-head image, detecting straight lines in the pile-head image by using a straight line detection algorithm, and obtaining a straight line set on each pile-head image, wherein each straight line is represented as y = Ax + b, A is the slope of each straight line, and b is the offset of each straight line;
5.2 Constructing a side-shot image and a forward-shot image training set, acquiring a linear slope on each collected pile-head image as the training set, wherein the label of the forward-shot image is 1, the label of the side-shot image is 0, and training a forward-shot side-shot two-classification model;
5.3 Whether the pile head image to be calculated is a side shot image is judged by using the trained two classification models, if the pile head image to be calculated is the side shot image, the pile head image is directly returned, the calculation of the ground pile area is not carried out, and the calculation deviation caused by the side shot is avoided.
6. The method of pile head footprint calculation according to claim 5, wherein said line detection algorithm is Hough line detection method and said binary model is support vector machine or random forest.
7. The method of claim 1, wherein in the calculation of the length of the head, a plurality of front pictures of the head are collected and the length is calculated, and the average value of the calculated lengths of the plurality of pictures is taken as the length of the head; meanwhile, in the calculation of the width of the pile head, a plurality of pictures of the side face of the pile head are collected, the width is calculated, and the average value of the calculated widths of the plurality of pictures is taken as the width of the pile head.
8. A system for pile head footprint calculation comprising a storage medium and a calculation module, wherein the storage module is adapted to store a program for performing the method of any of claims 1 to 7, and the calculation module is adapted to execute the program stored in the storage module.
9. An apparatus for pile head footprint calculation, comprising a camera for acquiring an image, a memory for storing a program for performing the method of any one of claims 1 to 7, and a processor for executing the program stored in the memory.
CN202211178396.XA 2022-09-27 2022-09-27 Method and system for calculating head-stacking occupied area Active CN115272446B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211178396.XA CN115272446B (en) 2022-09-27 2022-09-27 Method and system for calculating head-stacking occupied area

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211178396.XA CN115272446B (en) 2022-09-27 2022-09-27 Method and system for calculating head-stacking occupied area

Publications (2)

Publication Number Publication Date
CN115272446A CN115272446A (en) 2022-11-01
CN115272446B true CN115272446B (en) 2023-01-24

Family

ID=83757669

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211178396.XA Active CN115272446B (en) 2022-09-27 2022-09-27 Method and system for calculating head-stacking occupied area

Country Status (1)

Country Link
CN (1) CN115272446B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116311219B (en) * 2023-01-29 2024-02-09 广州市玄武无线科技股份有限公司 Ground pile article occupied area calculation method and device based on perspective transformation
CN116721338A (en) * 2023-03-13 2023-09-08 广州市玄武无线科技股份有限公司 Land pile product occupied area calculation method and device based on target detection

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109784172A (en) * 2018-12-14 2019-05-21 上海扩博智能技术有限公司 Ground heap commodity occupied area evaluation method, system, equipment and medium in market

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20080026003A (en) * 2006-09-19 2008-03-24 삼성전자주식회사 Apparatus and method for tagging id in photos by utilizing geographical positions
CN113034574B (en) * 2021-05-25 2021-09-14 广州市玄武无线科技股份有限公司 Commodity ground heap area calculation method and system based on target detection

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109784172A (en) * 2018-12-14 2019-05-21 上海扩博智能技术有限公司 Ground heap commodity occupied area evaluation method, system, equipment and medium in market

Also Published As

Publication number Publication date
CN115272446A (en) 2022-11-01

Similar Documents

Publication Publication Date Title
CN115272446B (en) Method and system for calculating head-stacking occupied area
CN107292248B (en) Commodity management method and system based on image recognition technology
WO2020048492A1 (en) Commodity state identification
CN111695486B (en) High-precision direction signboard target extraction method based on point cloud
CN111310706B (en) Commodity price tag identification method and device, electronic equipment and storage medium
CN110910314B (en) Splicing method and device for shelf scene images
CN111161346B (en) Method and device for layering commodities in goods shelves and electronic equipment
CN115249356B (en) Identification method, device, equipment and storage medium
CN115601672B (en) VR intelligent shop patrol method and device based on deep learning
CN112378333A (en) Method and device for measuring warehoused goods
CN111428743B (en) Commodity identification method, commodity processing device and electronic equipment
CN116188544A (en) Point cloud registration method combining edge features
CN111160450A (en) Fruit and vegetable weighing method based on neural network, storage medium and device
CN110472085B (en) Three-dimensional image searching method, system, computer device and storage medium
CN112614176A (en) Belt conveyor material volume measuring method and device and storage medium
CN116311219B (en) Ground pile article occupied area calculation method and device based on perspective transformation
CN109409387B (en) Acquisition direction determining method and device of image acquisition equipment and electronic equipment
CN113469216B (en) Retail terminal poster identification and integrity judgment method, system and storage medium
CN113298871B (en) Map generation method, positioning method, system thereof, and computer-readable storage medium
CN115830604A (en) Surface single image correction method, device, electronic apparatus, and readable storage medium
CN116935198B (en) Method and device for detecting shortage of goods on goods shelf and robot
CN112801055B (en) Unmanned supermarket positioning and tracking method based on film pressure sensor array
RU2805760C1 (en) Method and device for retail goods identification
CN117124302B (en) Part sorting method and device, electronic equipment and storage medium
JP2014127068A (en) Corresponding point search device, camera posture estimation device and program for these

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