JP2001088912A - Inventory management method and inventory system by image recognition - Google Patents

Abstract

(57) [Summary] An object of the present invention is to propose a method of performing inventory management by taking images and recognizing images of products displayed on actual product display shelves with a plurality of cameras and an inventory system using image recognition. A plurality of cameras 1 for taking images of inventory goods, a camera moving means 2 for moving the camera 1 to a position of a predetermined goods shelf S, a camera control means 50, and a camera selection means 51 for selecting and inputting a camera image. Image processing is performed by performing image processing on the selected and input image to extract contour data, and image extracting means 52 for judging and outputting contour feature data such as a contour shape, and a contour having the same shape as a feature data file 53 in which feature data is registered and stored in advance. The inventory device 5 includes a counting means 54 for counting the number of individual commodities based on the characteristic data and a communication control means 61 for communicating with the in-store communication line 6.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an inventory system such as a POS system, and more particularly, to an inventory management method and an inventory system for actually counting the number of inventory products by image recognition from a plurality of camera images.

[0002]

2. Description of the Related Art A conventional stocktaking method will be described by using a part of a schematic system configuration diagram of an embodiment of a stocktaking system according to the present invention shown in FIG. In stores such as convenience stores and warehouse shops, inventory of merchandise displayed on merchandise shelves and warehouse shelves is performed many times a day. At the time of stocktaking, a clerk visually counts hundreds or more of commodities, one-by-one remaining quantities, and performs stocktaking data entry and input operations on stocktaking forms and portable terminals. The collected inventory data is manually input from the input unit 42 or input to the store server 41 from a portable terminal interface unit (not shown), and updates the inventory data in the inventory data master 41. When the remaining quantity of the inventory data becomes equal to or less than a preset quantity, a purchase status of the corresponding product is generated and additionally registered in the order master. This means that the sales clerk must leave the sales operation when the goods are stocked.

[0003] As described above, daily visual inventory requires a great deal of labor and time. Therefore, in many stores, the purchase quantity (the supply quantity to the product shelf) and the sales quantity data from the POS terminals 3, 3 'are often used. The inventory difference is used as planned inventory data, and simultaneous inventory is performed at the end of the month or at the end of the sales period, using holidays and non-business hours. The difference between the visual inventory quantity and the planned inventory quantity is accounted for as the difference quantity. In this case, there is a problem that it is not possible to know when the difference amount has occurred and it is not possible to timely execute an improvement measure. Construction of an automatic inventory system utilizing surveillance cameras installed in many stores has been desired.

[0004]

SUMMARY OF THE INVENTION In view of the above-mentioned problems and problems, the present invention takes stock and recognizes products displayed on actual product display shelves with a plurality of cameras to manage inventory. The purpose is to propose an inventory system using methods and image recognition.

[0005]

SUMMARY OF THE INVENTION In a stocktaking system comprising a plurality of POS terminals and a store server having a stock data master for managing stock data of the store, which is connected by communication through a store communication line, an image of stock products is taken. A plurality of cameras, a camera moving unit for moving the camera to a position such as a predetermined product shelf, a warehouse shelf, a camera control unit for driving and controlling the camera moving unit, and a camera selecting unit for selectively inputting a plurality of camera images. Image extracting means for applying image processing to an image selected and input by the camera selecting means, extracting contour data, comparing the contour data with the feature data, and judging and outputting contour feature data such as a contour shape; A feature data file pre-registered and stored, a counting means for counting the number of individual products based on the contour feature data, and a communication connection with the in-store communication line. An inventory system comprising a communication control means and an inventory device having a communication control means, wherein an inventory of merchandise display shelves or warehouse shelves in a store is imaged by controlling the movement of a plurality of cameras to a predetermined shelf position, Select and input a captured image of the camera, apply image processing to the selected and input image to extract contour data, compare the contour data with the feature data of the feature data file, and determine and output contour feature data such as a contour shape. Counting the number of individual products based on the contour feature data of the same shape, inputting the count output to the inventory data master in the store server via the in-store communication line, and the store server Update inventory data with product quantity data.

[0006] The image extracting means extracts contour data from a selected and input image using an edge extraction processing function for performing edge enhancement processing of a pre-stored image, and extracts pre-stored shape data from the contour data. Characteristic contour data is extracted using an extraction function, and contour characteristic data is determined and output from the characteristic contour data with reference to characteristic data stored in advance in a characteristic data file.

[0007] From the stacking shape data of the products stacked on the shelf, the quantity of the shaded portion on the image is inferred and output to the counting means using an inference function stored in advance.

[0008] A learning means for learning individual products and characteristic data of the products in an actual stacking state is additionally installed in the stocktaking device, and actual data such as a stacking shape, a stacking quantity, and a stacking level are stored in the stocktaking device. Is input, and the extraction function of the characteristic data is inversely calculated and stored and registered.

[0009] Transmission / reception means for receiving a camera image and transmitting / receiving a camera control signal to / from the camera moving means are additionally installed in the stocktaking apparatus, and the camera or the camera moving means and the stocktaking apparatus are connected by wireless communication.

[0010] The camera moving means comprises a turning mechanism for turning and controlling a pan / tilt of the camera and a zoom driving mechanism for driving and controlling a zoom lens of the camera, and the turning mechanism and the zoom driving mechanism are controlled by the camera controlling means. X-installed on a merchandise display shelf or the like that allows a predetermined camera imaging position to be selected by drive control, and allows the camera to be slidably driven in a horizontal direction (X direction) and a vertical direction (Y direction). A ceiling or wall surface which is a Y movement mechanism, and which is driven and controlled by camera control means, or which can be slidably driven along a guide rail or guideway installed on the ceiling or wall surface of the store. The ceiling or wall moving mechanism is driven and controlled by the camera control means.

[0011] A transmitting / receiving means for transmitting / receiving a camera control signal or the like to / from the camera moving means, a modulation / demodulation unit for modulating or demodulating the camera control signal or the like, and a camera for driving / controlling the camera moving means with the received / demodulated camera control signal. A control unit is additionally installed to enable wireless connection between the camera moving unit and the inventory device.

[0012] The inventory system is provided with a plurality of POS terminals,
In an inventory system composed of a store server having an inventory data master that manages inventory data of stores connected to the in-store communication line, a plurality of cameras for imaging inventory products, and the cameras are moved to predetermined product shelf positions. Camera moving means, and an X-axis for controlling XY movement of the camera moving means.
A Y-movement control unit, a camera selection unit for selectively inputting image signals from a plurality of cameras, an A / D for converting the selectively input image signals into digital image data, and edge extraction necessary for image edge extraction processing A processing function memory that stores processing functions, an image processing unit that performs image processing of image data based on the image processing functions stored in the processing function memory, and a captured image from the processed image data that has undergone image processing A contour extracting unit for extracting contour data of the image, a feature extracting unit for extracting contour feature data such as a shape completed from the extracted contour data, and a feature storing a feature extracting function for processing by the feature extracting unit. Extraction function memory, a feature data file storing feature data such as shape, quantity, stacking shape, and number of stacking stages for each product on the display shelf, and feature data extracted from the feature data A feature data determining unit that determines and outputs accurate contour feature data in comparison with feature data for each individual product stored in advance in a file, and a counting unit that counts the number of contour feature data having the same shape from the contour feature data When,
An inventory product file storing planned inventory product data such as product name, product code, characteristic data, purchase quantity for each shelf number, and a communication control unit for communication connection with a store server via an in-store communication line. The system comprises a system memory storing an operation sequence program of each section in the apparatus, and an inventory control section having a stock control section for issuing a control command to each section in the apparatus in accordance with the operation sequence program.

[0013] A transmitting and receiving unit for wirelessly transmitting and receiving image data and control data to and from a camera, a camera moving unit, and the like, a modem for modulating and demodulating the transmitted and received signals,
A data conversion unit for performing data conversion between internal bus line data and modulation / demodulation data is additionally provided.

As the camera control means of the inventory device, a turning control unit for turning the camera and controlling the zoom lens is additionally provided, or an XY movement amount sensor for detecting the XY movement amount of the camera is additionally provided. .

An input / output terminal for the modem unit is additionally provided in the XY movement control unit.

An input terminal from the modem unit is additionally installed in the camera selection unit.

In the processing function memory, a color boundary processing function for performing color boundary processing of an image is additionally stored, a pattern boundary processing function for performing pattern boundary processing of an image is additionally stored, or
A character recognition processing function for performing image character recognition processing is additionally stored.

The feature extraction function memory additionally stores an overlap extraction function for extracting image overlap, additionally stores a stacking shape extraction function for extracting a stacking shape of a product, and a product for extracting the number of product stacking stages. Additionally memorize the upper row number extraction function,
Alternatively, a random stacking shape extraction function for extracting a stacking shape of randomly stacked products is additionally stored.

As a counting means of the inventory device, a shadow area inference unit for inferring the quantity in the shadow area of the stacked product, and an inference function storing a simple stacking operation function and the like required by the shadow area inference unit Install additional memory.

In the inference function memory, an operation function on a cone volume for inferring a cone shape or a pyramid shape is additionally stored, or a random accumulation operation function for inferring a random accumulation shape is additionally stored.

A feature data registration unit for registering feature data in a feature data file in the inventory device, and feature learning for learning feature data such as a stacking shape and the number of stacking stages from images displayed on the product shelves and inventory product file data. And a learning operation unit for operating the feature learning unit.

A learning operation terminal that wirelessly operates the feature learning unit is additionally installed in the learning operation unit.

The transmitting / receiving unit may be a communication unit using a weak radio wave, an infrared communication unit that modulates and transmits an infrared ray, or a spread spectrum communication unit that performs modulation and demodulation of a microwave in a spread spectrum type.

[0024]

BEST MODE FOR CARRYING OUT THE INVENTION

FIG. 1 is a schematic system configuration diagram of an embodiment of an inventory system according to the present invention. FIG. 2 is a block diagram of a main part of an embodiment of a stocktaking device according to the present invention.
The schematic configuration of the inventory system according to the present invention will be described with reference to FIG. The inventory system according to the embodiment includes a store server 40, an input unit 42, and a display device 43, which are connected by communication with a plurality of POS terminals 3, 3 'and an in-store communication line 6 and have an inventory data master 41 for managing inventory data of the store. In the store POS server system 4 composed of: a plurality of cameras 1, 1 'for taking images of inventory goods, camera moving means 2, 2 for moving the cameras 1, 1' to a position of a predetermined goods shelf S, A camera control means 50 for driving and controlling the moving means 2; a camera selection means 51 for selectively inputting a plurality of camera images; an image selected by the camera selection means 51; Image extracting means 52 for judging and outputting contour feature data such as a contour shape by comparing the feature data with the feature data; a feature data file 53 in which feature data is registered and stored in advance; Counting means 54 for counting the number of individual commodities based on collection data; learning means 5 for learning individual commodities and characteristic data of the commodities in an actual shelf state
6. An inventory product file 57 storing planned inventory product data such as a shelf number, a product name, a product code, characteristic data, and a purchase quantity of each product G displayed on the product shelf S; And a communication control means 61 for communication connection.

The detailed configuration of the inventory device according to the present invention will be described with reference to FIG. The inventory device 5 of the embodiment includes an XY movement control unit 50 that performs XY movement control of the camera movement unit 2.
a, a camera selection section 51a for selectively inputting image signals from a plurality of cameras 1, 1 ', an A / D 52a for converting a selectively input image signal into digital image data, and edge extraction processing required for image edge extraction processing A processing function memory 52c that stores functions and the like; an image processing unit 52b that performs image processing of image data based on an image processing function stored in the processing function memory 52c; A contour extraction unit 52d for extracting the contour data of
A feature extraction unit 52f for extracting contour feature data such as a completed shape from the extracted contour data;
A feature extraction function memory 52e that stores a feature extraction function for the feature extraction process, and a feature that stores feature data such as shape, quantity, stacking shape, and stacking level for each product on the display shelf S. A data file 53, a feature data determination unit 52g that compares the extracted contour feature data with the feature data of each individual product stored in the feature data file 53 in advance and determines and outputs accurate contour feature data; A counting operation unit 54a for counting the number of data, an inventory product file 5 storing planned inventory product data such as a product name, a product code, characteristic data, and a purchase quantity for each shelf number.
7, an operation unit 60 for operating the apparatus, an inventory data input unit 58a for inputting inventory data, and a communication control unit 6 for communication connection with the store server 40 (FIG. 1) via the in-store communication line 6.
1a, an operation unit 60 for inputting operation of the apparatus, a system memory 62 storing an operation sequence program for each unit in the apparatus, a stocktaking control unit 59 for issuing a control command to each unit in the apparatus according to the operation sequence program, It is configured.

Further, a transmitting / receiving unit 65 for wirelessly transmitting / receiving image data and control data to / from the camera 1 and the camera moving unit 2 and the like, and a modem unit 6 for modulating / demodulating the transmitted / received signal.
4. A data converter 63 for converting data between internal bus line data and modulation / demodulation data is additionally provided.

Further, a turning control unit 50b for turning the camera and controlling the zoom lens and an XY movement amount sensor 50c for detecting the XY movement amount of the camera are additionally provided.

Further, a feature data registration unit 55a for registering feature data in a feature data file, and a feature learning unit 56a for learning feature data such as a stacking shape and the number of stacking stages from images displayed on product shelves and inventory product file data. ,
A learning operation unit 56b for operating the feature learning unit 56a is additionally provided. Further, a learning operation terminal 56c that wirelessly operates the feature learning unit 56a is added.

The configuration of the camera moving means will be described with reference to the perspective views of various embodiments of the camera moving means according to the present invention shown in FIG. (A) is a general camera swivel device including a swivel mechanism 22 and zoom lens driving units 23 and 23a. In a large warehouse type store, the camera swivel device is installed on a high ceiling to provide a distant view. The image is taken in the imaging format, the field of view of the corresponding product shelf S is moved by the turning mechanism 22,
By zooming up with the zoom lens driving units 23 and 23a, an image of the product G can be captured.

(B) is a perspective view of the XY moving mechanism. A guide way 24 is installed in both sides of the product display shelf S in the vertical direction (Y direction), and a horizontal (X direction) X guide way 25 slidable in the Y direction from an opening 24a of the guide way. The guide part 25a is looking. The guide portion 25a is driven in the Y direction by a wire 25c via a pulley 25d driven and rotated by a deceleration drive motor 25b. The X guide way 25 has guide portions 25 at both ends.
a and is a mechanism slidable in the Y direction. The X guideway 25 has a cross-sectional shape of a substantially angle C,
From the opening, a camera mount 26 that holds the camera 2 and is slidable in the X direction is viewed. The camera mounting portion 26 has a drive wheel 26a, and is driven and controlled freely in the X direction by a drive mechanism (not shown).

(C) does not show details of a guideway 27 installed on a ceiling or a wall surface, a camera 2 with a zoom lens 23, a camera swivel mechanism 22, and a swivel device including a zoom lens drive unit 23 a. Moving drive mechanism 2
8. The guideway 27 is installed on a ceiling or a wall in a store along a product display shelf. Further, the camera 2 of the embodiment includes a camera control unit 30, a modulation / demodulation unit 31, a transmission / reception unit 32, and an antenna 29
Are additionally installed, and are moved to a predetermined shelf position using the movement drive mechanism 28 based on a radio command from the inventory device, and then the camera turning mechanism 22 and the zoom lens drive unit 2
The field of view of the corresponding product can be set by 3a.

In addition to the explanatory diagram of the inventory counting method according to the present invention shown in FIG. 3, the details of the inventory counting method by image recognition will be described. The inventory counting operation is started by inputting an inventory counting command from the operation unit 60. The inventory control unit 59 issues an operation control command to each unit according to a procedure registered in the system memory 62 in advance. When the product name, shelf number, and the like are read from the inventory product file 57 and sent to the camera control unit 50, the XY movement control unit 50a and the turning control unit 50b of the camera control unit 50 'To output a drive control signal (ST1). X added
-Y movement sensor 50c detects the corresponding position, and
The movement control unit 50a continuously outputs the drive control signal until the fed back XY value matches the commanded shelf position data. When the camera 1 moves to a predetermined position, an image of the product is captured and output (ST2). The camera selection unit 51a selectively inputs an image signal of the camera (ST3). The image signal selected and input is A / D 52a.
Is converted into digital image data (ST4). The image processing unit 52b generates image processing data for image recognition by the image processing unit 52b (ST5). The image processing data is performed by selectively using a processing function stored in the processing function memory 52c (ST51). This processing function includes an edge extraction processing function necessary for image edge extraction processing, a color boundary processing function for performing color boundary processing, a pattern boundary processing function for performing pattern boundary processing,
And a character recognition processing function for performing character recognition processing.

The image processing data is supplied to the contour extraction unit 52
d to extract the outline of the image and output the outline data (ST
6). According to the type of the above-described processing function, the extracted outline data is used to store the outline of the product (edge extraction processing function), the outline of the color boundary (color boundary processing function), and the outline of the pattern boundary of the product packaging box (pattern boundary processing function). ), Character recognition of a character or a bar code printed on the product packaging box (character recognition processing function), and the like are obtained.

Since the contour data is not necessarily completed contour data, the feature extracting unit 52f completes complete contour feature data by using a feature extraction function stored in advance in the feature extraction function memory 52e (ST71). At the same time, contour feature data such as product shape, overlap, stacking shape, and stacking stage number are extracted (ST7). The feature extraction function stored in advance in the feature extraction function memory 52e includes a shape extraction function for extracting the shape of a product, an overlap extraction function for extracting an overlap of images, a stack shape extraction function for extracting a stack shape, and a stack stage. And a random stacking shape extraction function for extracting the stacking shape of randomly stacked products. The selected feature extraction function is a product in the inventory product file 57 for each product. Described as part of the data.

The characteristic data determination unit 52g refers to the extracted contour characteristic data in the characteristic data file 53 in advance by referring to the characteristic data (ST81) and the commodity data (ST82) in the inventory commodity file 57, and counts the contour characteristic data. The judgment is output (ST8).

The counting section 54a calculates and outputs the number of items to be counted from the contour feature data for counting, the number of individual items, the number of stacking stages, and the like (ST9). At this time, depending on the stacked shape, it is not possible to count a portion (shadow area) hidden by the shadow of the image. Shadow area inference unit 54b additionally installed
Selects an inference function stored in the inference function memory 54c (ST91) and performs an inference operation on the total amount including the shadow area. The stored inference functions are a simple accumulation operation function for inferring a simple accumulation shape, a cone volume operation function for inferring a cone shape or a pyramid shape, and a random operation for inferring a random accumulation shape. Such as a stacking operation function.

The calculated and output inventory count is updated and input to the inventory master 41 in the store server 40 via the communication control unit 61a and the in-store communication line 6 while updating the inventory product file 57.

The feature learning method will be described with reference to FIG. 4 which illustrates the feature learning method according to the present invention. Data such as a shelf number, a product name, a product code, the number of stacking stages, and the total number of stacks are input from the learning operation unit 56b (ST10). Most of the data is quoted from the inventory product file 57. Hereinafter, the operation after the outline data is extracted (ST6) will be described because the common part is the same as the above-described flow of the inventory counting operation, so that the overlapping part is omitted. The contour data is input to the feature extracting unit 52f, and contour feature data is extracted (ST7). The feature data registration unit 55a corrects the feature data such as the shape, the overlap, and the stacked shape stored in the feature data file 53 using the contour feature data. By this learning operation, it is possible to constantly perform learning correction on a delicate shape, a stacking shape, a difference in the number of stacking stages, and the like of the product.

The attendant can use the portable learning operation terminal 56c to perform a feature learning operation near the display shelf of the product.

A camera controller 30 for driving and controlling the camera moving unit 2, a modulator / demodulator 31 for modulating and demodulating an image signal of the camera, a camera control signal, and the like, which are additionally installed in the camera installation unit, and wireless communication with the inventory device 5. Transmitting and receiving unit 3
The camera 2 can communicate and connect the camera 1 and the camera moving means 2 that freely move in the store with the stocktaking device 5 without cables.

Further, the transmitting and receiving unit 65 additionally installed in the inventory device 5 includes a camera unit and a portable learning operation terminal 5.
6c for communication connection by wireless means. The modulation and demodulation unit 64 modulates and demodulates control data and image data. The demodulated image data is input to an input terminal additionally provided in the camera selection unit 51a. On the other hand, the control data is subjected to data conversion by the data conversion unit 63 and input / output to / from the internal bus line of the inventory device 5.

[0043]

The present invention is embodied in the form described above, and has the following effects. A plurality of cameras are moved and controlled to a predetermined shelf position to take an image of a stock product on a product display shelf or a warehouse shelf in a store, and a captured image of the camera is selected and input, and image processing is performed on the selected and input image. Extracts the characteristic image data, compares the characteristic image data with the characteristic data in the characteristic data file, determines and outputs shape data such as a contour shape, counts and outputs the number of individual products based on the shape data, and outputs the count output in the store. Via the communication line, the data is input to the inventory data master in the store server, and the store server updates the inventory data with the input quantity data of the individual products, so that the products displayed on the actual product display shelf And a method for inventory management by recognizing and recognizing images with a plurality of cameras, and an inventory system using image recognition.

[Brief description of the drawings]

FIG. 1 is a schematic system configuration diagram of an embodiment of an inventory system according to the present invention.

FIG. 2 is a block diagram of a main part of an embodiment of a stocktaking device according to the present invention.

FIG. 3 is an explanatory diagram of an inventory counting method according to the present invention.

FIG. 4 is an explanatory diagram of a feature learning method according to the present invention.

FIG. 5 is a perspective view of various embodiments of a camera moving unit according to the present invention.

[Explanation of symbols]

 S Product shelf G Product 1, 1 'Camera 2, 2' Camera moving means 3, 3 'POS terminal 5 Inventory device 6 In-store communication line 22 Camera turning mechanism 23 Zoom lens 23a Zoom drive mechanism 24, 27 Guide way 25 X Guide way 26 Camera mounting unit 31 Camera control unit 32, 64 Modulation / demodulation unit 33, 65 Transmission / reception unit 40 Store server 42 Input unit 43 Display unit 50 Camera control unit 50a XY movement control unit 50b Turning control unit 50c XY movement amount sensor 51 Camera selection unit 51a Camera selection unit 52 Image extraction unit 52a A / D 52b Image processing unit 52c Processing function memory 52d Contour extraction unit 52e Feature extraction function memory 52f Feature extraction unit 52g Feature data determination unit 53 Feature data file 54 Counting unit 54a Counting Arithmetic unit 54b Shadow area inference unit 4c Inference function memory 55 Feature registration unit 55a Feature data registration unit 56 Feature learning unit 56a Feature learning unit 56b Learning operation unit 56c Learning operation terminal 57 Inventory product file 58 Inventory data input unit 59 Inventory control unit 60 Operation unit 61 Communication control unit 61a Communication control unit 62 System memory 63 Data conversion unit 70 Internal bus line.

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) G07G 1/12 321 G06F 15/24 5L096 1/14 15/62 380 9A001 H04N 7/18 15/70 450F Terms (Reference) 3E042 BA13 CD01 CD02 EA01 3F022 FF01 MM03 MM08 MM28 PP02 PP04 QQ17 5B049 BB11 CC27 DD00 DD02 EE02 EE03 EE05 EE07 EE12 FF03 FF04 GG01 GG03 GG04 GG07 5B057 AA02 DC03 DC03 BA03 DC03 DC DA07 EA01 EA03 EA05 EB05 EE00 EG08 FC14 FC15 HA01 HA05 5L096 CA05 CA25 DA02 FA06 FA52 HA09 JA11 JA22 KA04 9A001 HH23 HH28 JJ52 JZ58 KK37

Claims (30)

[Claims] 1. An inventory system comprising a plurality of POS terminals and a store server having a stock data master for managing the stock data of the store, the plurality of televisions being used to image stock products. John camera (hereinafter abbreviated as camera); camera moving means for moving the camera to a position such as a predetermined product shelf or warehouse shelf; camera control means for driving and controlling the camera moving means;
A camera selecting means for selecting and inputting the plurality of camera images, applying image processing to the image selected and input by the camera selecting means, extracting contour data, comparing the contour data with feature data, Image extraction means for judging and outputting contour feature data, a feature data file pre-registering and storing the feature data, counting means for counting the number of individual products based on the contour feature data, and the in-store communication line. An inventory system comprising an inventory device having communication control means for communication connection, and an inventory product on a product display shelf or a warehouse shelf in a store,
A plurality of cameras are controlled to move to a predetermined shelf position and imaged,
Selectively input a captured image of the camera, perform image processing on the selected and input image, extract contour data, and compare the contour data with the feature data of the feature data file to determine contour feature data such as a contour shape. Output, count and output the number of individual products based on the contour feature data of the same shape, input the count output to the inventory data master in the store server via the in-store communication line, Updating the inventory data with the input quantity data of the individual product, wherein the inventory management method by image recognition. 2. The image extraction means extracts edge data from the selected and input image using an edge extraction processing function for performing edge enhancement processing of an image stored in advance, and stores the edge data in advance from the edge data. Extracting characteristic contour data using a shape data extraction function or the like, and determining and outputting contour characteristic data by referring to characteristic data stored in advance in a characteristic data file from the characteristic contour data. 2. The inventory management method according to claim 1, wherein the inventory is managed by image recognition. 3. The method according to claim 1, wherein a quantity of a shadowed portion on the image is inferred from the stacking shape data of the goods stacked on the shelf to the counting means by an inference function stored in advance. 2. The inventory management method according to claim 1, wherein the inventory is managed by image recognition. 4. A learning means for learning individual commodities and characteristic data of the commodities in an actual shelving state is additionally installed in the stocktaking apparatus, and the stocktaking apparatus has a stacking shape, a stacking quantity, a stacking number, and the like. 2. The inventory management method using image recognition according to claim 1, wherein the actual data is input, and an extraction function of the characteristic data is inversely calculated and stored and registered. 5. The inventory device further includes transmission / reception means for receiving the camera image and transmitting / receiving a camera control signal to / from the camera movement means, and wirelessly communicates the camera or the camera movement means with the inventory device. 2. The inventory management method using image recognition according to claim 1, wherein communication is established. 6. The camera moving means comprises a turning mechanism for turning and controlling a pan / tilt of the camera and a zoom driving mechanism for driving and controlling a zoom lens of the camera, wherein the turning mechanism and the zoom driving mechanism are provided. 2. The inventory management method according to claim 1, wherein a predetermined camera imaging position can be selected by drive control by the camera control unit. 7. The camera moving means is installed on the product shelf, warehouse shelf, or the like, and the camera is moved in a horizontal direction (X direction).
And an XY moving mechanism capable of slidably driving in the vertical direction (Y direction), and by controlling the XY moving mechanism by the camera control means, a predetermined camera imaging position can be selected. 2. The inventory management method according to claim 1, wherein the inventory is managed by image recognition. 8. The camera moving means is a ceiling or wall moving mechanism which can be slidably driven along a guide rail or a guideway installed on a ceiling or a wall of a store. 2. The inventory management method according to claim 1, wherein a predetermined camera imaging position can be selected by drive control by the camera control unit. 9. A transmitting / receiving means for transmitting / receiving a camera control signal to / from the camera moving means, a modulation / demodulation unit for modulating or demodulating the camera control signal, and a drive control of the camera moving means by the received / demodulated camera control signal. 2. The inventory management method according to claim 1, wherein a camera control unit is additionally installed, and the camera moving unit and the inventory device can be wirelessly connected. 10. An inventory system comprising a plurality of POS terminals and a store server having an inventory data master for managing inventory data of the store, which is communicatively connected by an in-store communication line, and wherein a plurality of cameras for imaging inventory products are provided. A camera moving unit for moving the camera to a predetermined shelf position; an XY movement control unit for performing XY movement control of the camera moving unit; and selectively inputting image signals from the plurality of cameras. A / D (Analog Digit) for converting a selected and input image signal into digital image data.
al converter), a processing function memory storing edge extraction processing functions required for image edge extraction processing, and performing image processing of the image data based on the image processing functions stored in the processing function memory. An image processing unit, a contour extraction unit that extracts contour data of a captured image from the processed image data that has been subjected to the image processing, and a feature extraction unit that extracts contour feature data such as a complete shape from the extracted contour data. A feature extraction function memory storing a feature extraction function for processing by the feature extraction unit, and storing feature data such as a shape, a quantity, a stacking shape, and a stacking level for each product on the display shelf. Feature data for comparing the extracted feature data with the feature data for each individual product stored in the feature data file in advance to determine and output accurate feature data. A determination unit, a counting operation unit that counts the number of contour feature data of the same shape from the contour feature data, and stores planned inventory product data such as a product name, a product code, feature data, and a purchase quantity for each shelf number. An inventory product file, a communication control unit that communicates with the store server via the in-store communication line, a system memory storing an operation sequence program for each unit in the device, And an inventory control unit for issuing a control command to each unit in the apparatus. 11. A transmission / reception unit for wirelessly transmitting / receiving image data and control data to / from the camera, the camera moving means, etc. to the inventory device, a modulation / demodulation unit for modulating / demodulating the transmission / reception signal, and data in an internal bus line. The inventory system according to claim 10, further comprising a data conversion unit for performing data conversion with modulation / demodulation data. 12. The stocktaking system according to claim 10, wherein a turning control unit for turning a camera and controlling a zoom lens is additionally provided as a camera control means of the stocktaking device. 13. The inventory system according to claim 10, wherein an XY movement amount sensor for detecting an XY movement amount of a camera is additionally provided as a camera control unit of the inventory device. 14. The inventory system according to claim 10, wherein an input / output terminal for the modulation / demodulation unit is additionally provided in the XY movement control unit. 15. The apparatus according to claim 1, wherein an input terminal from the modulation / demodulation unit is additionally provided in the camera selection unit.
An inventory system according to 0 or 11. 16. The inventory system according to claim 10, wherein a color boundary processing function for performing color boundary processing of an image is additionally stored in the processing function memory. 17. The inventory system according to claim 10, wherein a pattern boundary processing function for performing pattern boundary processing of an image is additionally stored in the processing function memory. 18. The inventory system according to claim 10, wherein a character recognition processing function for performing character recognition processing of an image is additionally stored in the processing function memory. 19. The inventory system according to claim 10, wherein an overlap extraction function for extracting an overlap of images is additionally stored in the feature extraction function memory. 20. The inventory system according to claim 10, wherein a stacking shape extraction function for extracting a stacking shape of a product is additionally stored in the feature extraction function memory. 21. The inventory system according to claim 10, wherein a stacking stage number extraction function for extracting a stacking stage number of a product is additionally stored in the feature extraction function memory. 22. The feature extraction function memory further stores a random stacking shape extraction function for extracting a stacking shape of randomly stacked products.
The inventory system according to item 0. 23. As a counting means of the stocktaking device, a shadow area inference unit for inferring a quantity in a shadow area of a stacked product, a simple stacking operation function required by the shadow area inference unit, and the like are stored. 11. The stocktaking system according to claim 10, further comprising an additional inference function memory. 24. The inventory system according to claim 23, wherein an operation function on a cone volume for performing a productive inference of a cone shape or a pyramid shape is additionally stored in the inference function memory. 25. The inventory system according to claim 23, wherein a random stacking operation function for performing stacking inference of a random stacking shape is additionally stored in the inference function memory. 26. A feature data registration unit for registering feature data in the feature data file in the stocktaking device, and feature data such as a stacking shape and the number of stacking stages based on an image displayed on a product shelf and the stocktaking product file data. 12. A feature learning unit for learning the feature learning and a learning operation unit for operating the feature learning unit are additionally installed.
Inventory system described in 1. 27. The inventory system according to claim 26, wherein a learning operation terminal for operating the feature learning unit wirelessly is additionally installed in the learning operation unit. 28. The inventory system according to claim 11, wherein the transmission / reception unit is a communication unit using weak radio waves. 29. The inventory system according to claim 11, wherein said transmission / reception unit is infrared communication means for modulating and transmitting infrared light. 30. The inventory system according to claim 11, wherein said transmission / reception unit is a spread spectrum communication means for performing spread spectrum modulation / demodulation of microwaves.