Inventory system, program for inventory and inventory device

Abstract

To provide an inventory system for efficiently performing inventory work about products piled up in an inventory storage place.SOLUTION: An inventory system photographs respective mountains in an inventory storage place at prescribed intervals, acquires photographed images recorded in association with current positions when photographed, generates an ortho-image about the entire inventory storage place on the basis of the photographed images, extracts the mountains from the ortho-image on the basis of a learned model, attaches identification information to each extracted mountain, generates point group data related to the inventory storage place on the basis of data for point group data acquired by an unmanned mobile body, extracts point group data of each mountain in the point group data about the inventory storage place from each mountain, specifies products piled up in each mountain, specifies the number of the products from a shape of the point group data of each extracted mountain with reference to product master information storing information of the size of the products and other product related information in advance, and outputs the number of the products of each mountain and the number of the products in the entire inventory storage place.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to an inventory system, an inventory program, and an inventory device for carrying out an inventory of products piled up in an inventory area. [Background technology]

[0002] Conventionally, product inventory work is a task that companies must perform periodically, and various efforts have been made to improve the efficiency of inventory work.

[0003] For example, Patent Document 1 discloses a technology that uses a depth sensor installed on the ceiling or the like in a warehouse to measure the distance to placed products and calculates the number of products in stock based on the measured distance. Also, Patent Document 2 discloses a technology that enables a self-propelled robot to automatically grasp the inventory status by reading RFID tags attached to books and other items under management using an RFID reading antenna provided on the robot when the robot is made to travel a predetermined route. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2022-088305 [Patent Document 2] Japanese Patent Application Publication No. 2019-189380 Summary of the Invention [Problem to be solved by the invention]

[0005] Incidentally, some civil engineering products, such as concrete products like U-shaped gutters, and building materials like concrete blocks, are stacked outdoors for inventory management. Unlike warehouses where inventory can be neatly arranged using shelves, these products are often stacked in irregular locations on the premises when managed outdoors. Because these products are managed outdoors, it is difficult to grasp inventory quantities based on the distance from the ceiling, as in Patent Document 1, and the environment is not suitable for grasping inventory by using a self-propelled robot to travel a predetermined route, as in Patent Document 2. Even indoors, it is difficult to apply Patent Document 1 to warehouses with uneven ceilings, and it is also difficult to apply Patent Document 2 to inventory stacked irregularly in indoor warehouses.

[0006] The present invention has been made in consideration of the above problems, and has as its object to provide an inventory system, an inventory program, and an inventory device for efficiently carrying out inventory work on products piled up in an inventory area. [Means for solving the problem]

[0007] The inventory system according to the present invention includes an unmanned mobile body equipped with a camera device and capable of automatic control, and a server device capable of acquiring images taken by the unmanned mobile body via a communication network, and is an inventory system for automatically counting inventory for inventory purposes in an inventory area where one or more piles of at least one type of product are present, wherein the server device comprises: an image acquisition means for acquiring images taken by the unmanned mobile body equipped with a means for recording a current position, the image acquisition means acquiring images of each pile in the inventory area at predetermined intervals and correlating the current position at the time of taking the images; an orthoimage generation means for generating an orthoimage of the entire inventory area based on the captured images; and a server device for extracting piles of products from the orthoimage based on a trained model that has been trained to extract the piles of products from the orthoimage. the number of products piled up from the shape of the extracted point cloud data for each pile by referring to product master information that stores product size information and other product-related information in advance; and inventory quantity output means that outputs the number of products for each pile with identification information and the number of products in the entire inventory area.

[0008] Furthermore, in the inventory system of the present invention, the data for generating point cloud data is the photographed images, and the point cloud data generation means generates point cloud data based on the photographed images taken at two or more different locations, and combines the multiple point cloud data to generate point cloud data related to the inventory storage area.

[0009] Furthermore, in the inventory system of the present invention, the data for generating point cloud data is point cloud data acquired by a LiDAR equipped on the unmanned mobile body, and the point cloud data generation means combines multiple point cloud data acquired by the LiDAR to generate point cloud data related to the inventory storage area.

[0010] Furthermore, in the inventory system according to the present invention, a unique identification symbol or identification marker is placed near each pile in the inventory area or on the top surface of each pile, and the product master information has product information corresponding to the unique identification symbol or identification marker registered in advance, and the quantity determination means refers to the product master information based on the unique identification symbol or identification marker to identify the products for each pile and obtain the product-related information.

[0011] The inventory program of the present invention is an inventory program that causes a computer to execute a process for automatically counting stock for inventory purposes in an inventory area where one or more piles of at least one type of product exist, and the computer is provided with an image acquisition function that acquires images of each pile in the inventory area at predetermined intervals taken by an unmanned mobile object that is equipped with a camera device, is capable of automatic control, and has a means for recording its current position, and that associates and records the images with the current position at the time of taking the images; an orthoimage generation function that generates an orthoimage of the entire inventory area based on the captured images; and a function that extracts the piles from the orthoimage based on a trained model that has been trained to extract the piles of products from the orthoimage. the number of products piled up from the shape of the extracted point cloud data for each pile by referring to product master information that stores product size information and other product-related information in advance; and an inventory quantity output function that outputs the number of products for each pile with identification information and the number of products in the entire inventory area.

[0012] The inventory device according to the present invention is an inventory device for automatically counting stock for inventory purposes in a stockroom where there is one or more piles of at least one type of product, and includes an image acquisition unit that acquires images taken by an unmanned mobile body that is equipped with a camera device, is capable of automatic control, and has a means for recording its current position, and that acquires images of each pile in the stockroom at predetermined intervals and records the images in association with the current position at the time of taking the images; an orthoimage generation unit that generates an orthoimage of the entire stockroom based on the captured images; and a trained model that has been trained to extract piles of products from the orthoimage, and performs a pile detection process for each extracted pile. a point cloud data generation unit that generates point cloud data related to the inventory area based on point cloud data generation data acquired by the unmanned mobile body; a point cloud data extraction unit that extracts point cloud data for each pile to which the identification number is assigned in the point cloud data related to the inventory area, a quantity determination unit that identifies products stacked in each pile to which the identification number is assigned and determines the number of products stacked from the shape of the extracted point cloud data for each pile by referring to product master information that stores product size information and other product-related information in advance; and an inventory quantity output unit that outputs the number of products in each pile to which the identification information is assigned and the number of products in the entire inventory area. [Effects of the Invention]

[0013] According to the present invention, point cloud data is generated based on images captured by an unmanned mobile body, and the point cloud data is used to count the number of products in stock, making it possible to efficiently carry out inventory work even for products piled up in a stockroom. [Brief explanation of the drawings]

[0014] [Figure 1] 1 is a system configuration diagram showing an example of an inventory system according to an embodiment of the present invention. [Figure 2]2 is a block diagram showing an example of the configuration of a server device in an inventory system according to an embodiment of the present invention. FIG. [Figure 3] An image diagram showing an example of a flight route of an unmanned mobile body in an inventory system according to one embodiment of the present invention, and an example of an orthoimage generated based on images taken by the unmanned mobile body. [Figure 4] FIG. 1 is an image diagram showing an enlarged portion of an orthoimage generated in an inventory system according to an embodiment of the present invention. [Figure 5] 1 is an image diagram showing an example of an image captured by an unmanned moving body in an inventory system according to an embodiment of the present invention. FIG. [Figure 6] 10 is an image diagram showing another example of an image captured by an unmanned vehicle in an inventory system according to an embodiment of the present invention. FIG. [Figure 7] 1 is a table showing an example of inventory information output in an inventory system according to an embodiment of the present invention; [Figure 8] 1 is a flowchart illustrating an example of the flow of inventory processing executed by a server device of an inventory system according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0015] An embodiment of the present invention will now be described with reference to the drawings.

[0016] FIG. 1 is a block diagram showing an example of the configuration of an inventory system 100 according to the present invention. As shown in FIG. 1, the inventory system 100 includes at least a server device 10, user terminals 201-20n, and an unmanned mobile object 30. The server device 10 and the user terminals 201-20n (including cases where the term "user terminal 20" is used in situations where it is not necessary to specify the terminal) are connected to each other via a communication network 40 such as the Internet. Furthermore, data such as captured images acquired by the unmanned mobile object 30 may be configured to be acquired by the user terminal 20 via a recording medium and then transmitted to the server device 10. Alternatively, the server device 10 and the unmanned mobile object 30 may be configured to be connectable via the communication network 40, and data such as captured images may be transmitted directly from the unmanned mobile object 30 to the server device 10.

[0017] The server device 10 is a device that executes inventory processing based on the acquired photographed images and also appropriately communicates with the user terminals 201 to 20n. The functional configuration of the server device 10 will be described in detail later.

[0018] The user terminals 201 to 20n are terminals that can communicate with the server device 10 and are used by users who use the inventory system.

[0019] The unmanned mobile body 30 is a mobile body equipped with a camera device, capable of automatic control, and equipped with a means for recording its current location. The unmanned mobile body 30 may be any type capable of photographing products to be inventoryed from above, but an unmanned aerial vehicle, or so-called drone, may be used, for example. The unmanned mobile body 30 may be configured to be operated by an operator to move and photograph, or may be configured to automatically move along a predetermined route and photograph appropriately while traveling along the set route. The unmanned mobile body 30 may also be equipped with a GPS device, for example, as a means for determining its current location, and may appropriately acquire latitude and longitude information to recognize the route and identify the location at the time of photographing. The unmanned mobile body 30 does not necessarily have to be a single unit; multiple units may share the responsibility of photographing.

[0020] Furthermore, the unmanned mobile body 30 may employ LiDAR (Light Detection and Ranging) as another example of a means for generating point cloud data, which will be described later. Alternatively, any other means may be employed as long as it is capable of generating point cloud data, and a ToF camera (Time of Flight camera) or the like may also be employed.

[0021] The server device 10 is assumed to be equipped with a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), memory, a hard disk drive, and other storage devices that would normally be equipped in a general computer or server device (not shown). Needless to say, various processes are executed by programs to make these general computers and server devices function as the server device 10 and user terminal 20 of this example.

[0022] The server device 10 and the user terminal 20 use the data read into the memory as needed to execute information processing required for various processes using a CPU, and store the obtained processing results in a storage device as needed. The storage device functions as a storage medium for storing various information. The configuration of the storage device is not particularly limited, but examples include an HDD and an SSD.

[0023] In addition, in the example of Figure 1, the server device 10 is used as an example for explanation, but as a minimum configuration, a standalone computer terminal may be configured to integrate functions similar to those of the server device 10 and function as an inventory device.

[0024] 2 is a block diagram showing an example of the configuration of a server device in an inventory system according to an embodiment of the present invention. As shown in FIG. 2, server device 10 includes at least a photographed image acquisition unit 11, an orthoimage generation unit 12, an identification information assignment unit 13, a point cloud data generation unit 14, a point cloud data extraction unit 15, a quantity specification unit 16, an inventory quantity output unit 17, and a storage unit 18.

[0025] The captured image acquisition unit 11 has a function of acquiring captured images taken by an unmanned mobile body equipped with a means for recording its current position, capturing images of each pile in the inventory storage area at predetermined intervals, and recording the captured images in association with the current position at the time of capturing. For example, while autonomously flying the unmanned mobile body 30 along a preset travel route, the unit captures images at predetermined intervals, and associates and registers the position information at the time of capturing the images.

[0026] The orthoimage generation unit 12 has the function of generating an orthoimage of the entire inventory storage area based on the captured images. This orthoimage generation unit 12 generates an orthoimage that looks as if it were captured from directly above based on multiple captured images taken by the unmanned mobile vehicle 30. Each captured image is registered in association with location information at the time of capture, so by using this location information, location information can also be associated with the orthoimage. In other words, location information is associated so that coordinate information of any point on the orthoimage can be obtained.

[0027] The identification information assigning unit 13 has a function of extracting piles of products from an orthoimage based on a trained model trained to extract piles of products from an orthoimage and assigning identification information to each extracted pile to distinguish the pile. Although the orthoimage shows the entire inventory area, the identification information assigning unit 13 extracts a group of piles of products from the inventory area, distinguishing them from other piles. Any method can be used as long as it can accurately distinguish between piles. For example, a trained model may be used. That is, a trained model for pile extraction may be prepared by pre-training the process of distinguishing and extracting each pile from an input orthoimage and outputting it, and the piles may be extracted using the trained model for pile separation. The identification information assigning unit 13 then assigns identification information to each extracted pile. Here, identification information refers to information for identifying the pile to be targeted in the inventory process.

[0028] The point cloud data generation unit 14 has a function of generating point cloud data related to the inventory area based on point cloud data generation data acquired by the unmanned mobile object. Here, the point cloud data generation data refers to a plurality of captured images captured by the unmanned mobile object 30 in this example. The point cloud data generation unit 14 may generate point cloud data based on two or more captured images (i.e., captured images captured at two or more different locations) captured at a predetermined interval, for example, by stereo matching to extract parallax information to each point, and then calculate distance information to each point based on the principles of triangulation. The generated plurality of point cloud data are then combined to generate point cloud data related to the entire inventory area. Note that if there are areas in the inventory area where the mountains are distributed far apart, the point cloud data may be generated in two or more parts.

[0029] The data for generating point cloud data may be point cloud data acquired by a LiDAR equipped on the unmanned vehicle 30. Since the point cloud data acquired by the LiDAR can be used as is, multiple acquired point cloud data may be combined to generate point cloud data for the entire inventory area.

[0030] The point cloud data extraction unit 15 has a function of extracting point cloud data for each pile assigned an identification number from the point cloud data for the entire inventory area. Since the point cloud data for the entire inventory area generated by the point cloud data generation unit 14 does not distinguish between piles, point cloud data for each pile assigned an identification number is extracted. During extraction, since the position information for each pile can be identified from the orthoimage, point cloud data for a position that matches the identified position information is extracted as point cloud data corresponding to that pile.

[0031] The quantity determination unit 16 has the function of identifying the products stacked for each pile assigned an identification number, and by referencing product master information that stores product size information and other product-related information in advance, determining the number of products stacked from the shape of the extracted point cloud data for each pile. The quantity determination unit 16 first identifies the stacked products. While any means for identifying the products is acceptable, in this example, identification symbols or identification markers are considered. Unique identification symbols or identification markers are placed near each pile in the inventory area or on the products within the pile so that they appear in the captured image. Furthermore, information that can identify the products stacked in each pile with the identification symbol or identification marker is registered in association with the identification symbol or identification marker. This allows information identifying the products stacked in that pile to be obtained from the identification marker or identification symbol appearing in the captured image.

[0032] Next, the quantity determination unit 16 references the product master information based on the identified product information and reads out the product size information and other product-related information. Among the product-related information, particularly important information includes product dimension information and shape information. By knowing the width, depth, and height values ​​of the product, it is possible to determine how many products exist in the point cloud data of the pile. Furthermore, some products may be stacked so that a portion of each product overlaps the product below. For such products, it is necessary to determine the product height by subtracting the overlapping dimension, and this value may be recorded in the product master information.

[0033] After acquiring the product size information, the quantity specification unit 16 specifies the number of products stacked from the shape of the point cloud data for each extracted pile. The number of products included in the point cloud data for the extracted pile is specified by specifying how many products are lined up horizontally and how many are stacked vertically from the width, depth, and height values ​​of each part of the point cloud data for each extracted pile. The number of products specified for each pile in this way is registered by associating the product name with the specified number of products in identification information for identifying the pile. This process of specifying the number is performed for all piles.

[0034] The inventory quantity output unit 17 has the function of outputting the number of products for each pile with identification information and the number of products in the entire inventory area. Because the type and number of products for each pile are identified in the quantity identification unit 16, it is possible to create inventory information for each pile, and by aggregating this information for the entire inventory area, it is possible to output the type and number of products for the entire inventory area.

[0035] The storage unit 16 has a function of storing data necessary for executing various processes performed in the server device 10 and data obtained as a result of the processes. The storage unit 16 may also be configured to store a trained model for peak extraction.

[0036] Note that some of the functions realized by the configuration of the server 10 may be realized in the user terminal 20 instead of the server 10.

[0037] Figure 3 is an image diagram showing an example of an unmanned mobile body's flight route in an inventory system according to an embodiment of the present invention, and an example of an orthoimage generated based on images taken by the unmanned mobile body. As shown in Figure 3, a movement route is set in advance so that all piles of products can be photographed in an inventory storage area where there are multiple piles of products, and photography is performed at predetermined intervals while automatically moving along the set movement route. Then, an orthoimage such as that shown in Figure 3 is generated based on the photographed images taken at the predetermined intervals.

[0038] Figure 4 is an image diagram showing an enlarged portion of an orthoimage generated in an inventory system according to an embodiment of the present invention. The image shown in Figure 4 is an enlarged image of the dashed line portion in the orthoimage of Figure 3. There are four mountains in Figure 4. The identification information assigning unit 13 extracts the four mountains based on the trained model for mountain extraction, and assigns identification information 01 to 04, for example, to each mountain and registers them.

[0039] Figure 5 is an image diagram showing an example of a captured image taken by an unmanned mobile body in an inventory system according to an embodiment of the present invention. The captured image shown in Figure 5 is an image showing the pile of identification information 01 in Figure 4, but also shows the pile of identification information 02 and the pile of identification information 03. As shown in the captured image shown in Figure 5, the number of products stacked varies from pile to pile, so point cloud data extraction unit 15 extracts point cloud data for each pile from the point cloud data of the entire inventory area generated by point cloud data generation unit 14, and identifies the number of products in each pile.

[0040] Figure 6 is an image diagram showing another example of an image captured by an unmanned moving body in an inventory system according to an embodiment of the present invention. The image shown in Figure 6 is an image of a pile of products of a different type from that shown in Figure 5. As such, since the stacking method differs for each product, when performing processing to identify the number of products from point cloud data, it is necessary to take into account the differences in the stacking method for each product recorded in the product master information.

[0041] Figure 7 is a table showing an example of inventory information output by an inventory system according to an embodiment of the present invention. The example in Figure 7 shows (a) output information showing the type and quantity of each product for the four piles shown in Figure 4, and (b) output information about the total inventory quantity for the four piles. By summing up the output information shown in Figure 7 for the entire inventory area shown in Figure 3, output information for the entire inventory area can be obtained.

[0042] FIG. 8 is a flowchart illustrating an example of the flow of inventory processing executed by a server device of an inventory system according to an embodiment of the present invention. As shown in FIG. 8, the inventory processing begins by acquiring a photographed image in the server device 10 (step S101). Next, the server device 10 generates an orthoimage based on the photographed image (step S102). Next, the server device 10 extracts piles of products from the orthoimage, assigns identification information to each pile, and registers them in association with each other (step S103). Next, the server device 10 generates point cloud data for the entire inventory area (step S104). Next, the server device 10 extracts point cloud data for each pile from the point cloud data for the entire inventory area (step S105). Next, the server device 10 identifies the products stacked in each pile with the assigned identification information, acquires product size information and other product-related information by referring to product master information, and identifies the number of products stacked in the point cloud data using the product-related information (step S106). Then, the server device 10 outputs the inventory information for each pile and the inventory information for the entire stock area, which is the sum of all the piles (step S107), and ends the inventory process.

[0043] As described above, according to the inventory system of the present invention, an unmanned mobile vehicle is used to photograph each pile in the inventory area at predetermined intervals, and the photographed images are recorded in association with the current position at the time of photographing. An orthoimage of the entire inventory area is generated based on the photographed images. Piles are extracted from the orthoimage based on a trained model that has been trained to extract piles of products from the orthoimage. Identification information for distinguishing the piles is assigned to each extracted pile. Point cloud data related to the inventory area is generated based on the point cloud data generation data acquired by the unmanned mobile vehicle. For each pile, corresponding point cloud data for the pile is extracted from the point cloud data related to the inventory area. For each pile, the products piled up are identified. Product master information that stores product size information and other product-related information in advance is referenced, and the number of products piled up is identified from the shape of the extracted point cloud data for each pile. The number of products in each pile and the number of products in the entire inventory area are output. This makes it possible to perform inventory work efficiently even with products piled up outdoors.

[0044] In other words, even if the situation is not necessarily regular, such as each product being stacked differently or the piles being positioned irregularly, if the unmanned vehicle can capture images of the entire inventory area evenly, point cloud data can be generated from the captured images, and the type and number of products can be identified and counted from the point cloud data, making it possible to perform efficient inventory work even in inventory areas with irregular conditions.

[0045] In the above embodiment, the case of an outdoor inventory storage area has been described as an example, but the present invention is not limited to this. The inventory system of the above embodiment can also be applied indoors as long as the environment allows movement of the unmanned mobile object 30.

[0046] In the above embodiment, the orthoimage generation unit 12 generates an orthoimage by stitching together multiple captured images taken by the unmanned vehicle 30 based on the position information associated with each captured image. However, this is not limited to this. For example, the point cloud data generation unit 14 may generate point cloud data for the entire inventory area, and then generate an orthoimage based on a 3D model generated by mapping the point cloud data into a 3D virtual space. Specifically, the 3D model obtained by mapping the point cloud data allows immediate identification of the coordinate information of the four corners of each mountain when viewed from above. Therefore, it is conceivable to generate an orthoimage by stitching together corresponding captured images based on this information. Furthermore, the captured images may be applied as textures to the 3D model obtained by mapping the point cloud data, thereby generating an image equivalent to the orthoimage when viewed from directly above. In other words, generating an orthoimage based on captured images includes not only directly stitching captured images together, but also generating point cloud data and using the point cloud data to generate an orthoimage. [Explanation of symbols]

[0047] 100 Inventory System 10 Server device 11. Image acquisition unit 12 Orthoimage generation unit 13 Identification information assignment unit 14 Point cloud data generation unit 15 Point cloud data extraction section 16 Quantity specification section 17 Stock quantity output section 18 Memory section 20, 201~20n User terminals 30 Unmanned Mobile Vehicles 40 Communication Network

Claims

1. An inventory system for automatically counting inventory for inventory purposes in an inventory storage area where there is one or more piles of at least one type of product, comprising: an unmanned mobile body equipped with a camera device and capable of automatic control; and a server device capable of acquiring images taken by the unmanned mobile body via a communication network; The server device an image capturing means for capturing images captured by the unmanned vehicle equipped with a means for recording a current position, the image capturing means capturing images of each pile in the inventory storage area at predetermined intervals and capturing the images in association with the current position at the time of capturing the images; an orthoimage generating means for generating an orthoimage of the entire inventory storage area based on the captured image; An identification information assigning means for extracting mountains from the orthoimage based on a trained model trained on extracting mountains of products from the orthoimage and assigning identification information to each extracted mountain for distinguishing the mountain; a point cloud data generating means for generating point cloud data relating to the inventory site based on point cloud data generation data acquired by the unmanned mobile body; a point cloud data extraction means for extracting point cloud data for each pile to which the identification number is assigned from the point cloud data relating to the inventory area; a number specifying means for specifying the products stacked for each pile to which the identification number is assigned, and for specifying the number of products stacked from the shape of the extracted point cloud data for each pile by referring to product master information that stores product size information and other product-related information in advance; an inventory quantity output means for outputting the number of products for each pile with identification information attached and the number of products in the entire inventory area; An inventory system comprising:

2. The point cloud data generation data is the captured image, The point cloud data generating means generates point cloud data based on the images captured at two or more different points, and combines the plurality of point cloud data to generate point cloud data relating to the inventory storage area. The inventory system according to claim 1.

3. The point cloud data generation data is point cloud data acquired by a LiDAR equipped in the unmanned moving body, The point cloud data generating means combines a plurality of point cloud data acquired by LiDAR to generate point cloud data relating to the inventory storage area. The inventory system according to claim 1.

4. A unique identification symbol or identification marker is placed near each pile or on the top surface of each pile in the inventory storage area; The product master information is pre-registered product information corresponding to the unique identification symbol or identification marker, The number specifying means specifies the products for each pile and acquires the product-related information by referring to the product master information based on the unique identification symbol or identification marker. The inventory system according to claim 1.

5. An inventory program for causing a computer to perform a process for automatically counting inventory for inventory purposes in an inventory storage area where one or more piles of at least one type of product exist, the process comprising: The computer, an image acquisition function for acquiring images taken by an unmanned mobile body equipped with a camera device, capable of automatic control, and equipped with a means for recording its current position, the image acquisition function acquiring images of each pile in the inventory storage area at predetermined intervals and recording the images in association with the current position at the time of the image acquisition; an orthoimage generation function for generating an orthoimage of the entire inventory storage area based on the captured image; An identification information assignment function that extracts mountains from the orthoimage based on a trained model that has been trained to extract mountains on which products are piled from the orthoimage and assigns identification information to each extracted mountain for distinguishing the mountain; a point cloud data generation function for generating point cloud data relating to the inventory storage area based on point cloud data generation data acquired by the unmanned mobile object; a point cloud data extraction means for extracting point cloud data for each pile to which the identification number is assigned from the point cloud data relating to the inventory area; a number specifying function that specifies the products stacked for each pile to which the identification number is assigned, and that specifies the number of products stacked from the shape of the extracted point cloud data for each pile by referring to product master information that stores product size information and other product-related information in advance; an inventory output function that outputs the number of products in each pile with identification information and the number of products in the entire inventory area; An inventory program characterized by realizing the above.

6. An inventory device for automatically counting inventory for inventory purposes in an inventory storage area where one or more piles of at least one type of product exist, an image acquisition unit that acquires images taken by an unmanned mobile body that is equipped with a camera device, can be automatically controlled, and has a means for recording its current position, and that acquires images of each pile in the inventory storage area at predetermined intervals and records the images in association with the current position at the time of the image capture; an orthoimage generating unit that generates an orthoimage of the entire inventory storage area based on the captured image; An identification information assigning unit that extracts mountains from the orthoimage based on a trained model that has been trained to extract mountains of products from the orthoimage and assigns identification information to each extracted mountain for distinguishing the mountain; a point cloud data generation unit that generates point cloud data related to the inventory storage area based on point cloud data generation data acquired by the unmanned mobile object; a point cloud data extraction means for extracting point cloud data for each pile to which the identification number is assigned from the point cloud data relating to the inventory area; a number specifying unit that specifies the products stacked for each pile to which the identification number is assigned, and that specifies the number of products stacked from the shape of the extracted point cloud data for each pile by referring to product master information that stores product size information and other product-related information in advance; an inventory quantity output unit that outputs the number of products for each pile with identification information and the number of products in the entire inventory area; An inventory device comprising:

Images