Search

JP-7854960-B2 - Imaging device and program

JP7854960B2JP 7854960 B2JP7854960 B2JP 7854960B2JP-7854960-B2

Inventors

  • 内藤 英浩

Assignees

  • 東芝テック株式会社

Dates

Publication Date
20260507
Application Date
20230508

Claims (6)

  1. Imaging unit, A storage means for storing multiple images captured by the imaging unit, A decoding means that generates decoded information by decoding the symbols attached to the products contained in the image stored by the storage means , The aforementioned storage means stores multiple images, and the association means associates the decoded information with an object image containing a product enclosed in a specified region . An output means that outputs association information associated by the association means, A transmission means for transmitting either product identification information contained in the decoded information or product identification information identified by general object recognition performed on an image captured by the imaging unit to a product sales data processing device, An imaging device equipped with [a specific feature].
  2. The system further comprises an extraction means for extracting the object image from the image captured by the imaging unit, The association means associates the object image with the decoded information. The imaging apparatus according to claim 1.
  3. The extraction means extracts an image within a predetermined range that includes the image of the symbol. The imaging apparatus according to claim 2.
  4. The extraction means extracts object images of a predetermined number of frames before and after the image of the symbol, or object images of a predetermined number of frames before the image containing the symbol, or object images of a predetermined number of frames after the image containing the symbol. The imaging apparatus according to claim 2.
  5. The imaging unit further comprises a first imaging unit and a second imaging unit installed at a different position and angle from the first imaging unit. The extraction means further extracts the object image from the image captured by the first imaging unit and the image captured by the second imaging unit. The association means associates the object image extracted from the image captured by the first imaging unit and the second imaging unit with the decoded information. The imaging apparatus according to any one of claims 2 to 4.
  6. A computer as an imaging device equipped with an imaging unit, A storage means for storing multiple images captured by the imaging unit, A decoding means that generates decoded information by decoding the symbols attached to the products contained in the image stored by the storage means , The aforementioned storage means stores multiple images, and the association means associates the decoded information with an object image containing a product enclosed in a specified region . An output means that outputs association information associated by the association means, A transmission means for transmitting either product identification information contained in the decoded information or product identification information identified by general object recognition performed on an image captured by the imaging unit to a product sales data processing device, A program to make it work.

Description

Embodiments of the present invention relate to an imaging device and a program. Recently, POS systems exist in stores such as shopping centers, mass retailers, and department stores that handle product registration and payment processing for items sold in the store. Some of these POS systems use imaging devices to capture images of products and then recognize the packaging to identify the product. By the way, in order to correctly identify a product from an image, it is necessary to improve the accuracy of the dictionary used to estimate the product from the image. To improve the dictionary's accuracy, it is necessary to collect many images related to the product in question and train the dictionary with them. For example, new products are released, and even for the same product, the packaging and other designs may change depending on the season, so it is necessary to train the dictionary on a daily basis. However, training the dictionary requires manual image capture of products, which places a burden on the workers. Figure 1 shows a system according to an embodiment of this system.Figure 2 is an explanatory diagram illustrating the principle of package recognition.Figure 3 shows an example of how an object detection model extracts product images.Figure 4 is a perspective view showing the external appearance of the POS terminal and imaging device.Figure 5 is a timing chart showing the processing flow in the system.Figure 6 is a block diagram showing the hardware configuration of a POS terminal.Figure 7 is a block diagram showing the hardware configuration of the imaging device.Figure 8 is a block diagram showing the hardware configuration of an edge device.Figure 9 is a block diagram showing the hardware configuration of the store server.Figure 10 is a flowchart showing the flow of control processing for a POS terminal.Figure 11 is a functional block diagram showing the functional configuration of the imaging device.Figure 12 is a flowchart showing the control process flow of the imaging device.Figure 13 is a flowchart showing the control process flow for edge devices.Figure 14 is a flowchart showing the flow of control processing for the store server. The embodiments of the present invention will be described below with reference to the drawings. In one embodiment, an imaging device that captures the customer's scanning of products in a system where the customer performs registration and accounting operations will be described as an example. However, the present invention is not limited to the first embodiment described below. Figure 1 shows a system according to an embodiment. In Figure 1, the system 10 is installed in a store T such as a supermarket, mass retailer, convenience store, or specialty store. The system 10 is a system installed within the store T and comprises one or more POS (Point of Sales) terminals 1 (product sales data processing devices), an imaging device 3 connected to each POS terminal 1, an edge device 5, and a store server 7. Each POS terminal 1, edge device 5, and store server 7 are connected to each other by a communication line L (e.g., LAN (Local Area Network)). Each imaging device 3 is connected to each POS terminal 1, for example, by a USB (Universal Serial Bus) cable U. Each imaging device 3 is also connected to the edge device 5 and the store server 7 by a communication line L. System 10 handles product registration, payment processing, and sales management for products sold at store T. System 10 also uses package recognition technology to perform product registration. Package recognition is a technology that identifies products by combining a technology that identifies a product by capturing and analyzing symbols such as barcodes and QR codes attached to the product (hereinafter referred to as "code recognition") with a known general object recognition technology that recognizes a product based on its shape and appearance as captured by the image. Regarding general object recognition, for example, see "Keiji Yanai, 'Current Status and Future of General Object Recognition,' Transactions of the Information Processing Society of Japan, Vol. 48, No. SIG16 [Retrieved August 10, 2010], Internet <URL: http://mm.cs.uec.ac.jp/IPSJ-TCVIM-Yanai.pdf>". Code recognition involves capturing an image of a symbol attached to a product and obtaining decoded information by decoding and analyzing that symbol. This decoded information includes product identification information that identifies the product to which the symbol is attached. This product identification information could be, for example, a product code, but any information that identifies the product is acceptable, not necessarily a product code. Next, we will explain general object recognition. Figure 2 is a diagram illustrating general object recognition. As shown in Figure 2, general object recognition uses three types of dictionaries—an object detection model 542, a feature extraction model 543, and a matching dictionary 544—to