KR-102962643-B1 - AI-Based behavior pattern analysis system, apparatus, and method for object recognition video processing in lodging facilities

KR102962643B1KR 102962643 B1KR102962643 B1KR 102962643B1KR-102962643-B1

Abstract

The present invention relates to a system for providing a behavioral pattern analysis solution through object recognition image processing related to accommodation facilities based on an artificial intelligence model, comprising: a video data receiving unit that receives video data from a plurality of CCTVs installed to film points where payment activities occur outside a guest room, in a hallway section, or in a parking area of the accommodation facility; a video analysis unit that includes an artificial intelligence model for each frame of the video input from the video data receiving unit, which detects payment-related objects including at least one of a person object and a banknote object, a coin object, a card object, or a mobile device object, and analyzes the change in distance or proximity between the payment-related objects and the person object to determine whether a payment event has occurred; an event recording unit that stores the frame number, location information, and time of occurrence of the payment event determined by the video analysis unit; a report generating unit that automatically generates a payment history report based on the payment event information stored in the event recording unit; a report transmission unit that transmits the payment history report generated by the report generating unit to an administrator terminal or a building owner terminal; and a POS device into which a payment amount is entered. The system includes a POS linkage comparison unit that communicates with the POS device, wherein the POS linkage comparison unit compares an event determined to be a payment event by the image analysis unit with payment amount data entered into the POS device and generates a warning signal if there is a discrepancy, and the system includes a bracket coupled to the CCTV, wherein an infrared sensor or a distance sensor for detecting hand movements of a human object is fixedly installed on the bracket, and when a hand movement is detected by the infrared sensor or the distance sensor, a payment analysis trigger signal is provided to the image analysis unit to control the analysis. The system provides a behavioral pattern analysis solution through object recognition image processing related to accommodation facilities based on an artificial intelligence model.

Inventors

이시형

Assignees

주식회사스테이지

Dates

Publication Date: 20260511
Application Date: 20251106

Claims (3)

In a system providing a behavioral pattern analysis solution through image processing for object recognition related to accommodation facilities based on an artificial intelligence model, A video data receiving unit that receives video data from multiple CCTVs installed to film points where payment transactions take place outside guest rooms, corridor sections, or parking areas of accommodation facilities; A video analysis unit comprising an artificial intelligence model that detects a payment-related object including at least one of a person object and a banknote object, a coin object, a card object, or a mobile device object for each frame of a video input from the video data receiving unit, and determines whether a payment event occurs by analyzing the change in distance or proximity between the payment-related object and the person object; An event recorder that stores the occurrence frame number, occurrence location information, and occurrence time of a payment event determined by the above-mentioned video analysis unit; A report generation unit that automatically generates a payment history report based on payment event information stored in the above-mentioned event record; A report transmission unit that transmits the payment history report generated by the above report generation unit to an administrator terminal or a building owner terminal; POS device for entering the payment amount; and It includes a POS interlocking comparison unit that communicates with the above POS device, and The above POS integration comparison unit is, When there is a discrepancy between an event determined as a payment event by the above-mentioned video analysis unit and the payment amount data entered into the above-mentioned POS device, a warning signal is generated, and The above system is, It includes a bracket coupled to the above CCTV, An infrared sensor or a distance sensor for detecting hand movements of a human object is fixedly installed on the above bracket, and When a hand movement is detected by the infrared sensor or distance sensor, a payment analysis trigger signal is provided to the image analysis unit to control the analysis. The above image analysis unit is, Determining as a cash payment candidate event the case where a banknote object or a coin object maintains a state of more than a preset frame continuously with the hand of a person object, and If the banknote object or coin object is not detected in the video frame in a subsequent frame, it is determined to be a final cash payment event, and A case in which the operation of the above card object approaching or being inserted into the insertion part of the POS device is detected is determined as a card payment event, and A mobile payment event is determined when a mobile device object is detected to be in close proximity to the NFC payment recognition area or wireless payment recognition area of the POS device, or when a QR code displayed on the mobile device is scanned by the POS device. When a receipt object is detected in the above POS device, it is determined as a payment completion event, and The above report generation unit is, Automatically generate a payment history report including the number of payment event occurrences, POS amount comparison results, and payment event occurrence time information, and A system providing a behavioral pattern analysis solution through object recognition image processing related to accommodation facilities based on an artificial intelligence model, comprising a payment proof generation unit that captures a frame in which a payment event occurred as a video image and generates it as payment proof data.
delete
In claim 1, The above system is, A sensor mat positioned in close proximity to a POS device where a payment transaction takes place, equipped with a pressure sensor or a metal detection sensor that generates a sensor signal by detecting pressure changes or the presence or absence of metal when a banknote or coin is placed on it; A sensor mat signal transmission unit that generates a payment event trigger signal according to the sensor output of the sensor mat and transmits it to the image analysis unit; A report transmission unit that automatically classifies the payment event as a payment omission event and transmits it to an administrator terminal if the payment amount is not entered into the POS device within a preset time after the payment event occurs; and It further includes a blockchain-based data storage unit that stores the above payment event information and payment proof data in blocks, and stores each block such that it includes the hash value of the previous block, thereby managing the stored blocks so that they can be continuously identified and verified through the hash values. The sensor mat above is, Housing body; An emergency request button on one side of the upper surface of the above housing body for transmitting a danger request signal to an emergency response agency terminal when a danger situation occurs; A mounting panel on the other side of the upper surface on which banknotes or coins are placed; A plurality of upwardly protruding anti-slip protrusions to prevent banknotes or coins placed on the upper surface of the above-mentioned mounting panel from sliding; and It includes a waterproof transparent cover formed to cover the upper surface of the housing body so that the pressure sensor or the metal detection sensor and the emergency request button are protected from liquid ingress, The above waterproof transparent cover is, It is detachably coupled to the above housing body, making maintenance or replacement easy, and A system providing a behavioral pattern analysis solution through object recognition image processing related to accommodation facilities based on an artificial intelligence model, wherein when the above emergency request button is pressed, the sensor mat signal transmission unit generates an emergency request signal and transmits it to a manager terminal, a building owner terminal, or an emergency response agency terminal.

Description

AI-Based behavior pattern analysis system, apparatus, and method for object recognition video processing in lodging facilities The present invention relates to a system for providing a behavioral pattern analysis solution through object recognition image processing related to accommodation facilities based on an artificial intelligence model. More specifically, the invention relates to a system for providing a behavioral pattern analysis solution that detects payment-related objects, including human objects and banknote, coin, card, or mobile device objects, from CCTV footage installed outside guest rooms, corridor sections, or parking areas of accommodation facilities using an artificial intelligence model to determine payment events, performs a function to detect missing or abnormal payments by comparing with payment amount data entered into a POS device, and analyzes fire events, violence events, or abnormal behavior events to automatically report them to a manager terminal, a building owner terminal, or an emergency response agency terminal. In addition, the present invention includes a payment trigger sensor module positioned in close proximity to a POS device where a payment transaction takes place, and provides a function to assist in video-based payment event determination or to transmit an emergency request signal when a dangerous situation occurs by utilizing a pressure sensor or a metal detection sensor and an emergency request button. In addition, the present invention belongs to the field of technology that prevents data forgery and ensures reliability by recording payment event information and payment proof data in block units and storing them in a blockchain-based data storage unit in a form including the hash value of the previous block. Meanwhile, the system of the present invention is not limited to accommodation facilities and can be extended to various industrial fields, such as automatically detecting unauthorized entry in unmanned stores (e.g., unmanned convenience stores, unmanned pet shops, etc.) in offline environments with high cash payment rates, or recognizing the behavioral patterns of human objects to automatically detect accident or dangerous behavior events and transmit notifications to control centers or emergency response agencies in environments where safety accidents or dangerous behaviors must be monitored, such as construction sites. Therefore, the present invention belongs to the field of smart accommodation operation management and payment fraud prevention system technology, which fuses AI-based object recognition technology, image analysis technology, POS payment integration technology, behavioral pattern analysis and automatic event reporting technology, and blockchain-based data security technology; additionally, it is a scalable technology applicable to offline commerce unmanned operation environments and industrial safety monitoring fields. In general, at accommodations such as motels, hotels, and serviced apartments, payment transactions often take place outside guest rooms, in hallways, or in parking areas. Particularly with the increase in unmanned accommodations, it is becoming common practice for customers to make payments directly using kiosks or POS terminals without front desk staff being present. In such an environment, there is a problem in that it is difficult for operators to verify the payment process in real time or monitor abnormal payment activities or payment omissions. For example, situations may arise where cash is exchanged between a customer and a third party during the settlement of room charges, where amounts are not actually entered into the POS device when transferring facility usage fees, or where the payment process is intentionally omitted. However, conventional CCTV systems installed in existing accommodations merely record video or stream it in real-time; they do not analyze human objects or payment-related objects, such as banknotes, coins, cards, or mobile devices, to determine payment status. In other words, standard CCTV systems are nothing more than simple video storage systems; they cannot identify what objects were exchanged or whether a monetary transaction actually took place during the payment process, and there are limitations in real-time capability and accuracy for accommodation operators to determine whether payments have been missed based solely on video footage. In addition, existing POS systems can check whether a payment amount has been entered, but they do not provide a function to compare or verify whether the payment amount has been entered with whether an actual physical exchange of money has taken place. In other words, since the exchange of money is confirmed solely through CCTV footage and the POS system only verifies whether the amount has been entered, there is no function to automatically detect or provide notifications for discrepancies between the video and the POS input data. For this reason, there is a