Search

KR-102961927-B1 - AI-based integrated water management remote monitoring instrumentation and control system

KR102961927B1KR 102961927 B1KR102961927 B1KR 102961927B1KR-102961927-B1

Abstract

The present invention relates to a system for integrated management of regional water management facilities such as reservoirs, freshwater lakes, pumping stations, water and sewage treatment plants, river weirs, check dams, and multi-purpose dams, comprising: a GPU-embedded IoT-CCTV detection means (100) having a first regional sensor (110) to an n-th regional sensor (130) comprising a GPU-embedded IoT (101) that detects abnormal vibrations, noise, heat, ambient temperature and humidity, pressure, and levels of various facilities and devices within the management area; a GPU-embedded CCTV (102) capable of capturing video and images of intruder movements and anomaly detection of facilities and equipment; and a GPU-embedded Drone (103) for aerially photographing a wide area within the management area where the various facilities and devices are installed, or for monitoring the entire area within the management area and surrounding areas as needed in the event of a disaster; An AI-based streaming data processing system (200) that continuously collects and preprocesses data from the above-mentioned GPU-embedded IoT-CCTV detection means (100), and predicts warnings and immediate response measures after comparing and analyzing anomaly detection with a pre-trained AI model (240) based on the collected and preprocessed data, the database (524) of the central/regional control center (500), the weather forecast (140) of the Korea Meteorological Administration, and the data (150) of the Integrated Water Management Council; A streaming analysis and processing system is provided, each comprising: a video streaming analysis (310) that receives unstructured data of video/image and voice, including measurement text analyzed and predicted from the AI-based streaming data processing system (200), and immediately streams abnormal events different from the normal patterns of a pre-trained AI model (240) to be detected, and analyzes security systems such as intrusion detection, behavior pattern analysis, face recognition, and vehicle license plate recognition from real-time video/image analysis captured by the GPU-embedded CCTV (102); a voice streaming processing (320) that recognizes abnormal vibrations, noise, and the voice of an intruder in various equipment and devices within the facility management area from real-time voice analysis detected through the GPU-embedded IoT (101) and the GPU-embedded CCTV (102); and a multimodal streaming AI (330) that comprehensively analyzes and processes measurement text, video/image, and voice detected by the first regional sensor (110) to the nth regional sensor (130). A device (300); and an edge computing streaming AI combined system (400) that maximizes the real-time nature of data streamed from the streaming analysis and processing device (300), deploys AI models mounted on the GPU-embedded IoT (101) and GPU-embedded CCTV (102) to an edge computing device, thereby notifying the central/regional control center (500) of the data prioritized to the location of regional anomaly detection, and simultaneously directly notifies the manager/field worker (410) or automatically controls the water management automatic control system (420) in case of emergency; By implementing an AI-based integrated water management remote monitoring instrumentation and measurement control system characterized by including a central/regional control center (500) that comprehensively monitors and analyzes regional anomaly detection data received through a wired/wireless communication network (430) and a server (510) connected to the edge computing and streaming AI combined system (400), stores the data in a database (D/B), develops an AI model for training, and establishes an organic cooperative system with the field manager (410) and relevant agencies (440), it is possible to simultaneously realize high speed and autonomy of the real-time integrated water management instrumentation and measurement control system through emergency notifications and response measures, and there is an effect applicable to various industrial sites.

Inventors

  • 박홍대

Assignees

  • 주식회사 젠탑

Dates

Publication Date
20260507
Application Date
20250516

Claims (8)

  1. In a system for integrated management of regional water management facilities such as reservoirs, freshwater lakes, pumping stations, water and sewage treatment plants, river weirs, check dams, and multi-purpose dams, A GPU-embedded IoT-CCTV detection means (100) having a first regional sensor (110) to an n-th regional sensor (130) comprising a GPU (Graphics Processing Unit) embedded IoT (101) that detects abnormal vibrations, noise, heat, ambient temperature and humidity, pressure, and levels in various facilities and devices within the management area, a GPU-embedded CCTV (102) capable of capturing video and images of intruder movements and anomaly detection of facilities and equipment, and a GPU-embedded Drone (103) for capturing a wide area within the management area where the various facilities and devices are installed from the air or for monitoring the entire area within the management area and surrounding areas as needed in the event of a disaster; An AI-based streaming data processing system (200) that continuously collects and preprocesses data from the above-mentioned GPU-embedded IoT-CCTV detection means (100), and predicts warnings and immediate response measures after comparing and analyzing anomaly detection with a pre-trained AI model (240) based on the collected and preprocessed data, the database (524) of the central/regional control center (500), the weather forecast (140) of the Korea Meteorological Administration, and the data (150) of the Integrated Water Management Council; A streaming analysis and processing system is provided, each comprising: a video streaming analysis (310) that receives unstructured data of video/image and voice, including measurement text analyzed and predicted from the AI-based streaming data processing system (200), and immediately streams abnormal events different from the normal patterns of a pre-trained AI model (240) to be detected, and analyzes security systems such as intrusion detection, behavior pattern analysis, face recognition, and vehicle license plate recognition from real-time video/image analysis captured by the GPU-embedded CCTV (102); a voice streaming processing (320) that recognizes abnormal vibrations, noise, and the voice of an intruder in various equipment and devices within the facility management area from real-time voice analysis detected through the GPU-embedded IoT (101) and the GPU-embedded CCTV (102); and a multimodal streaming AI (330) that comprehensively analyzes and processes measurement text, video/image, and voice detected by the first regional sensor (110) to the nth regional sensor (130). Device (300) and; An edge computing streaming AI combined system (400) that maximizes the real-time nature of data streamed from the streaming analysis and processing device (300), deploys AI models mounted on the GPU-embedded IoT (101) and GPU-embedded CCTV (102) to edge computing devices, thereby notifying the central/regional control center (500) of the data prioritized to the location of regional anomaly detection, and simultaneously directly notifies the manager/field worker (410) or automatically controls the water management automatic control system (420) in case of emergency; An AI-based integrated water management remote monitoring instrumentation and measurement control system characterized by including a central/regional control center (500) that comprehensively monitors and analyzes regional anomaly detection data received through a wired/wireless communication network (430) and a server (510) connected to the above edge computing and streaming AI combined system (400), databases the data, develops an AI model for training, and establishes an organic cooperative system with the field manager (410) and related organizations (440).
  2. delete
  3. In Article 1, The above AI-based streaming data processing system (200) comprises a real-time data collector (210) that continuously collects data from the above GPU-embedded IoT-CCTV detection means (100), and A real-time preprocessor (220) that performs noise removal, data normalization, window partitioning processing, etc. on data collected through the real-time data collector (210), and An AI analysis and prediction device (230) that uses a pre-trained AI model (240) to classify anomaly detection and accident classification logs into text subtitles, and simultaneously performs warnings and immediate response predictions, and the data pre-processed through the above real-time preprocessor (220). An AI-based integrated water management remote monitoring instrumentation and measurement control system characterized by each having an AI model (240) that is read and written in the AI analysis and prediction device (230) and is pre-trained with the measurement text, video/image patterns, and voice waveforms of the normal state of various facilities and devices in the management facility where the first regional sensor (110) to the nth regional sensor (130) are installed.
  4. delete
  5. delete
  6. delete
  7. delete
  8. In Article 1, The above central/regional control center (500) comprises a server (510) that receives anomaly detection streaming processing data detected by the first regional sensor (110) to the nth regional sensor (130), and A comprehensive situation room (520) that has a first monitoring monitor (521) connected via TCP/IP from the above server (510) and having up to 64 split screens, an electronic map (E-Map) showing a regional water management system, 3D design drawings of various equipment and devices within the facility management area, and a second monitoring monitor (522) showing repair methods and repair videos capable of displaying text subtitles on a page-by-page basis, and which assesses the situation and gives work instructions to the manager/field worker (410) or requests cooperation with relevant agencies (440) in the event of a major disaster; An analysis system room (523) that analyzes data, video/images, and audio transmitted from the above server (510) and general situation room (520), transmits the analyzed data to the manager/field worker (410), or transmits the analyzed data to relevant agencies (440) in the event of a major disaster and provides support and cooperation, and An AI-based integrated water management remote monitoring instrumentation and measurement control system characterized by each having a database (524) that stores data transmitted from the above server (510) and analysis system room (523) by creating a database and transmitting it to a pre-trained AI model (240) of the above AI-based streaming data processing system (200) for use as a learning AI model.

Description

AI-based integrated water management remote monitoring instrumentation and control system The present invention relates to an integrated water management system, and more specifically, to an AI-based integrated water management remote monitoring instrumentation and measurement control system capable of simultaneously realizing high speed and autonomy, and enabling anomaly detection regardless of climate conditions or disasters, through an integrated water management instrumentation and measurement control system that combines a GPU-embedded IoT-CCTV detection means, an artificial intelligence model, streaming, and edge computing, unlike the integrated water management systems centered on central/regional control centers of the prior art. Recently, the UN has predicted that nearly half of the world's population will face severe water shortages in the near future due to abnormal climate conditions. In Korea, the National Water Management Committee was established under the President in June 2019 to deliberate and decide on important matters regarding water management, and basin water management committees were established under the National Water Management Committee for the Han River, Geum River, Yeongsan River, Seomjin River, and Nakdong River basins. In addition, there are nine integrated water management councils, including the Korea Water Resources Corporation, Korea Rural Community Corporation, Korea Environment Corporation, Korea Hydro & Nuclear Power, Korea Environment Institute, Korea Rural Economic Institute, Korea Institute of Land and Housing, Korea Institute of Construction Technology, and the National Institute of Disaster and Safety Research, where various research related to water management and technology development utilizing AI are being carried out. However, the design of a streaming artificial intelligence (AI) system is required to process the vast amount of data produced by these nine integrated water management councils. When building an AI-based real-time streaming system, a balance of the three major elements—speed, accuracy, and stability—is essential. Particularly in large-scale real-time environments, considerations include scalability—designing a horizontal scaling structure capable of processing data without performance degradation even during rapid data surges and adopting microservice architectures based on Kafka, Flink, and Kubernetes—and fault tolerance—implemented through checkpoint and reprocessing strategies for automatic recovery in the event of system failure and data loss prevention, as well as utilizing Spark Structured Streaming and Flink State Backend. Additionally, lightweight models and stream compression technologies are applied to reduce latency to millisecond levels, and latency optimization is applied to eliminate bottlenecks between inference servers and streaming nodes. Furthermore, to enhance operational efficiency in industrial settings, such as fault prediction and quality anomaly detection, AI will be further expanded by combining with edge devices, multimodal sensors, and autonomous systems. Meanwhile, an integrated facility management system using a regional data collection and recording device has been proposed in Korean Registered Patent Publication No. 10-0990362 (B1) filed and registered by the applicant of the present invention (October 29, 2010). However, the above registered patent technology is a method in which regional data acquired from CCTV/Web-Camera and still cameras is collected in a PC-based Data Acquisition (DAQ) Unit, video and audio data is analyzed for anomaly detection based on a central/regional control center, and databased to display information on intruders and the location of incidents only when a monitor clicks an icon. This not only fails to process data in real time but also fails to satisfy the balance of the three major elements—speed, accuracy, and stability—such as AI-based real-time streaming systems, and has the problem of being unable to simultaneously achieve high speed and autonomy. FIG. 1 is a drawing showing the prior art. FIG. 2 is a block diagram showing the overall technical configuration of an AI-based integrated water management remote monitoring instrumentation and measurement control system according to a preferred embodiment of the present invention. FIG. 3 is a block diagram showing a central/regional control center for an AI-based integrated water management remote monitoring instrumentation and measurement control system according to a preferred embodiment of the present invention. FIG. 4 is a drawing showing a first monitor of a central/regional control center for an AI-based integrated water management remote monitoring instrumentation and measurement control system according to a preferred embodiment of the present invention. FIG. 5 is a drawing showing a second monitor of a central/regional control center for an AI-based integrated water management remote monitoring instrumentation and measurement