NTT DOCOMO, NTT Communications, and NTT Comware have jointly announced the immediate availability of a trial version of an AI-based system designed to monitor water levels in rivers and reservoirs. This trial is open to companies and local governments across Japan, with the aim of refining the system’s AI capabilities and user experience, ultimately leading to a commercial launch by March 2024.
A noteworthy aspect of this system is its elimination of physical water-level gauges. Instead, it utilizes virtual gauges and AI technology to provide real-time visualization of water levels through video imagery and time-series data displayed on a screen. Furthermore, the system can automatically alert authorities when a predetermined water level threshold is exceeded.
Key Features of the System:
- Virtual Gauge and AI Analysis: The system utilizes cameras and specialized equipment installed near water bodies to capture video images. AI video recognition technology overlays a virtual gauge on the video feed and accurately determines the water level in real time. An innovative algorithm and AI image segmentation technology analyze the ratio of water surface to land area displayed on the virtual gauge, with the option to use multiple virtual gauges for enhanced accuracy.
- Comprehensive Management Screen: Authorized personnel responsible for monitoring rivers and reservoirs can access a cloud-based management screen to view water levels, configure AI settings, and analyze data through video and graphs. The system also offers the capability to live-stream the information to local residents, ensuring timely alerts and updates.
- Integration with EDGEMATRIX®: The system incorporates NTT Com’s EDGEMATRIX® video analytics platform, which processes large volumes of video data in an Edge AI Box (edge computer). This enables efficient video analysis, water-level determination, and streaming, enhancing the system’s overall performance.
By introducing this AI-based water level monitoring system, NTT Group aims to revolutionize traditional monitoring practices and provide more efficient and accurate insights to support disaster prevention efforts. The trial phase will allow for fine-tuning and optimization, ensuring that the system meets the diverse needs of companies and local governments across Japan.
