On November 20, the award ceremony for the 2025 Metaverse AI Data Elements "Golden Apricot Award" national competition successfully concluded in Guangzhou. The project titled *High-Quality Industrial Dataset Construction—Metering Innovation in the AI Era: Application of METER MIND Visual Recognition Technology in the Water Industry*, submitted by Hangzhou Water Meter, was honored with the highest accolade of the competition—the "Golden Apricot First Prize"—for its outstanding achievements in deeply integrating metaverse data elements with industrial scenarios.
As an authoritative competition among the first batch certified by the National Data Administration, the "Golden Apricot Award" was organized by the China Mobile Communications Association and jointly guided by esteemed experts, including Academician Shen Changxiang and Academician Zheng Weimin, underscoring its high industry credibility and benchmark significance. After multiple rounds of rigorous evaluation—including material reviews, video presentations, expert commentaries, and supplementary defenses—14 exceptional projects were selected as finalists from 289 submissions to compete for the "Metaverse Benchmark—Golden Apricot Award." With only two first prizes awarded, METER MIND's visual recognition technology distinguished itself with outstanding performance and emerged victorious.
Centered on "AI vertical models + metaverse data elements," METER MIND's visual recognition technology addresses long-standing challenges in the water industry, such as difficulties in meter reading, verification, and high costs associated with meter upgrades. By enabling low-cost, high-precision data collection, the technology constructs high-quality industrial datasets, providing a critical foundation for transforming water utility management from experience-driven to data-driven approaches. It lays a solid technical groundwork for implementing various smart water management scenarios, including intelligent scheduling, district metering, risk, and even the construction of digital twin systems.
In terms of technical approach, METER MIND adopts a "low-cost retrofitting + multimodal large model recognition" method to achieve high-precision identification of mechanical water meter readings, with an accuracy rate of up to 99.99% and excellent robustness. Multidimensional data such as readings and temperature collected by the system can be endowed with new contextual value within the metaverse framework, supporting water utility operational decision-making, leakage预警, and user service optimization. This fully unleashes the core driving force of data in virtual simulation, predictive analysis, and other critical processes.
In terms of technical capabilities, METER MIND demonstrates three core strengths:
1. **Zero-shot cold start and continuous iteration capabilities**: The model continuously self-optimizes within the metaverse data environment.
2. **High degree of localization in hardware**: It utilizes domestic chips, such as those from HiSilicon, with a localization rate exceeding 95%.
3. **Full intellectual property ownership of the software model**: This ensures technological security and controllability.
METER MIND is not only a breakthrough in the application of AI technology in the water industry but also a significant practice of empowering industrial scenarios through metaverse data elements. As this technology is promoted nationwide, it will further advance the evolution of water systems toward a new form of virtual-real integration and data-driven operations, providing a replicable model for the implementation of the industrial metaverse.
