海康观澜大模型落地见效 覆盖工业质检、新能源电力等多领域

Core Insights - Hikvision focuses on technological innovation in IoT perception, artificial intelligence, and big data, providing a comprehensive range of smart IoT hardware and software products, as well as services for complex intelligent IoT system construction [1] Group 1: Technology and Product Offerings - The company offers perception technologies that extend from visible light to millimeter waves, infrared, X-rays, and sound waves, integrating AI and big data into its smart IoT products, IT infrastructure, platform services, data services, and application services [1] - The "Guanlan" large model framework consists of a foundational model, industry models, and scenario application models, integrating multimodal data such as visual, acoustic, thermal, and X-ray data, and has been recognized for its maturity and reliability in model development [1] Group 2: Applications in Industrial Manufacturing - In the industrial manufacturing sector, the application of the Guanlan large model effectively addresses efficiency bottlenecks in traditional production processes, achieving an 80% increase in detection efficiency for soldering defects in circuit board production compared to traditional methods [2] - The model aids in real-time analysis of the entire product packaging process at Hikvision's manufacturing base, addressing common industry pain points related to flexible manufacturing and component misplacement [2] Group 3: Applications in Renewable Energy - In the renewable energy sector, the Guanlan large model's wind power prediction unit exemplifies its practical application, predicting future power generation based on wind direction and speed, with a 1% increase in prediction accuracy translating to significant revenue changes [3] - The wind power prediction unit utilizes a self-developed time series model that reduces long-term prediction errors, maintaining high accuracy for forecasts ranging from 15 minutes to ten days ahead, which lowers management and scheduling costs for the grid [3] - In a 300MW offshore wind farm in East China, the annual reduction in power prediction deviation assessment costs exceeded 1.2 million yuan, a decrease of over 15%, while in an 80MW mountainous wind farm in South China, the reduction exceeded 800,000 yuan, a decrease of over 46% [3] Group 4: Future Focus - The company aims to continue focusing on technological innovation and promote the large model's large-scale application across more industry scenarios, providing technical support for digital and intelligent transformation in various industries [3]