22-08-2023 дата публикации
Номер: CN116630784A
Принадлежит:
The invention provides a machine vision iron notch state identification method, and relates to the technical field of metallurgical blast furnace ironmaking, and the method comprises the following steps: S1, image acquisition and data preparation: installing a camera at a blast furnace mouth, and getting through a network; s2, developing a deep learning image classification model, collecting an iron notch image to establish a data set, training an AI model, and testing and tuning; and S3, model deployment is in butt joint with the system, the AI model with the verified precision is deployed, a post-processing program is developed according to the tap hole time precision requirement, and the tapping time is written into the database. According to the method, intelligent, automatic and unmanned judgment of the blast furnace taphole in the taphole opening/closing state is realized, the tapping time is accurately positioned, and data support is provided for subsequent furnace number generation ...
Подробнее