Coal preparation plant visual production monitoring method and system based on visual language model

By analyzing video data from coal preparation plants using visual language models and large language models, the problems of low efficiency in manual monitoring and strong limitations of intelligent systems in coal preparation plant video monitoring have been solved. This has enabled in-depth understanding of content from multiple video sources and upstream and downstream correlation analysis, thereby improving fault handling efficiency.

CN122116274APending Publication Date: 2026-05-29TIANJIN MEITENG TECH CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
TIANJIN MEITENG TECH CO LTD
Filing Date
2026-02-14
Publication Date
2026-05-29

AI Technical Summary

Technical Problem

Video monitoring in coal preparation plants suffers from problems such as low efficiency of manual monitoring, strong limitations of existing intelligent systems, and insufficient value mining of video data. It cannot achieve in-depth understanding of content from multiple video sources and upstream and downstream correlation analysis, resulting in low efficiency in fault location and untimely handling.

Method used

A visual language model-based approach is adopted, which analyzes the equipment operation videos of multiple production equipment through visual language model (VLM) and large language model (LLM) to generate a single equipment production monitoring report. Based on the upstream and downstream topology, abnormal event correlation analysis is performed to construct a real-time heterogeneous graph for fault reasoning and generate a multi-equipment production monitoring report.

Benefits of technology

It enables in-depth understanding of content from multiple video sources and upstream and downstream correlation analysis, improving the intelligence level of coal preparation plant production monitoring and fault handling efficiency, reducing the workload of manual monitoring, and improving the accuracy and efficiency of fault location and handling.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122116274A_ABST
    Figure CN122116274A_ABST
Patent Text Reader

Abstract

The present application provides a kind of coal preparation plant visual production monitoring method and system based on visual language model, it is related to abnormal monitoring technical field, comprising: obtaining the equipment running video corresponding to multiple production equipment;Determine the single equipment production monitoring report corresponding to multiple production equipment based on equipment running video through visual language model, and single equipment production monitoring report at least includes the state data and abnormal event data corresponding to single production equipment;According to the upstream and downstream topological relationship between single equipment production monitoring report and multiple production equipment, abnormal event correlation analysis is carried out, and multiple equipment production monitoring report is obtained, and multiple equipment production monitoring report at least includes abnormal correlation analysis result.The present application can realize the depth understanding of multiple video source content, upstream and downstream correlation analysis and the comprehensive research of production state, improve the intelligent level of coal preparation plant production monitoring and fault processing efficiency.
Need to check novelty before this filing date? Find Prior Art