Building defect intelligent detection and consultation method and system based on large model and machine vision

The intelligent building defect detection system, which combines large-scale models with machine vision, solves the problems of low efficiency and limited accuracy in existing technologies, and achieves efficient and accurate building defect detection and consultation, supporting mobile access and on-site processing.

CN118941533APending Publication Date: 2024-11-12CHINA INST OF BUILDING STANDARD DESIGN & RES
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411007477.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-25
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

Current building quality inspection relies on manual experience, which is inefficient and costly. It lacks the ability to combine and analyze image and text information, resulting in limited recognition accuracy and an inability to achieve flexible analysis and response.

Method used

This method employs a large model and machine vision approach. By acquiring image and text datasets of building defects, the data is cleaned, labeled, and fused to train a large language model and an intelligent image recognition model. Combined with basic building information, the method performs real-time detection and consultation, and generates a detection report.

Benefits of technology

It improves detection efficiency, reduces resource costs, enables intelligent analysis and unattended interactive consultation of building defects, enhances the accuracy and flexibility of defect assessment, supports mobile access, and facilitates on-site handling.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118941533A_ABST
    Figure CN118941533A_ABST
Patent Text Reader

Abstract

The invention relates to a building defect intelligent detection and consultation method and system based on a large model and machine vision. The method comprises the following steps: acquiring a building defect image and a text data set; performing cleaning, labeling and fusion learning on the building defect image and text data set to obtain a processed data set, and training, adjusting and integrating a large language model by adopting the data set to obtain a building defect intelligent detection and consultation platform; and searching in the platform according to the building image and the basic information of the building acquired in real time, identifying defects and performing prediction analysis, and generating a building detection report. According to the method, information fusion of defect image intelligent identification and defect language description is comprehensively utilized, deep learning and algorithm improvement are combined, accurate and efficient identification of building defects is achieved, professional analysis and prediction are provided, and important technical support is provided for building detection.
Need to check novelty before this filing date? Find Prior Art