Automatic inspection method, device and computer equipment for building facade

By using multi-target detection models and anomaly reasoning strategies, combined with image verification technology, the problems of low efficiency and misjudgment in traditional inspections have been solved, and high-precision automatic inspection of defects on building facades has been achieved.

CN122453818APending Publication Date: 2026-07-24LIGHT TRAP (BEIJING) TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
LIGHT TRAP (BEIJING) TECH CO LTD
Filing Date
2026-06-01
Publication Date
2026-07-24

AI Technical Summary

Technical Problem

Traditional manual inspections are inefficient and costly, and existing image recognition technologies have low accuracy in detecting defects on building facades, easily misjudging reflections, perspective objects, etc. as abnormalities.

Method used

A multi-class target detection model is used to identify the range of suspicious images. Combined with anomaly reasoning strategy and image verification strategy for building facades, the abnormal information of building facades is screened and generated. The multi-class target detection model is used to classify and identify image data of different anomaly types. Anomaly reasoning and image verification strategies are used to improve detection accuracy.

Benefits of technology

It effectively identifies and locates abnormalities on the exterior facade of damaged buildings, reduces misjudgments, improves the accuracy and comprehensiveness of inspections, and generates visualized inspection results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122453818A_ABST
    Figure CN122453818A_ABST
Patent Text Reader

Abstract

The application relates to an automatic inspection method and device for a building facade and a computer device. The method comprises the following steps: acquiring current image data of a building facade, identifying suspicious image ranges through a multi-class target detection model based on the current image data, identifying abnormal information of each suspicious image range through a building facade anomaly reasoning strategy based on each suspicious image range, screening abnormal image ranges based on the abnormal degree of each suspicious image range, identifying building facade abnormal information of the building facade through an image review strategy based on each abnormal image range, and generating a current inspection result of the building facade based on the building facade abnormal information. The method can improve the inspection accuracy of building facade abnormalities.
Need to check novelty before this filing date? Find Prior Art