Industrial space intelligent-oriented factory three-dimensional digital twin model reconstruction and incremental updating method

By using a bifurcated neural implicit field network and an incremental update method, the problems of low efficiency and inconsistent updates in the full reconstruction of factory 3D digital twin models are solved, achieving efficient, agile and high-fidelity 3D digital twin model updates, which are suitable for the reconstruction and incremental updates of 3D digital twin models of factories for industrial space intelligence.

CN121746644BActive Publication Date: 2026-06-16BEIJING FEIDU TECH CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN Β· China
Patent Type
Patents(China)
Current Assignee / Owner
BEIJING FEIDU TECH CO LTD
Filing Date
2026-02-26
Publication Date
2026-06-16

AI Technical Summary

Technical Problem

The existing full reconstruction of factory 3D digital twin models is inefficient, and the full scan triggered by local changes leads to high data redundancy, which cannot meet the needs of industrial scenarios for agile model evolution. Furthermore, geometric distortion or physical breakage is prone to occur during the update process, which cannot guarantee the long-term logical rigor.

Method used

A bifurcated neural implicit field network is adopted to generate an initial 3D digital twin base through multi-view image training. The structural change area is located by real-time image difference analysis of the inspection robot. The radiation field and SDF network weights are adjusted for incremental updates. Combined with teacher-student distillation loss and multi-resolution hash encoding, a high-fidelity explicit 3D digital twin model of the factory is generated.

Benefits of technology

It enables agile updates of the factory's 3D digital twin model, increases the model evolution frequency from "days" to "minutes", reduces operation and maintenance costs, ensures geometric consistency and high-fidelity rendering effects, and enhances perception robustness and version differential maintenance capabilities.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121746644B_ABST
    Figure CN121746644B_ABST
Patent Text Reader

Abstract

The application provides a plant three-dimensional digital twin model reconstruction and incremental updating method for industrial space intelligence, comprising: acquiring plant multi-view basic images for a plant panoramic scene to train a bifurcated structure neural implicit field network containing a radiation field branch and an SDF branch, and generate an initial three-dimensional digital twin base; acquiring plant real-time inspection images collected by an inspection robot at different inspection poses, rendering the initial three-dimensional digital twin base by using the radiation field branch to obtain a virtual reference view, and performing difference analysis on the plant real-time inspection images to locate a structural change area in the plant; determining an axis-aligned bounding box of a physical range of a changed entity in the area as an effective boundary of a local space patch and determining matched radiation field network weights and SDF network weights according to the effective boundary; adjusting the radiation field network weights and the SDF network weights and updating the network in an incremental updating manner to generate a plant three-dimensional digital twin model based on an implicit three-dimensional geometric field.
Need to check novelty before this filing date? Find Prior Art