Spatially intelligent dynamic panoramic scene rendering method fusing depth and motion constraints

By mapping multimodal spatial scene data into tensor structures, generating dense depth maps and 3D motion fields, and performing multidimensional analysis and semantic optimization, the problems of dynamic object deformation and projection errors in panoramic rendering are solved, achieving high-precision panoramic image consistency and coherence.

CN122244260APending Publication Date: 2026-06-19BEIJING FEIDU TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN Β· China
Patent Type
Applications(China)
Current Assignee / Owner
BEIJING FEIDU TECH CO LTD
Filing Date
2026-03-23
Publication Date
2026-06-19

AI Technical Summary

Technical Problem

Existing panoramic rendering solutions lack explicit physical constraints on the motion trends of objects in complex dynamic environments. This results in deformation and artifacts of dynamic objects, and the effects of projection errors and velocity change rates are not effectively optimized, leading to poor consistency of panoramic images, flickering, breaks, and discontinuous motion trajectories.

Method used

By mapping multimodal spatial scene data into tensor structure data, analyzing spatial and motion features, generating dense depth maps and 3D motion fields, analyzing image projection errors, motion trajectories, and velocity change rates, rendering is performed in conjunction with semantic optimization constraints, and finally, full-angle panoramic image mapping and correction are performed in a spherical coordinate system.

Benefits of technology

It improves the spatial consistency and visual quality of panoramic images, solves the problems of dynamic object deformation, projection drift and speed discontinuity, and ensures the geometric closure and light and shadow continuity of panoramic images.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122244260A_ABST
    Figure CN122244260A_ABST
Patent Text Reader

Abstract

This application provides a spatial intelligent dynamic panoramic scene rendering method that integrates depth and motion constraints. The method includes: mapping multimodal spatial scene data into tensor structure data and parsing spatial feature data and motion feature data; generating a dense depth map based on the spatial feature data and a three-dimensional motion field based on the motion feature data to generate a first modeling image; performing image projection error, motion trajectory, and velocity change rate analysis on the first modeling image and performing constraint optimization to obtain a second modeling image; performing spatiotemporal consistency optimization and geometric semantic feature rendering to obtain a rendered spatial modeling image; mapping to a spherical coordinate system to generate a full-angle panoramic image and performing mapping error correction to obtain a panoramic image of the spatial scene. This application significantly alleviates the spatiotemporal distortion and geometric inconsistency problems in dynamic scene rendering through joint modeling of depth and motion factors and multi-dimensional dynamic constraint optimization, improving the continuity and visual quality of the panoramic image.
Need to check novelty before this filing date? Find Prior Art