A dynamic sample database construction method, device and medium for a remote sensing cross-scale interpretation large model

By constructing a dynamic sample database for a large-scale remote sensing interpretation model, the problems of inconsistent category systems, chaotic management of multi-source data, low degree of automation in annotation, and poor retrieval reusability in remote sensing sample databases have been solved. This has enabled efficient and dynamic sample management and large-scale model training optimization, thereby improving remote sensing interpretation performance.

CN122364480APending Publication Date: 2026-07-10ZHONGKE XINGTU DIGITAL EARTH HEFEI CO LTD
View PDF 1 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
ZHONGKE XINGTU DIGITAL EARTH HEFEI CO LTD
Filing Date
2026-03-26
Publication Date
2026-07-10

AI Technical Summary

Technical Problem

Existing remote sensing sample databases suffer from technical bottlenecks such as inconsistent category systems, chaotic management of multi-source data, low automation of sample annotation, poor retrieval reusability, and static architecture that is difficult to adapt to dynamic business needs, resulting in low efficiency in the training and application of large remote sensing interpretation models.

Method used

We construct a dynamic sample database for large-scale remote sensing interpretation models. Through a knowledge graph of land use samples, unified data standardization, human-computer collaborative annotation, and an iterative mechanism driven by feedback from large models, we achieve automated and accurate annotation, dynamic updating and optimization of samples, forming a unified semantic space and an efficient retrieval system.

Benefits of technology

It improves the cross-scale adaptability and interpretation accuracy of large remote sensing interpretation models, reduces the cost of acquiring high-quality samples, enhances the dynamic adaptation capability of sample databases and the training efficiency of large models, and supports multi-task adaptation and continuous iteration.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122364480A_ABST
    Figure CN122364480A_ABST
Patent Text Reader

Abstract

This invention discloses a method, device, and medium for constructing a dynamic sample database for a large-scale remote sensing interpretation model. It relates to the field of remote sensing image sample database construction technology. The method includes: constructing a land cover sample knowledge graph to clarify the interpretation marker information of each land cover sample; constructing a standardized land cover sample database and storing land cover sample labels by category; initializing land cover segmentation masks and labels based on the SAM3 visual interpretation model, and performing hierarchical quality checks on the land cover segmentation masks and labels; dynamically iteratively updating the database, retrieving land cover samples from the database to generate a training set, and fine-tuning or optimizing the interpretation model. This invention constructs a knowledge graph adapted to the large-scale remote sensing interpretation model, forming a unified and scalable category system and interpretation marker information. This system can adapt to different regions and task requirements, providing a unified semantic foundation for the training and validation of the large-scale remote sensing interpretation model and solving the problem of inconsistent category systems.
Need to check novelty before this filing date? Find Prior Art