3D Map Descriptor Retrieval With Asymmetric Binary Matching

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing 3D map retrieval methods require large computational resources and time due to the large data volume, affecting user experience in applications like VR, AR, MR, autonomous driving, and robotic navigation.

Innovation Solution

Implement asymmetric binary data retrieval using 3D map descriptors and retrieval descriptors of varying lengths to improve retrieval performance by optimizing storage and transmission overheads while maintaining accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If original 3D map data is used for retrieval, then retrieval accuracy is maintained, but computational resources and time consumption increase significantly

Engineering Contradiction:

Improveretrieval accuracyVSAvoidretrieval speed

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The patent segments the 3D map data into multiple descriptors (e.g., geometric descriptors, appearance descriptors, semantic descriptors) that can be processed independently. This segmentation allows the system to retrieve only relevant descriptor components rather than processing the entire large-scale 3D map data, thereby improving retrieval speed while maintaining accuracy through selective descriptor matching.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent extracts essential features from the original 3D map data to create a condensed representation. By extracting only the most discriminative descriptors (such as key geometric features or distinctive appearance characteristics) and discarding redundant information, the system achieves fast retrieval with reduced computational load while preserving the critical information needed for accurate matching.

Inventive Principle:

Principle #2Taking out (Extraction)

2Quantity of substance

If 3D map data is compressed to reduce data volume, then storage and transmission overheads are reduced, but retrieval accuracy may deteriorate

Engineering Contradiction:

Improvedata volumeVSAvoidretrieval accuracy

Core Design Contradiction:

Quantity of substanceVSMeasurement precision

Solution Approach 1:

The patent applies different compression strategies to different types of descriptors based on their importance and characteristics. Critical geometric descriptors that define the fundamental structure of 3D map points are compressed with higher precision to preserve matching accuracy, while less critical appearance or semantic descriptors use more aggressive compression. This local quality differentiation ensures that compression reduces data volume without uniformly degrading retrieval accuracy across all descriptor types.

Inventive Principle:

Principle #3Local quality

Data Source

PatentUS12536207B2Method and apparatus for retrieving three-dimensional (3D) map

Publication Date: 2026.01.27 HUAWEI TECH CO LTD

AI summary

A method and an apparatus are disclosed for retrieving a 3D map by obtaining binary data of a plurality of 3D map descriptors, obtaining binary data of a retrieval descriptor, and performing retrieval in the binary data of the plurality of 3D map descriptors based on binary data of the retrieval descriptor, to obtain at least one target 3D map descriptor. The plurality of 3D map descriptors correspond to at least one 3D map point of the 3D map. The retrieval descriptor is a feature that corresponds to a real environment and that is extracted from visual information collected by a sensor of an electronic device. A length of binary data of each of the plurality of 3D map descriptors is different from a length of the binary data of the retrieval descriptor.