Cross-Scene AR Relocalization via Intermediary Map Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing multi-person Augmented Reality (AR) technologies face challenges in relocalization, requiring devices to be in the same scene and angle, which complicates the process and leads to poor user experience.
Innovation Solution
A data processing method where a first device acquires an image of a target object in its real scene, builds a map, and transmits this data to a second device. The second device then performs relocalization based on the received map data and its own target object, even if the devices are in different real scenes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If devices perform relocalization based on being in the same scene, then relocalization accuracy is improved, but device complexity and operational difficulty increase
Solution Approach 1:
The patent introduces map data as an intermediary element that mediates between devices in different scenes. Instead of requiring direct scene overlap, the first device generates map data from its scene, transmits it to the second device, and the second device uses this map data to perform relocalization in its own scene. This intermediary map data enables cross-scene relocalization while maintaining accuracy.
Solution Approach 2:
The patent transitions from spatial co-location (devices in the same scene) to data-space interaction (devices sharing map data). By changing the dimension of interaction from physical space to data space, devices can perform relocalization without being in the same physical scene, thereby simplifying the operational requirements.
2Measurement precision
If devices require being in the same scene for relocalization, then relocalization precision is improved, but ease of operation deteriorates
Solution Approach 1:
Map data serves as an intermediary that eliminates the need for devices to be in the same scene. The first device creates map data representing its scene, transmits this data to the second device, and the second device uses it to achieve precise relocalization in its own scene. This intermediary mechanism maintains precision while greatly improving ease of operation.
Solution Approach 2:
The first device performs preliminary actions by generating and transmitting map data before the second device needs to perform relocalization. This preliminary preparation of scene information allows the second device to quickly perform relocalization without needing to physically co-locate with the first device, thereby improving operational ease.
3Adaptability or versatility
If devices are in different real scenes, then adaptability is improved, but relocalization difficulty increases
Solution Approach 1:
The patent uses map data as an intermediary to bridge different real scenes. The first device generates map data from its scene, transmits it to the second device in a different scene, and the second device uses this external scene data to perform relocalization. This intermediary mechanism enables cross-scene adaptability while managing the complexity of detecting and measuring features across different environments.
Solution Approach 2:
The patent solves the cross-scene relocalization problem by transitioning from direct spatial feature matching to data-space feature matching. Devices compare features against transmitted map data rather than requiring direct visual contact with the same physical features, enabling adaptability across different scenes while reducing measurement difficulty.
Data Source
AI summary
A data processing method and an electronic device are provided. The data processing method is implemented by a first device. In the method, an image including a first target object is acquired, the first target object being located in a first reality scene where the first device is located. Based on the image, a map corresponding to the first reality scene is built, and data of the map is obtained. The data of the map is transmitted to a second device, the data of the map being used to instruct the second device to perform relocation based on the data of the map and a second target object, the first and second target objects having at least two identical feature points, the second target object being in a second reality scene where the second device is located, and the second reality scene being different from the first reality scene.