Edge 3D Reconstruction With Dynamic TSDF Update Frequency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing wireless communication systems, particularly 5G NR, face challenges in achieving high-fidelity three-dimensional scene reconstruction on edge devices with low latency and low power consumption, especially in environments where frequent image processing is required.
Innovation Solution
Implementing a method on edge devices to update a feature volume and truncated signed distance function (TSDF) volume based on a stream of posed images, using an adaptive frequency strategy that increases frequency initially and decreases it after the scene is fully observed, enabling fast and efficient 3D reconstruction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the TSDF volume is updated frequently to achieve high-fidelity 3D reconstruction, then the reconstruction quality is improved, but the power consumption and processing latency increase
Solution Approach 1:
The patent implements dynamic update frequency adjustment for the TSDF volume based on scene complexity and observation progress. The system transitions from frequent updates during initial scene capture to less frequent updates as the scene becomes fully observed, optimizing the balance between reconstruction quality and power consumption on edge devices
Solution Approach 2:
The system changes the update frequency parameter of the TSDF volume over time. Initially, updates occur at high frequency to capture scene geometry accurately, then the update frequency is reduced when the scene is fully observed, directly addressing the power consumption vs. reconstruction quality trade-off
2Loss of time
If the TSDF volume is updated at high frequency to reduce latency, then the observation latency is reduced, but the power consumption increases
Solution Approach 1:
The update frequency is made dynamic rather than static. During initial scene capture, high-frequency updates reduce latency for observing reconstructed surfaces. As the scene becomes fully observed, the frequency is reduced, maintaining low latency during critical phases while conserving power during stable observation phases
3Measurement precision
If continuous high-frequency image processing is performed, then the reconstruction fidelity is improved, but the power consumption on edge devices increases
Solution Approach 1:
The system uses periodic action with varying frequencies. Image processing is performed at high frequency during initial scene capture to ensure reconstruction fidelity, then transitions to lower frequency periodic updates when the scene is fully observed, reducing power consumption while maintaining acceptable fidelity
Solution Approach 2:
The processing frequency parameter is changed over time based on scene observation progress. The system adjusts the frequency of image processing operations to match the reconstruction needs at different stages, optimizing the balance between fidelity and power consumption
Data Source
AI summary
Aspects presented herein may improve the overall performance of three-dimensional (3D) scene reconstruction on edge devices by enabling a fast high fidelity 3D reconstruction for edge devices. In one aspect, a UE receives, from a camera, a stream of posed images. The UE updates, based on each posed image in the stream of posed images, a feature volume recursively. The UE updates, based on the updated feature volume, a truncated signed distance function (TSDF) volume. The UE outputs an indication of the updated TSDF volume. In some example, the UE may also receive a stream of depth maps associated with the stream of posed images, and the TSDF volume may be updated further based on the stream of depth maps.