Systems, methods, and apparatuses for aligning image frames

By downsampling and reducing the image frames of autonomous vehicles, combined with image pyramid technology, the problem of low processing efficiency of high-resolution image frames is solved, enabling faster and more accurate image data alignment and object detection.

CN115908869BActive Publication Date: 2026-07-07WAYMO LLC
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
WAYMO LLC
Filing Date
2022-09-30
Publication Date
2026-07-07

AI Technical Summary

Technical Problem

Autonomous vehicles require high computational resources and are time-consuming when processing high-resolution image frames, resulting in low image data processing efficiency and difficulty in quickly and accurately detecting changes in objects in the environment.

Method used

By downsampling or reducing image frames into multiple versions of different sizes, motion estimation and alignment are performed using image pyramid techniques, reducing computational load and improving processing efficiency. Alignment information is used to align larger resolution image frames.

Benefits of technology

It improves the efficiency and accuracy of image processing, reduces processing latency, enables faster detection of object changes between image frames, and enhances signal-to-noise ratio and image quality.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115908869B_ABST
    Figure CN115908869B_ABST
Patent Text Reader

Abstract

Examples described relate to an apparatus comprising a memory to store image frames and at least one processor. The at least one processor can be configured to receive a plurality of image frames from an image capture device and downsize each of the plurality of image frames to generate a plurality of versions of each image frame at a plurality of different sizes. The at least one processor can be further configured to determine alignment information for a first version of a first image frame. The alignment information can comprise a first alignment vector to identify image data in a first version of a second image frame that corresponds to image data in the first version of the first image frame. Moreover, the at least one processor can be configured to determine, based at least on the first alignment vector, a first initial alignment vector to identify image data in a first version of a third image frame.
Need to check novelty before this filing date? Find Prior Art