Feature Tracking for Autonomous Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing feature tracking systems, both software-based and hardware-accelerated, face challenges in efficiently tracking numerous features in real-time, especially in applications like automotive systems where hundreds or thousands of features need to be tracked, due to resource constraints and limitations in spatial location precision.
Innovation Solution
The system merges detected features from a feature detector with tracked features from a feature tracker to identify and prioritize important features for tracking, allowing for real-time or near real-time feature tracking without excessive resource usage, and enabling tracking at subpixel locations for increased accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If software is used to track features through a sequence of images, then feature tracking can be implemented, but the run time becomes linearly dependent on the number of features making it prohibitive for real-time applications
Solution Approach 1:
The patent replaces software-based feature tracking with hardware-accelerated processing. Specifically, it uses a hardware-accelerated processor to perform optical flow estimation and feature tracking operations that would otherwise be executed through software, thereby achieving real-time performance for tracking hundreds or thousands of features without linear time increase
2Productivity
If hardware-accelerated processors are used to estimate optical flow, then processing speed improves, but the system cannot track features at subpixel locations reducing accuracy
Solution Approach 1:
The patent merges the advantages of both hardware-accelerated processing and software-based precision methods. It combines hardware-accelerated optical flow estimation with software-based subpixel refinement techniques, allowing the system to achieve both real-time processing speed and subpixel location accuracy by integrating both approaches in a unified feature tracking system
3Reliability
If all detected features are tracked, then comprehensive feature coverage is achieved, but resource consumption increases excessively
Solution Approach 1:
The patent applies local quality by differentiating the treatment of features based on their importance and location. It prioritizes tracking of features in specific regions (such as those identified as important by the hardware-accelerated processor) while reducing or eliminating tracking of less important features, thereby achieving comprehensive coverage of critical features while reducing overall computational resource consumption
Data Source
AI summary
In various examples, feature tracking for autonomous or semi-autonomous systems and applications is described herein. Systems and methods are disclosed that merge, using one or more processes, features detected using a feature tracker(s) and features detected using a feature detector(s) in order to track features between images. In some examples, the number of merged features and/or the locations of the merged features within the images are limited. This way, the systems and methods are able to identify merged features that are of greater importance for tracking while refraining from tracking merged features that are of less importance. For example, if the systems and methods are being used to identify features for autonomous driving, a greater number of merged features that are associated with objects located proximate to the driving surface may be tracked as compared to merged features that are associated with the sky.