Feature Pattern Tracking System Reducing Computational Load
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image-based object tracking methods are burdensome due to the need for extensive image recognition in high-resolution images, increasing the cost and complexity of tracking objects.
Innovation Solution
A tracking system and method that utilizes a trackable device with a feature pattern and optical sensing modules to rapidly identify and track objects by capturing images with multiple cameras, employing deep learning algorithms and inertial measurement units to predict and refine the location of feature patterns across images, reducing the need for full image recognition in each frame.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional image recognition methods are used to track objects by analyzing many consecutive images, then the tracking can be performed with standard resolution images, but the computational burden and cost increase significantly
Solution Approach 1:
The patent divides the image processing task into two segments: first detecting feature patterns (such as QR codes or distinctive markers) which are small and easy to recognize, then using the detected positions to guide the tracking process. This segmentation allows the system to avoid processing entire high-resolution images frame by frame, instead focusing computation only on relevant regions, thereby reducing computational burden while maintaining tracking accuracy
Solution Approach 2:
The patent performs preliminary feature pattern detection before the main tracking process. By pre-identifying distinctive features and their positions in advance, the system establishes reference points that guide subsequent tracking operations. This preliminary action reduces the need for exhaustive image analysis in each frame, improving tracking efficiency without sacrificing precision
2Measurement precision
If high-resolution images are used for object tracking, then the image quality and detail are improved, but the computational cost and processing time increase
Solution Approach 1:
The patent extracts only the essential feature patterns from the images rather than processing the entire image data. By identifying and isolating distinctive features (such as specific patterns or markers), the system achieves accurate tracking while minimizing the amount of data that needs to be processed, thereby reducing processing time without losing recognition accuracy
Solution Approach 2:
The patent uses simplified representations or copies of the target object's feature patterns rather than processing the complete high-resolution image data. By working with extracted feature patterns as simplified copies, the system maintains sufficient information for accurate tracking while significantly reducing computational requirements and processing time
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present disclosure provides a tracking system and method thereof. The tracking system includes a trackable device with an appearance including a feature pattern and a tracking device. The tracking device includes an optical sensor module configured to capture a first image which covers the trackable device. The tracking device further includes a processor coupled to the optical sensor module. The processor is configured to retrieve a region of interest (ROI) of the first image based on the feature pattern, and locate a position of each of a plurality of feature blocks in the ROI, where each feature block contains a portion of the feature pattern. The processor further calculates a pose data of the trackable object according to the positions of the feature blocks.