Object Tracking via Edge Detection and Coordinate Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional object tracking methods face challenges in balancing tracking speed and accuracy, often resulting in time-consuming verification processes that can lead to inaccuracies when increasing tracking speed.
Innovation Solution
The method involves edge detection to generate edge images of template and candidate images, calculating pixel values in horizontal and vertical directions to create reduced-data edge images, and position-matching to determine coordinate information, ensuring accurate object tracking by comparing edge images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional pixel-by-pixel correlation methods are used for object tracking, then tracking accuracy can be maintained, but tracking speed becomes slow and time-consuming
Solution Approach 1:
The patent extracts only the edge information from the template image and candidate image, discarding the redundant pixel-by-pixel correlation data. By using edge detection to generate edge images and then calculating correlation only on these reduced edge images, the system maintains tracking accuracy while significantly improving tracking speed. This extraction of essential features (edges) resolves the contradiction between accuracy and speed.
2Productivity
If block correlation scheme with correlation window is used, then tracking speed can be improved, but verification of whether the tracked object is indeed the correct object becomes time-consuming
Solution Approach 1:
The patent extracts edge information from the images and performs correlation calculation only on these edge images rather than on the full images or large correlation windows. This extraction of essential edge features reduces the verification time while maintaining the ability to accurately identify and verify tracked objects, thus resolving the contradiction between tracking speed and verification time.
3Productivity
If tracking speed is increased to process images faster, then productivity improves, but tracking accuracy deteriorates and object identification becomes unreliable
Solution Approach 1:
The patent changes the parameter used for correlation calculation from full pixel data to edge image data. By transforming the input data from complete images to edge-extracted images, the system achieves faster processing (improved productivity) while maintaining reliable object identification (preserved accuracy). This parameter change enables simultaneous improvement of both tracking speed and accuracy.
Data Source
AI summary
A method and an electronic apparatus for comparing tracking object are provided. The method applies an edge detection to a template image and a candidate image to generate a template edge image and a candidate edge image respectively. A vertical template edge image and a horizontal edge template image caused by summing (or averaging) the pixel values of the template image horizontally and vertically are used to compare with the summing (or averaging) pixel values of the horizontal/vertical candidate edge images. A measurement in the comparison is based on the positions of partial horizontal candidate edge image and the partial vertical candidate edge image. The method determines that a region of the candidate image has the tracking object according to the vertical coordinate information and the horizontal coordinate information.