Adaptive Template Matching for Vehicle Tracking Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object tracking systems face challenges in accurately tracking vehicles when their orientation or size changes significantly in images, particularly due to camera attitude adjustments or long distances, leading to reduced pattern matching accuracy.
Innovation Solution
An image processing apparatus that uses multiple templates of varying sizes and shapes to track objects across frames, adjusting based on detected movement and orientation changes, and determines tracking continuation or termination based on matching scores and edge intensity thresholds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single template is used for pattern matching to track vehicles, then the device complexity is low, but the tracking accuracy deteriorates when vehicle orientation or size changes significantly
Solution Approach 1:
The patent applies dynamics by making the template adjustable and adaptable rather than fixed. The template is dynamically updated based on detected vehicle characteristics and orientation changes, allowing the tracking system to adapt to varying vehicle appearances, orientations, and sizes while maintaining accurate tracking performance
Solution Approach 2:
The patent changes the parameters of the template (such as size, shape, and orientation) based on detected vehicle characteristics. By modifying template parameters according to actual vehicle appearance variations, the system maintains high tracking accuracy despite changes in vehicle orientation, distance, or perspective
2Area of stationary object
If the camera attitude is adjusted to track moving vehicles, then the coverage area increases, but the object size in the image changes causing pattern matching failure
Solution Approach 1:
The template is dynamically adjusted in response to camera attitude changes and resulting object size variations. The system adapts the template parameters to match the current object size in the image, maintaining pattern matching accuracy even as the camera pans, tilts, or zooms to track moving vehicles across different areas
Solution Approach 2:
The system uses feedback from the pattern matching process to detect when tracking accuracy deteriorates due to object size changes. Based on this feedback, the template parameters are automatically adjusted to compensate for size variations caused by camera attitude adjustments, ensuring continuous accurate tracking
3Reliability
If the vehicle turns or changes orientation on the road, then the vehicle movement is accurately reflected, but the pattern matching fails due to sight change
Solution Approach 1:
The template is dynamically updated to reflect changes in vehicle orientation and appearance. As the vehicle turns or changes direction, the template adapts to the new vehicle posture and orientation, maintaining reliable movement detection and accurate pattern matching throughout the vehicle's maneuver
Solution Approach 2:
The patent changes the template parameters (such as orientation angle, size, and shape) to match the vehicle's current orientation and appearance. By adjusting these parameters in response to vehicle turns and maneuvers, the system maintains both reliable movement detection and accurate pattern matching
Data Source
Figure 1
Figure 2
Figure 3A~3B
AI summary
An image processing apparatus for tracking a target object to be tracked included in a plurality of consecutively captured images comprises an image acquisition unit configured to acquire an image; and a tracking unit configured to detect a target object to be tracked included in the image by matching and acquire over a plurality of the images an amount of movement of the target object to be tracked and a direction of the movement of the target object to be tracked, wherein the tracking unit performs first matching by using a first template corresponding to the target object to be tracked, and performs second matching by using a second template, which is different in size or shape from the first template, in a case where a score of a result of the first matching is lower than a first threshold value.