Object Tracking Device with Adaptive Motion State Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional object tracking methods fail to accurately detect and track objects that repeatedly move and stop, as they assume simple motion patterns, leading to missed detections when objects exhibit complex motion states like temporary resting or moving.
Innovation Solution
An object tracking device and method that differentiate between moving and resting objects by estimating the current state of a target object based on its past state, using a combination of background subtraction, long-period, and short-period background models, and particle filters to prioritize tracking information accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a typical tracking method assumes simple motion patterns (constant velocity or constant acceleration), then the tracking algorithm is computationally simple, but it fails to detect objects that repeatedly move and stop
Solution Approach 1:
The patent applies dynamics by making the motion model adaptive rather than static. The system dynamically switches between constant velocity model and constant acceleration model based on the object's current state (moving or resting). This allows the tracking algorithm to adapt to complex motion patterns while maintaining computational efficiency, resolving the contradiction between detection accuracy and model complexity.
Solution Approach 2:
The patent changes the motion parameters (velocity, acceleration) based on the object's state. When an object is detected as moving, the system uses constant velocity parameters; when resting, it uses constant acceleration parameters. This parameter switching enables accurate tracking of objects with repeated motion and resting states without requiring a single complex motion model.
2Reliability
If the system uses background subtraction with long-period background model to detect temporarily resting objects, then it can detect resting objects, but it may fail to track objects that move again after resting
Solution Approach 1:
The patent implements feedback by continuously monitoring the object's state and using this information to adjust the tracking approach. When an object transitions from resting to moving, the system receives feedback about this state change and switches from the long-period background model to the constant velocity model, ensuring continuous tracking without information loss.
Solution Approach 2:
The system dynamically adjusts the background model based on the object's current state. For resting objects, it uses the long-period background model to maintain detection reliability. When the object starts moving again, it transitions to using motion-based tracking with constant velocity or acceleration models, ensuring tracking continuity throughout the entire motion-resting cycle.
Data Source
AI summary
An object tracking device, method and non-transitory computer-readable storage medium are disclosed. An object tracking device may include an input terminal configured to receive an input image, and at least one processor connected to the input terminal wherein the at least one processor is configured to store a past state of a target object in a past image, detect a first object which is moving in the input image, detect a second object which is resting in the input image, estimate a current state of the target object based on the past state of the target object, wherein the current state comprises a current moving state when the target object is currently moving in the input image, and wherein the current state comprises a current resting state when the target object is currently resting in the input image, determine whether to prioritize tracking the target object with information corresponding to the first object or with information corresponding to the second object based on the current state of the target object.