Object Tracking Device Occlusion Handling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object tracking methods struggle when multiple objects are hidden behind another object, leading to incorrect identification numbers being assigned upon re-emergence, especially when conditions of the hidden objects change.
Innovation Solution
An object tracking method and apparatus that estimates the positional relationship between multiple objects before they enter a hidden area, and upon re-emergence, uses this information to correctly assign identification numbers based on pre-hidden states.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Duration of action of stationary object
If estimated information about objects is used to determine reappearance and assign identification numbers, then object tracking can continue during occlusion, but wrong identification numbers may be assigned when multiple objects are hidden and their conditions change
Solution Approach 1:
The system performs preliminary actions by detecting objects before they enter the hidden area and recording their identification numbers and positions. This preliminary detection and recording enables the system to restore correct identification information when objects reappear, preventing misidentification during occlusion events.
Solution Approach 2:
The system introduces an intermediary mechanism (the detection and recording unit) that captures and stores object information before occlusion occurs. This intermediary record serves as a reference to match against reappearing objects, ensuring accurate identification even when multiple objects are hidden behind obstacles.
2Quantity of substance
If multiple objects are tracked simultaneously, then comprehensive monitoring is achieved, but complexity of managing identification during occlusion increases
Solution Approach 1:
The system segments the tracking process into distinct functional units: a detection unit that identifies objects before occlusion, a recording unit that stores identification information, and a matching unit that restores identities after occlusion. This segmentation simplifies the management of multiple objects by handling each object's identification independently through standardized steps.
Solution Approach 2:
The system creates copies of object identification information during the detection phase and stores these copies for later use. When objects reappear, the system compares the current object data with the stored copies to restore correct identification numbers, avoiding the need for complex real-time tracking management during occlusion.
3Speed
If objects are detected in real-time, then timely response is achieved, but correct identification during and after occlusion becomes difficult
Solution Approach 1:
The system performs preliminary detection and recording of object identification information before occlusion occurs. This advance recording ensures that even though real-time detection continues at high speed, the identification precision is maintained by referring to the pre-recorded data when objects reappear from the hidden area.
Solution Approach 2:
The system implements a feedback mechanism where detected objects are compared against stored identification records. When an object reappears after occlusion, the system uses feedback from the stored records to verify and restore the correct identification number, ensuring precise identification despite rapid detection cycles.
Data Source
Figure 1~2A
Figure 2B~2C
Figure 3
AI summary
In a case where multiple objects detected by an object detector (10) enter a hidden area calculated by a hidden area calculator (32), a positional relationship estimator (33) estimates a positional relationship between the multiple objects in the hidden area based on: information about each of the multiple objects; and lanes and routes into which each of the multiple objects may change their courses, and which are calculated by a route calculator (34). When the objects come out of the hidden area, an identification number assigner (35) determines identification information about each object based on the positional relationship estimated by the positional relationship estimator (33).