Queue Monitoring Using Depth Detection for Accurate Object Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for monitoring objects in a 3D space, such as queues, are complex and unreliable due to the need for calibration and mapping of 3D spaces to 2D images, which becomes difficult especially when objects change size or move, leading to decreased accuracy and reliability in tracking.
Innovation Solution
A system using a depth detection device like a TOF camera that receives image information with depth data, calculates actual object heights, filters cluster information, and tracks objects by comparing patterns, allowing for accurate recognition and tracking without mapping a 3D space to an image, thereby simplifying the process and improving reliability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If 3D space mapping to 2D image is used for object monitoring, then object recognition is possible, but system complexity increases significantly due to calibration requirements
Solution Approach 1:
The patent extracts and removes the complex 3D-to-2D mapping calibration process from the system. Instead of requiring camera calibration parameters (height, angle, FOV) and complex spatial mapping, the system directly uses 2D image coordinates with simple height threshold filtering to recognize and track objects, thereby eliminating the complicated calibration subsystem while maintaining object recognition capability
Solution Approach 2:
The patent inverts the traditional approach by not mapping 3D space to 2D image, but rather working directly in 2D image space with height information. Instead of converting 3D coordinates through complex calibration, the system uses 2D image coordinates combined with height threshold comparison to achieve object recognition, reversing the conventional mapping direction and simplifying the process
2Reliability
If 3D space mapping is used to track moving objects, then object tracking is possible, but tracking reliability decreases when objects change size or position
Solution Approach 1:
The patent implements dynamic object tracking by continuously updating object positions in the queue based on current image frames. The system dynamically adjusts tracking by comparing current frame object positions with previous positions, using height threshold filtering to maintain accurate object identification even as objects move, change size, or reposition within the queue
Solution Approach 2:
The patent changes the approach from fixed 3D spatial parameters to dynamic 2D image parameters. Instead of relying on fixed calibration parameters that become invalid when objects move, the system uses dynamic height threshold comparison in 2D image space, allowing accurate tracking regardless of object position or size changes in the queue
3Measurement precision
If detailed calibration parameters are set for each object, then object recognition precision improves, but system operation becomes more difficult
Solution Approach 1:
The patent implements self-service by automatically determining object heights and positions through simple height threshold filtering in 2D image space. The system does not require manual calibration parameter input or complex setup procedures; instead, it automatically processes image data and applies height-based filtering to recognize and track objects, making the system easy to operate without specialized calibration knowledge
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enables accurate recognition and continuous tracking of objects in a queue, even when they are close or moving, by using depth information to determine actual heights and queue patterns, enhancing the system's reliability and efficiency in monitoring queue situations.
Implementation Method 1
an image receiving unit that receives image information including depth information of objects and a background of a photographing space through a depth detection device
Data Source
AI summary
The present invention relates to a system and a method of monitoring a queue that allow for exactly recognizing objects in a queue and exactly monitoring the situation of the queue by tracking the recognized objects, using depth detection device such as a TOF camera. According to the present invention, it is possible to map a space and an image using depth information provided through a depth detection device without mapping a 3D space to an image taken by a camera and exactly measure the actual heights of objects from the ground using depth information of the objects and the background and initial parameters of the depth detection device so that desired objects can be exactly detected and recognized in filtering. Accordingly, the system can be less complicated. Further, waiting time is assigned to recognized objects so that they can be easily tracked.