Queue Counting via Seed Location Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing queue counting systems struggle to accurately count objects waiting in a single queue for multiple electronic points of service, as they often incorrectly include individuals already being served or fail to adapt to situations where multiple service locations share a single queue line.
Innovation Solution
A system that determines an initial seed location for a single queue and adjusts based on operational statuses of electronic points of service (EPOSs) to identify the actual queue start point, excluding those already being served, and assigns objects exhibiting queuing behavior to the correct queue, using an object tracking system and processor system to count only waiting objects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing queue learning algorithms are used to learn queue paths through a service point area, then queue path recognition is achieved, but the system cannot accurately discriminate between customers waiting to be served and those already in transaction
Solution Approach 1:
The system segments the service point area into multiple service locations, each with its own seed location. The queue path learning algorithm is applied separately for each service location, allowing accurate discrimination of customers waiting at each specific EPOS while maintaining the ability to handle multiple service locations simultaneously.
Solution Approach 2:
The patent introduces seed locations as intermediary reference points at each service location. These seed locations serve as mediators between the tracking system and the queue counting logic, enabling the system to accurately identify and count only those customers waiting to be served at each specific EPOS, rather than all customers in the service point area.
2Measurement precision
If a separate queue is created for each electronic point of service, then queue tracking at individual EPOSs is achieved, but the system is not applicable to situations where multiple service locations are fed from a single queue line
Solution Approach 1:
The system implements a universal queue management approach where each service location has its own seed location and queue learning algorithm, but these can be configured to share a common queue path. This multi-functional design allows the same system architecture to handle both separate queues for each EPOS and single queues serving multiple EPOSs, providing versatility without sacrificing counting accuracy.
3Quantity of substance
If all objects in the service point area are counted, then comprehensive object tracking is achieved, but customers already being served are incorrectly included in the queue count
Solution Approach 1:
The system applies local quality by defining specific seed locations at each service location and learning queue paths localized to each EPOS. This allows the system to count only objects exhibiting queue-like behavior relative to each specific service location, automatically excluding customers already being served at other EPOSs while maintaining comprehensive tracking capability across the entire service point area.
Data Source
AI summary
A system for counting objects in a queue waiting for service at one of a plurality of electronic points of service “EPOSs” includes an object tracking system and a processor system for processing data. The processor system may have memory means storing data indicative of the positions of each of said EPOSs within a service point area, and be adapted to receive data from each EPOS indicating its operational status and to receive from said tracking system data relating to positions of objects within the service point area. The method of counting objects in a queue comprises firstly determining an initial seed location as a point at which a single queue of objects forms. Then, determining a new seed location as a point at which the single queue forms Once the new seed location is determined, tracked objects may be assigned to the single queue and then counted.