Hockey Puck Tracking via Dual-Process Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Automating the tracking of a hockey puck in sports games is challenging due to its small size, similar shape to other objects, high speed, and occlusion by players and sticks, making it difficult for camera operators and broadcast systems to accurately capture and broadcast the game while also hindering the collection of reliable gaming statistics.
Innovation Solution
A system and method that utilize a combination of image processing techniques, including correlation filters, shape comparators, and Kalman estimators to track the puck through a first object location process and a second object location process, which includes controlled and free movement tracking routines, to handle occlusions and shape changes, ensuring accurate puck detection and tracking even in complex environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a single object location process is used to track the puck, then the system complexity is low, but the tracking reliability fails when the puck is occluded or moves at high speed
Solution Approach 1:
The tracking system is segmented into two distinct object location processes: a first process for normal tracking conditions and a second process for difficult conditions (occlusion, high speed). This segmentation allows each process to be optimized for its specific purpose, improving overall reliability without requiring a single overly complex system to handle all scenarios.
Solution Approach 2:
The system dynamically switches between the first and second object location processes based on tracking conditions. When the puck becomes occluded or moves at high speed, the system transitions from the first process to the second process, adapting to changing conditions to maintain reliable tracking.
2Productivity
If automated tracking is implemented, then broadcast operational costs are reduced, but the difficulty of detecting and measuring the puck increases due to its small size and similar shape to other objects
Solution Approach 1:
The system utilizes color information to distinguish the puck from other objects. By analyzing color characteristics and changes in the video feed, the tracking system can identify and follow the puck even when its small size and similar shape to other objects make detection difficult.
Solution Approach 2:
The system employs intermediate processing steps including template matching and feature extraction that serve as mediators between the raw video data and the final tracking decision. These intermediary processes help overcome the difficulty of detecting the small puck among similarly shaped objects.
3Speed
If the first object location process is used continuously, then the processing speed is maintained, but the measurement precision decreases when the puck is occluded by players or sticks
Solution Approach 1:
The system incorporates feedback mechanisms that monitor tracking quality and occlusion conditions. When occlusion is detected, the feedback loop triggers a switch to the second object location process, which is optimized for maintaining precision under occlusion conditions while managing processing speed.
Solution Approach 2:
The system dynamically adjusts its processing approach based on occlusion conditions. During normal conditions, the first process maintains high processing speed. When occlusion is detected, the system transitions to the second process that prioritizes measurement precision through alternative detection methods.
Data Source
AI summary
A system and method for tracking an object comprising: an image input gateway arranged to receive a series of image frames from a stream of images, wherein the stream of images includes representations of the object being spatially displaced over the series of image frames; an object tracker arranged to process the series of image frames to track the object with a first object location process arranged to locate the object in the series of image frames; and, when upon the object is unable to be located by the first object location process in any of the image frames within the series of image frames, use a second object location process to locate the object in the image frames or subsequent image frames for tracking the object through the remaining series of image frames.