Kinematic-Pass: An Open-Source HSV-Based Ball Tracking and Trajectory Mapping System for Skill Assessment
Description
Abstract: This dataset encompasses the cross-platform source code, algorithmic tracking logic, and interface architecture for an automated computer vision framework engineered to evaluate motor skills and frequency metrics in table tennis through the standardized wall-pass test. Developed using a decoupled dual-stack layout—utilizing a high-performance Python back-end powered by OpenCV and a responsive web-based HTML5/JavaScript dashboard—this system automates the diagnostic evaluation of athletic coordinates without laboratory-bound sensors. The computational core leverages real-time color segmentation within the Hue-Saturation-Value (HSV) color space combined with morphological noise filtration (erosion/dilation) to isolate ball dynamics based on high-circularity contour parameters. Concurrently, a localized pixel threshold matrix defines the dynamic "wall target zone." The tracking engine utilizes a bounded deque memory buffer to reconstruct fading 2D spatial trajectories while applying mathematical cross-over logic to count successful passes as the ball centroid bisects the calibrated target quadrants. This data package includes complete backend multi-color tracking configurations, configuration scripts, and styling templates, offering a foundational toolkit for sports scientists, performance analysts, and table tennis coaches looking to digitize reactive training paradigms. Keywords: Computer Vision in Sports, HSV Color Segmentation, Object Tracking, Table Tennis Analytics, Wall Pass Test
Files
Institutions
- Sabaragamuwa University of Sri LankaSabaragamuwa Province, Ratnapura