This project is a computer vision application to track a ball's trajectory and detect its bounces relative to the boundaries of a sports field. The system is implemented in Python and utilizes OpenCV, numpy, and imutils for image processing, with real-time video input handling.
-
Real-Time Ball Tracking: Detects and tracks a green ball in video streams using HSV color space.
-
Predictive Tracking: Employs a Kalman filter to estimate the ball's position, enhancing the stability of tracking.
-
Dynamic Frame Adjustments: Resizes frames adaptively based on the ball's distance from the center.
-
Contour and Shape Analysis: Uses contour approximation to ensure accurate detection of circular objects.
-
Customizable Parameters: Allows fine-tuning of HSV color ranges, tracking buffer size, and contour thresholds.
-
In the project there is also a test-video used to see how it works not in webcam.
- Python 3.6+
- OpenCV
- numpy
- imutils