Edgefleet

Cricket Ball Detection and Trajectory Tracking

This project detects the centroid of a cricket ball in each frame of a video captured from a static camera. It generates a processed video with the ball trajectory and a CSV file containing per-frame annotations.

---

Features

• Detects cricket ball using color-based segmentation.
• Tracks the ball using temporal consistency.
• Generates trajectory overlay on output video.
• Produces per-frame centroid annotations.

---

Requirements

Install the required libraries:

pip install opencv-python numpy pandas

---

Input

• Cricket video from a static camera.
• Supported format: .mp4

If your video is in .mov format, convert it:

ffmpeg -i input.mov input.mp4

---

How to Run

1. Place your video file in the same directory as the script.

2. Open the Python script and set:

video_path = "your_video.mp4"

3. Run the script:

python cricket_ball_detection.py

---

Output

For each video, the following files are generated:

• output_with_trajectory.mp4 → Processed video with ball trajectory

• ball_annotations.csv → Per-frame ball centroid data

CSV format:

frame_index, x_centroid, y_centroid, visibility_flag

If the ball is not visible:

x = -1, y = -1, visibility_flag = 0

---

Multi-Video Processing

To process multiple videos, place them in a folder and modify the main script to loop through all files.

---

Author

Dhana Lakshmi AI22BTECH11012