AlphaBit - OpenML

Getting Started

1. Visual Studio Code

1. Download Visual Studio Code. (Recommended)

Download

2. Download Python 3.7 or newer (tested on Python 3.8/3.9/3.10/3.11)

Download

3. Configure Visual Studio Code & Terminal

1. Open Folder (This is the folder where you want to have the OpenML project)

2. Open New Terminal

3. Install the dependencies

4. Dependencies (use pip3 for python3)

pip install opencv-python ultralytics numpy

pip install opencv-python==4.8.0.76
											pip install ultralytics==8.0.196
											pip install numpy==1.24.4

If OpenCV does not work try to reinstall it using the method above (or retry using the methods below)

pip install opencv-python

Example

5. Hardware Requirements

A Webcam connected to the device is required (the code uses cv2.VideoCapture(0), make sure camera index 0 is correct for your configuration.)

Sufficient CPU power for real-time inference (a CUDA-enabled graphics card is recommended for better performance) for Testing and Control Hub is also accepted.

6. Test Camera Python Script

Create a Python script named camera_test.py and add the following code to it

camera_test.py (Click to download)

import cv2
	cap = cv2.VideoCapture(0) #If it doesn't work, increment the number by 1 until the camera works and appears on the screen
	ret, frame = cap.read()
	print("Camera working:", ret)
	cap.release()

Then

python camera_test.py

7.If the camera works, download the ML model from Resources

If you have a Very Low / Low quality camera download the first ML

If you have a Medium / Very Good quality camera download the second ML

Make sure you download the model with [Python Testing] for your camera quality [It matters a lot]

8. Calibrate The Camera

Access Camera Calibration, and then return after you have finished calibrating the camera.

9.🎉If you have successfully completed all steps, you can proceed to Python Code For Detection to test the OpenML model 🎉

Access Python Code For Detection

10. Additional Notes

The code uses math and cv2 modules for geometric calculations and camera operations. These are included in the dependencies mentioned above.

If you encounter CUDA-related errors, ensure you have compatible GPU drivers and PyTorch/CUDA is installed (Ultralytics YOLO usually handles this automatically).

Adjust the fov_degrees, first_angle and y values based on your camera calibration.

Support -> Discord

Choose Language / Alege Limba

Choose Detection Method