General idea is to calculate only the direction of the object ball (3) and projecting that on the table (7). To make this work you need the point of impact (6) of the cue-ball. This can be done by calculating the position of the cueball (2) when going near the object ball. We need to use intersection of the path of the cue-ball with the object-ball, where a virtual object-ball (4) is defined by twice the ratio from the center (3) of the object-ball. The direction of the shot can be determined by sticking red tape on the tip of the cue-sticki (1). To detect the cue-ball we select the whitest ball. Calculate the shot-line using the center of the cue-ball (2) to and the center of the cue tip (1).
- Install Python
- Install the following packages
pip install cv2
pip install imutils
pip install mahotas
brew install freeimageBasic circle detection based on 7 non-solid circles
python Werkende-cirkels.pyDetect balls in a high res stock photo
python pool-table-stock-werkend.pyCircle detection:
- https://docs.opencv.org/3.3.1/da/d53/tutorial_py_houghcircles.html
- https://www.pyimagesearch.com/2014/07/21/detecting-circles-images-using-opencv-hough-circles/
Binarize images:
Optimizing images
- Contrast and Brightness: https://docs.opencv.org/2.4/doc/tutorials/core/basic_linear_transform/basic_linear_transform.html