Development of a Fast Universal Ph Paper Value Predictor Using Computer Vision and Random Forest

Accurate <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{p H}$</tex> measurement is critical across scientific and industrial domains, but traditional methods (pH meters/paper) face limitations. This study develops a lightweight, Computer Vision-based system for automated pH prediction using universal <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{p H}$</tex> paper. The proposed method captures standardized images of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{p H}$</tex> paper, extracts multi-colorspace features (RGB, HSV, Normalized), and trains a Random Forest model optimized for accuracy and speed. Evaluated via cross-validation on a custom dataset, the system achieves high performance (MAE, RMSE, <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{R}^{\mathbf{2}}$</tex> metrics) with low computational cost. Unlike prior Computer Vision approaches focused on generic color recognition, this work specifically targets <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{p H}$</tex> prediction, offering real-time applicability in laboratories, environmental monitoring, and education. Results demonstrate a practical, cost-effective alternative to conventional methods, balancing precision and efficiency.