Development and statistical validation of an image processing system for total ammonia nitrogen monitoring in aquaculture

Aquaculture has become essential for global food security and economic growth, but its productivity relies heavily on effective water quality management. Total ammonia nitrogen (TAN) concentration is a key parameter, as high concentrations can harm aquatic life and disrupt cultivation systems. Conventional TAN measurement methods including colorimetric tests, spectrophotometry, electrochemical techniques, and biological assays offer high analytical accuracy but remain costly, complex, and often inaccessible for small-scale farmers. To overcome these limitations, this study developed a low-cost, TAN monitoring prototype using image processing techniques integrated with a Raspberry Pi microcontroller. Standard TAN solutions (0-5 mg/L) were prepared to construct regression models based on red (R) green (G) blue (B) color values extracted from captured images. Validation experiments showed that although both the R and G channel models provided strong predictive capability, the G channel model demonstrated superior practical performance. The G channel achieved the coefficient of determination (R²)=0.969, mean absolute error (MAE)=0.535 mg/L, root mean squared error (RMSE)=0.805 mg/L, a sensitivity of 1.397, and a limit of detection (LOD) of 0.215 mg/L, delivering stable and consistent predictions in the concentration range of 0.2-2.0 mg/L. In contrast, the R channel model, despite achieving higher numerical accuracy, exhibited greater drift at higher TAN concentrations, reducing its reliability in operational settings. The prototype system displays TAN values on a liquid crystal display (LCD) screen and sends notifications via the LINE application, enabling aquaculturists to respond promptly to potential water quality risks. By offering an affordable, accessible, and user-friendly alternative to conventional analytical methods, the proposed system supports more effective water quality management and contributes to the sustainable development of aquaculture practices.