Comparing ACES AP0 and AP1 color primaries for smartphone colorimetry in detecting soil CO₂ leakage with a beetroot...

Effective monitoring of CO₂ leaks from carbon capture and storage (CCS) sites is vital to mitigate environmental risks, including soil acidification and ecosystem disruption. This study presents a novel, sustainable biosensor based on betanin from Beta vulgaris (beetroot) extract for pH-responsive colorimetric detection of soil CO₂ leakage, coupled with smartphone imaging for quantitative analysis. Central to the investigation is a comparative evaluation of Academy Color Encoding System (ACES) AP0 and AP1 primaries against sRGB and DaVinci Wide Gamut color spaces, highlighting AP1's advantages in perceptual uniformity, reduced metamerism, and enhanced computational efficiency for resolving subtle color gradients (effective LOD ~ 0.015 wt% or ~150 ppm using PLS on G-channel), while AP0's expansive virtual gamut excels in archival fidelity but introduces practical artifacts, resulting in higher effective LOD (~0.017 wt% or ~170 ppm). Sensors were fabricated with dialysis membranes, calibrated in a simulated chamber across 0-4.0 wt% CO₂ levels, and imaged using a Realme 9 smartphone. RGB data were extracted via Python libraries (Pillow, OpenCV, NumPy, Colormath, Colour-Science), with machine learning models (partial least squares (PLS), multiple linear regression (MLR), and support vector regression (SVR)) optimized through K-fold cross-validation. The optimal ACES AP1 PLS model on the green channel demonstrated superior performance (R² = 0.9984, RMSE = 0.0519, MAPE = 17.80%), achieving <5% deviation from UV-Vis benchmarks (base LOD ~ 0.0024 wt% or ~24 ppm) during field validation at an Indonesian CCS facility. This low-cost, portable system advances CCS surveillance by integrating natural biosensors with advanced colorimetry, offering scalable, real-time solutions for environmental monitoring.