Prediction of compost organic matter via color sensor.

ABSTRACT

Composted materials serve as an effective soil nutrient amendment. Organic matter in compost plays an important role in quantifying composted materials overall quality and nutrient content. Measuring organic matter content traditionally takes considerable time, resources, and various laboratory equipment (e.g., oven, muffle furnace, crucibles, precision balance). Much like the quantitative color indices (e.g., sRGB R, sRGB G, sRGB B, CIEL*a* b*) derived from the low-cost NixPro2 color sensor have proven adept at predicting soil organic matter in-situ, the NixPro2 color sensor has the potential to be effective for predicting organic matter in composted materials without the need for traditional laboratory methods. In this study, a total of 200 compost samples (13 different compost types) were measured for organic matter content via traditional loss-on-ignition (LOI) and via the NixPro2 color sensor. The NixPro2 color sensor showed promising results with an LOI-prediction model utilizing the CIEL*a* b* color model through the application of the Generalized Additive Model (GAM) algorithm yielding an excellent prediction accuracy (validation R2 = 0.87, validation RMSE = 4.66 %). Moreover, the PCA scoreplot differentiated the three lowest organic matter compost types from the remaining 10 compost types. These results have valuable practical significance for the compost industry by predicting compost organic matter in real time without the need for laborious, time-consuming methods.