Design of a Low-cost Universal Color Sensor to Support Rural Healthcare

Medical tests such as urinalysis, and other colori-metric assays require the measurement of changes in color. For such applications, this paper presents an innovative, low-cost color sensor, based on reflective color sensing. The proffered system is constituted of a light-intensity sensor and an RGB (red, green, blue) LED, placed on strategic geometries. Various regression models for calibration of the color sensor were investigated. Using cross validation with a training set of 160 colors, a cubic-spline regression model was selected. The accuracy and robustness of the novel sensor are also discussed.