In this work, a novel approach that can calibrate the colors obtained with a cellphone camera was proposed for the colorimetric sensor array. The variations of ambient light conditions, imaging positions and even cellphone brands could all be compensated via taking the black and white backgrounds of the sensor array as references, thereby yielding accurate measurements. The proposed calibration approach was successfully applied to the detection of glucose in urine by a colorimetric sensor array. Snapshots of the glucose sensor array by a cellphone camera were calibrated by the proposed compensation method and the urine samples at different glucose concentrations were well discriminated with no confusion after a hierarchical clustering analysis. Display omitted •The glucose in urine was detected using a colorimetric sensor array.•A calibration approach for cellphone camera imaging was proposed.•Snapshots of the glucose sensor array by a cellphone camera were calibrated.•Glucose concentrations were well discriminated with no confusion.