How it works
Chemical Sensors
-
Filter paper strips impregnated with anthocyanin and curcumin solutions;
-
Glycerol added to retain moisture and improve the color-change process;
-
A 3D-printed PLA frame to keep the filter paper strips with indicators in a fixed position.
Computer Vision Algorithm
-
ArUco markers printed on the sensor frame for automatic localization;
-
Real-time video stream analyzed in the HSV color space (Hue–Saturation–Value);
-
The algorithm extracts the average hue over a predefined area of the sensor.
Classification
Based on calibrated thresholds, the system assigns one of three freshness states:
-
FRESH – low pH, pigments within their initial color range
-
WARNING – intermediate pH, visible change in hue
-
SPOILED – high pH, strong color change, confirmed especially by curcumin
​​
The result is displayed live on the screen, transforming natural color changes into a clear digital signal.

Experimental Results
-
In a sealed testing chamber with controlled ammonia vapors, the sensors showed clear and repeatable color transitions.
-
Anthocyanin responded in approximately 15–30 seconds, transitioning from fresh to warning/spoiled.
-
Curcumin showed changes in about 10–20 seconds, confirming spoilage.
-
Graphs of digital hue versus time/pH showed a smooth, nearly linear trend, indicating reliable freshness tracking.