In this project we see how to build a device that detects maturation stages based on color with a neural network model. As fruits and vegetables ripen, they change color due to the four families of pigments: chlorophyll (green), carotenoids (yellow, red, orange), flavonoids (red, blue, purple), betalain (red, yellow, purple).
These pigments are groups of molecular structures that absorb a specific set of wavelengths and reflect the rest. Unripe fruits are green due to the chlorophyll in their cells. As they mature, the chlorophyll breaks down and is replaced by orange carotenoids and red anthocyanins. These compounds are antioxidants that prevent the fruit from spoiling too quickly in the air.
After doing some research on color change processes during fruit and vegetable ripening, we decided to build an artificial neural network (ANN) based on the classification model to interpret the color of fruit and vegetables and predict ripening stages.
Before building and testing the neural network model, we developed a web application in PHP (running on a Raspberry Pi 3B+) to collect the color data generated by the AS7341 visible light sensor and create a dataset on the maturation stages . We used an Arduino Nano 33 IoT to send the produced data to the web application.
After completing the dataset, we built the artificial neural network (ANN) with TensorFlow.