Ripe Stage Detection and Optimal Ripe Hours Prediction of a Banana Using Modified Fractional Order Colpitts Oscillator

Accurate assessment of ripe-stages and reliable prediction of ripe hours of banana will ensure efficient supply chain management, extended shelf-lives, reduction of food waste, and optimal nutrients to the consumers. Bioimpedance spectroscopy (BIS) is an accurate and intrinsic method to track banana ripeness but not suitable for a commercial portable meter. This work presents a compact and reliable solution to this challenge by designing a modified fractional-order Colpitts oscillator (FOCO) as a novel electronic sensor to ripeness detection. When a banana sample is connected to the proposed circuit its frequency output indicates the ripe stage of the banana: “Green,” “Ripe,” “Overripe,” and “Decay.” In addition to that, the proposed sensor includes another novel feature; it uses a time-series model to predict “hours to Ripe” and “hours to Overripe” for a banana in Green stage, from the same FOCO output. The model is trained with 540 banana sample data and validated with 180 test data. The work reports 94.4% ${F}1$ score in ripe stage identification and 82.5% ${F}1$ score in the ripe hour prediction.