Entropy-Aided Assessment of Amla (Emblica officinalis) Quality Using Principal Component Analysis

The texture analysis-based scheme for identifying non-consumable amla fruit (Emblica officinalis) samples is proposed here. The method uses entropy analysis to detect wrinkles and irregularities developed on the fruit surface with progression in time. Since entropy is one of the major tools used to detect the randomness of data, it is used here to identify these surface irregularities, which are almost absent in fresh samples. Based on these features, the edibility of the samples is predicted. Principal component analysis (PCA) further analyzes these entropy features to enhance the most important directions of variations, followed by a threshold-based segmentation scheme to detect rotten samples. The method possesses less computational burden as it applies PCA and entropy only; it is highly efficient to yield a high detection accuracy of 93.33%; hence, it is easy for real-life implementation.