Tea category identification based on optimal wavelet entropy and weighted k-Nearest Neighbors algorithm

摘要: Tea category classification is of vital importance to industrial applications. We developed a tea-category identification system based on machine learning and computer vision with the aim classifying different tea types automatically accurately. 75 photos three categories were obtained 3-CCD digital camera, they are green, black, oolong. After preprocessing, we 7 coefficient subbands using 2-level wavelet transform, extracted entropies from as features. Finally, weighted k-Nearest Neighbors algorithm was trained for classification. The experiment results over 5 × 5-fold cross validation showed that proposed approach achieved sensitivities 95.2 %, 90.4 %, 98.4 %, oolong, black tea, respectively. an overall accuracy 94.7 %. average time identify new image merely 0.0491 s. Our method accurate efficient in identifying categories.