Machine learning & computer vision-based optimum black tea fermentation detection
Anuja Bhargava,Atul Bansal,Vishal Goyal,Aasheesh Shukla
DOI: https://doi.org/10.1007/s11042-023-15453-3
IF: 2.577
2023-05-03
Multimedia Tools and Applications
Abstract:The world's high-value crop is tea, its aspect plays a powerful role in its marketability. Tea is the utmost extensively absorbed aromatic beverage with the legion of health benefit and respond as a remedy for various disease like neurological disorder and cardiovascular. The identification of tea fermentation is onerous due to the heavy production of products. The manual investigation is expensive, laborious, and inconsistent. Thus, an automated machine learning-based algorithm is proposed for the grading of tea fermentation (fermented, over-fermented, under-fermented). Firstly, images are pre-processed by Gaussian filtering to enhance the quality of the image and removing of noise. Then various features namely, statistical, textural, color, geometrical, laws texture energy, the histogram of gradients, and discrete wavelet transform are extracted (150) and selected from feature vector by PCA. Lastly, k-NN, SRC, and SVM are used to make decisions for the detection of tea fermentation levels. The performance of the system has been validated by the k (10) fold cross-validation technique. The proposed algorithm achieves 87.39% (k-NN), 89.72% (SRC), and 98.75% (SVM) for tea fermentation level detection. The proper feature selection shows the enhanced performance of the system. Among three different classifiers, SVM shows more efficient results that are promising and comparable with the literature. This paper also includes the analytical comparison of distinct approaches proposed by the different researchers for tea fermentation level detection. This potential fermentation level detection may guide the detection of tea products which further promotes the development of the food industry.
computer science, information systems, theory & methods,engineering, electrical & electronic, software engineering