Search

In this paper, we present an efficient technique of electronic tongue (ET) signal processing using the sparse decomposition method for the prediction of tea quality. Sparse model coefficients of an under complete dictionary are considered as characteristic attributes of the ET signals obtained for characterization of tea samples. The effectiveness of the proposed method is established by performing experiment on three different types of pulse voltammetry. The ability of the sparse coefficients, as features of ET signal, is tested by employing three different classifiers. High classification accuracy of all the classifiers on three types of voltammetric measurement data validates the usefulness of this technique. The present model-based the tea quality prediction method is very fast, simple, and straightforward. This tea quality prediction method is able to address the problem more efficiently in comparison to other techniques found in the literature.

Liquor characteristics of black cut, tear, and curl tea mostly depend on two biochemical components like theaflavin (TF) and thearubigin (TR). Evaluation of tea quality can be done efficiently by estimating the concentration of TF and TR without using biochemical tests as it takes much time, which requires laborious effort for sample preparation, storage, and measurement. Moreover, the required instruments for this test are very costly. In this paper, we have proposed an efficient method of TF and TR prediction in a given tea sample using electronic tongue (ET) signal. Combinations of transformed features, like discrete cosine transform, Stockwell transform (ST), and singular value decomposition, of ET signals are fused to develop regression models to predict the contents of TF, TR, and TR/TF. Three different regression models such as artificial neural network, vector-valued regularized kernel function approximation, and support vector regression are used to evaluate the performance of the proposed method. High prediction accuracy using fusion of features ensures the effectiveness of the proposed method for prediction of TF and TR using ET signals.

A voltammetric electronic tongue (E-tongue) has been used to detect the optimum fermentation time (OFT) of crush tear curl black tea. An E-tongue with an array of five noble metal working electrodes has been developed for this purpose. A suitable large amplitude pulse voltammetric waveform is used and a simple feature extraction technique based on the charge activity of working electrodes is proposed. The principal component analysis reveals a significant correlation between the E-tongue responses and the biochemical changes during various stages of fermentation. Principal component regression models are developed to assess the effectiveness of the features derived from E-tongue response to model the temporal variations of theaflavins, a key quality compound of tea, produced during the fermentation process. The OFTs detected by the E-tongue correlate by a factor of 0.97 with the estimates obtained from an ultraviolet-visible spectrophotometer-based reference method.

Electronic tongue (ET) system is under extensive development for automatic analysis and prediction of quality of different industrial end products. Each sensor in an ET system generates a specific electronic response in presence of different organic or inorganic compounds in the sample. The vital part of the ET system is the discrimination of the complex pattern generated by the sensor array. In this paper, a novel technique of black tea quality estimation is using the ET signals. A moving window is used to extract discrete wavelet transform coefficients from the transient response of ET. The energy in different frequency bands are used as the features of the ET signal for different positions of the window. The prediction of a new sample is performed by the highest score obtained by a particular class by testing all the patterns generated by windowing ET signal. The performance of the proposed technique is verified to estimate black tea quality using two kernel classifiers, namely support vector machine and recently proposed vector valued regularized kernel function approximation method. High prediction accuracy of both the classifiers confirms the effectiveness of the proposed technique of tea quality estimation using ET signals.

Tea quality assessment is a difficult task because of the presence of innumerable compounds and their diverse contribution to tea quality. As a result, instrumental evaluation of tea quality is not practiced in the industry, and tea samples are assessed by experienced tea tasters. There had been a very few reports where an electronic tongue has been used for the discrimination of taste of tea samples. In this paper, a voltammetric electronic tongue instrument is described, which can declare tea-taster-like scores for black tea. The electronic tongue is based on the principle of pulse voltammetry and consists of an array of five working electrodes along with a counter and a reference electrode. The five working electrodes are of gold, iridium, palladium, platinum, and rhodium. The voltage equivalent of the output current from between the working electrode and the counterelectrode generated out of the tea liquor when excited with pulse voltage between the working electrode and the reference electrode has been considered for data analysis. First, the sampled data have been compressed using discrete wavelet transform (DWT) and are then processed using principal component analysis (PCA) and linear discriminant analysis (LDA) for visualization of underlying clusters. Finally, different pattern recognition models based on neural networks are investigated to carry out a correlation study with the tea tasters' score of five different grades of black tea samples obtained from a tea garden in India. The efficacy of the classifier has been established using tenfold cross-validation methods.

Commonly used classification algorithms are not capable of incremental learning. When a new pattern is presented to such a computational model, it can either classify the unknown pattern based on its legacy training or declare the pattern as an outlier if such a provision is built into the associated algorithm. In the case of the pattern being an outlier to the existing training model, it is desirable that the same could be seamlessly included in the training model with appropriate class labels so that a universal computational model may be evolved incrementally. To this end, classifiers having the incremental-learning ability can be of great benefit by automatically including the newly presented patterns in the training data set without affecting class integrity of the previously trained system. In the present treatise, an incremental-learning fuzzy model for classification of black tea using electronic nose measurement is proposed. For application in black tea grade discrimination, an attempt has been made to correlate the multisensor aroma pattern of electronic nose with sensory panel (tea tasters) evaluation. However, this problem is associated with 2-D complexities. On one hand, the aroma of tea depends on the agroclimatic condition of a particular location, the specific season of flush, and the clonal variation for the tea plant. On the other hand, the sensory evaluation is completely human dependent that often suffers from subjectivity and nonrepeatability. In our pursuit of developing a universal computational model capable of objectively assigning tea-taster-like scores to tea samples under test, it has been felt that an incremental approach could be extremely beneficial for electronic-nose-based tea quality estimation. To this end, the proposed incremental-learning fuzzy model promises to be a versatile pattern classification algorithm for black tea grade discrimination using electronic nose. The algorithm has been tested in some tea gardens of northeast India, and encouraging results have been obtained.

Tea is an extensively consumed beverage worldwide with an expanding market. The major quality attributes of tea are flavor, aroma, color, and strength. Out of these, flavor and aroma are the most important attributes. Human experts called ldquotea tastersrdquo conventionally evaluate tea quality, and they usually assign scores to samples of tea that are under evaluation on a scale of 1 to 10, depending on the flavor, the aroma, and the taste of the sample. This paper presents a study where, first, the selection of appropriate sensors was carried out based on sensitivity with the major aroma-producing chemicals of black tea. Then, this sensor array was exposed to black tea samples that were collected from the tea gardens in India, and the computational model has been developed based on artificial neural network methods to correlate the measurements with the tea taster's scores. With unknown tea samples, encouraging results have been obtained with a more than 90% classification rate.