Abstract
Susceptibility to mechanical injury and fast decay rates are currently two main problems of litchi fruit after harvesting. To achieve better postharvest management of litchi fruit, this study aimed to find an effective method of litchi fruit supervision during the circulation process that included mechanical injury detection and storage quality detection. For mechanical injury detection, injury-free litchis without any treatment and litchis with mild and severe mechanical injuries were dropped from 80 and 110 cm high, respectively. The electronic nose (E-nose) response, total soluble solid (TSS), and titratable acidity (TA) of samples were tested on days 0, 1, 2, 3, 4, and 5 after injury at room temperature. For storage quality detection, normal litchis were stored in a cold environment. The E-nose response, TSS, and TA of samples were tested on storage days 0, 3, 6, 10, 15, 19, and 24. The experimental results showed that mechanical injury not only accelerated pericarp browning but also accelerated flavor (TA and TSS) loss. The browning index quickly increased during storage, and the TSS and TA of defect-free litchis changed only barely at room temperature and during cold environment storage. After feature extraction, mechanical injury of litchi can be well-detected by E-nose from day 1 to day 4 after injury. The best mechanical injury detection time of litchi fruit is at day 4 after injury under room temperature storage conditions. After singular sensor elimination and comprehensive feature extraction, the storage time and browning degree, but not TSS and TA, of litchi fruit can be detected by E-nose. E-nose data preprocessing should differ according to the litchi variety and detection target.
Litchi (Litchi chinensis) is a subtropical to tropical fruit with an attractive red appearance, great taste, and rich nutritive value that has been highly enjoyed by consumers worldwide for many years (Ali et al., 2016; Gong et al., 2014). However, postharvest litchi is very fragile, which is mainly indicated by its susceptibility to mechanical injury (Chen et al., 2014) and high decay rate (Zhang and Quantick, 1997). The fragility of postharvest litchi has been given much attention by researchers in past decades, but there are still many problems that require further research.
Mechanical injury primarily occurs during harvest and the transportation process, which has been given less attention. Due to the thin pericarp, thick flesh, and high water content of litchi fruit, cell rupture, cell disruption, and cell separation are easily caused by collision and extrusion (Chen et al., 2013a). Mechanical injury also opens a channel for pathogenic bacteria to enter more easily, which increases the decay rate of fruit. Litchi fruit with serious decay due to mechanical injury loses its commercial value and should be removed so they do not infect the surrounding litchis with their pathogenic bacteria. The influence of mechanical injury on litchi pericarp (Chen et al., 2013a, 2013b) has been analyzed, but how mechanical injury affects the storage quality of litchi is still unknown. Accurate detection of the initial mechanical injury to litchi is crucial for postharvest litchi management; however, it has not yet been reported.
In addition to mechanical injury, the natural decay rate of litchi fruit is incredibly fast after harvesting (Dharini et al., 2008). The red color of postharvest litchi pericarp rapidly fades and turns fully brown within a few days if stored at room temperature due to the degradation of anthocyanin in its pericarp (Hu et al., 2004). Although the current preservation technology can slow the decay rate of litchi to some extent (Khan et al., 2012), the reality of the fast decay of postharvest litchi fruit is still a problem. Fully brown litchi has worse resistance to pathogenic bacteria and almost zero commercial value. Therefore, a rapid and accurate litchi storage quality detection method should be developed for sellers to handle litchi fruit storage timely and accurately.
Fresh food with less storage time has an increasingly important role in consumer habits because of better standards of living (Elshiekh and Habiba, 1996). Because cold storage extends the storage life of litchi fruit, the outward appearance changes less, especially during the initial stage. Many consumers want to know the storage time of litchi.
At present, there are two traditional methods of litchi quality detection: the sensory detection method (Alves et al., 2011) and the physicochemical detection method (Huang et al., 2016). The sensor detection method evaluates the qualities of litchi fruit such as the pericarp color, flavor, and fragrance based on multiple human perceptive organs. The physicochemical detection method detects the total soluble solid content, titratable acidity, and weight using chemical analysis or physical measurements. The sensor detection method provides direct evaluation results from humans but is flawed because it is time-consuming, labor-intensive, and easily affected by human subjectivity. The physiochemical detection method is objective and accurate, but it is destructive, complicated, and time-consuming. Therefore, the traditional ways cannot meet the requirements of the progressing litchi industry. Even though machine vision (Xiong et al., 2011) and spectrum technologies (Xiong et al., 2018) have allowed intelligent and fast detection of many agricultural products, they are unsuitable for stored litchi quality detection because litchi fruits cover each other during storage.
The electronic nose (E-nose), also known as a bionic olfaction instrument, acquires sample information by mimicking the human olfactory system (Röck et al., 2008). An E-nose is usually composed of a sampling and cleaning channel, gas-sensitive sensor array, and pattern recognition subsystem. Furthermore, the E-nose has a sensor array that contains several gas sensors that are sensitive to different substances, which gives the entire sensor array the ability to detect simple and complex odors (Pearce et al., 2006). The E-nose is a portable tool that can detect sample quality easily, quickly, and intelligently. Compared with sensory and physicochemical detection methods, the E-nose can overcome the flaws associated with time and labor requirements, destruction, complications, and human subjectivity. Compared with the E-tongue, the E-nose can nondestructively detect characteristics of samples (Zhang and Tong, 2005). Compared with other machine detection methods like machine vision and spectrum, the E-nose can overcome the limit of the visual angle. Therefore, the E-nose is more suitable than other detection methods for litchi quality supervision.
Accordingly, this study applied an E-nose to detect the quality of litchis with mechanical injuries and normal litchis after harvesting to determine a feasible method of expanding litchi quality supervision during the postharvest circulation process. After E-nose sampling, browning indexes, total soluble solid content, and titrable acidity were recorded by sensory detection, a soluble solids refractometer, and acidity titration, respectively. The objectives of this research were to 1) to test the impact of mechanical injury and storage time on the quality of postharvest litchi; 2) to test the feasibility of using the E-nose to detect mechanical injury of litchi; and 3) to find an efficient way to detect the quality of litchi fruit during storage.
Materials and Methods
Litchi samples
Samples of ‘Guiwei’ litchis for mechanical injury detection experiments were harvested at Conghua litchi orchard (located in Guangzhou, China) at 80% to 90% maturity and then shipped to the laboratory within 2 h. After their stems and leaves were removed, litchi samples were divided into three groups: injury-free, mild injury, and severe injury. The injury-free group did not undergo any treatment. The mild and severe mechanical injury groups were dropped from heights of 80 and 110 cm, respectively. Injury was not apparent on the mildly mechanically injured litchis; however, the severely mechanically injured litchis had noticeable cracks on the pericarp. All litchi samples were packed in perforated polyethylene bags (300 mm × 200 mm × 0.05 mm; perforation ratio of 5%). There were 10 litchi samples in each bag, and the bags were stored at 25 °C (room temperature).
Samples of ‘Yuhebao’ litchis for storage quality detection experiments were also harvested at Conghua litchi orchard at 80% to 90% maturity and shipped to the laboratory within 2 h. All litchi samples were stored in a cold environment (5 °C and 90% humidity) after their stems and leaves were removed. Before storage, all litchi samples were pre-cooled in 4 to 5 °C ice water for 5 min (Ruan et al., 2012) and packed in perforated polyethylene bags (300 mm × 200 mm × 0.05 mm; perforation ratio of 5%); there were 25 litchi samples in each bag.
E-nose set-up
A portable commercial E-nose (PEN3; Airsense Inc., Schwerin, Germany) was used to perform sampling of volatile litchis. This E-nose is mainly composed of sampling and cleaning channels, a sensor array, and data collection and processing subsystems. The sensor array contains 10 metal oxide gas sensors that are sensitive to various volatiles, which makes the entire E-nose capable of detecting simple and complex odors. The parameters of those 10 sensors are shown as Table 1 (Cardozo and Londoño, 2013). The response data of each sensor were represented as G/G0, where G was the response value of the sensor contacting the sample volatile and G0 was the response value of the sensor contacting the zero gas (ambient air filtered through standard active carbon).
Parameters of sensors of the PEN3 electronic nose
Experimental sampling
E-nose sampling.
Each litchi sample comprised one litchi fruit in a 100-mL glass beaker that was sealed with double-layer preservative film. It was placed in the storage environment immediately to avoid decay during the volatile collection process. After 0.5 h, the E-nose was applied to test the headspace of each sample. Before each test, all sensors of the E-nose were cleaned and restored by zero gas. Litchi samples for mechanical injury detection were tested on days 0, 1, 2, 3, 4, and 5 for each mechanical injury group; four samples were used for each test day. Litchi samples used for storage quality detection were tested on days 0, 3, 6, 10, 15, 19, and 24; 20 samples were used for each test day.
Sampling of quality parameters.
where L is the browning degree of a single litchi, NL is the amount of litchis with the Lth grade, and M is the number of litchis in the batch.
After detecting the browning degree, each litchi sample was peeled to acquire the flesh, which was then homogenized and filtered to obtain the litchi juice to determine the total soluble solid content (TSS) and titratable acidity (TA). The TSS of litchi samples was tested using a soluble solids refractometer (PR-32α; ATAGO Inc., Tokyo, Japan), and the TA of litchi samples was determined as described by previous research (AOAC Official Method 942.15, 1965; Jiang et al., 2004). The TA was defined as the percentage of citric acid determined by titration with 0.1 M NaOH.
Data analysis methods.
The linear discriminant analysis (LDA) (Gorjichakespari et al., 2016) is one of the most commonly used classification procedures. This method maximizes the variance between categories and minimizes the variance within each single category. It usually has better classification ability than the principal component analysis and can show relationships among groups.
The K-nearest neighbors (KNN) method (Wang et al., 2017) is a nonparametric method based on the distance between objects in a space with a dimension equal to the number of variables explored. The class to which the sample is assigned is that of the samples in the training group closest to it. Only the objects closest to K are used to make the assignments.
The partial least-squares regression (PLSR) method (Tian et al., 2015) is a technique involving data containing correlated predictor variables. This technique constructs new predictor variables, known as components, as linear combinations of the original predictor variables. PLSR constructs these components while considering the observed response values, thus leading to a parsimonious model with reliable predictive power. Currently, PLSR has been widely used for quantitative analysis modeling.
In this study, LDA was applied to determine if the E-nose is able to detect mechanical injury of litchi fruit, determine the optimal mechanical injury detection time for litchi, select the optimal sensors and feature values, and detect the storage time of litchi fruit. KNN was used to further determine the detection effect of the E-nose on storage time of litchi. PLSR was applied to determine the ability of the E-nose to quantify the quality parameters of litchi fruit during storage. Data statistics were performed using Excel 2007 (Microsoft Corporation, Redmond, WA). Data analysis and figure output were performed using Matlab 2017a (MathworksInc., Natick, MA).
Results
Mechanical injury affects litchi quality parameters during storage
Changes in the quality of litchis after different mechanical injuries during storage are shown in Fig. 1. The BI of litchi increased constantly during storage, with severely mechanically injured litchis increasing the fastest, followed by mildly mechanically injured litchis and injury-free litchis (Fig. 1A). The TSS (Fig. 1B) and TA (Fig. 1C) of injury-free litchis changed slightly during storage (normal fluctuation, no obvious increase or decrease). However, the TSS (Fig. 1B) and TA (Fig. 1C) of both mildly and severely mechanically injured litchis decreased with storage time; severely mechanically injured litchis decreased the fastest, followed by mildly mechanically injured litchis. Therefore, mechanical injury not only accelerates the pericarp browning but also accelerates flavor loss.
Influence of mechanical injury on storage qualities of litchi. (A) Browning index. (B) Total soluble solid. (C) Titratable acidity.
Citation: HortScience horts 55, 4; 10.21273/HORTSCI14750-19
E-nose for mechanically injured litchi classifications
LDA classifications based on E-nose data for litchis with different degrees of mechanical injury during storage are shown in Fig. 2. The response data of E-nose sensors at 115 s were selected as the feature value for analysis. On storage day 0 (Fig. 2A), litchis with different mechanical injuries can be classified. However, injury-free litchis are similar to mildly mechanically injured litchis, which may lead to misclassification during detection. From storage days 1 to 4 (Fig. 2B to E, respectively), all mechanical injury degrees of litchi can be well-classified. The contribution rate distributions are not uniform on the first main axis (LD1) or the second main axis (LD2), especially on day 1 (Fig. 2B); they are 97.38% and 1.41%, respectively. However, the contribution rate distributions of LD1 and LD2 become more uniform with time after injury. They were 56.14% and 31.61%, respectively, on day 4 (Fig. 2E). The more uniform the contribution rate distribution, the stronger the robustness of the detection model. On storage day 5, mildly mechanically injured litchis overlap with severely mechanically injured litchis, which cannot be classified. Therefore, we can infer that mechanical injury of litchis can be well-detected during days 1 to 4 after injury; however, the best time to detect mechanical injury of litchi is on day 4 after injury during room temperature storage.
Linear discriminant analysis (LDA) classification results of different mechanical injury degrees during storage. (A) Day 0. (B) Day 1. (C) Day 2. (D) Day 3. (E) Day 4. (F) Day 5.
Citation: HortScience horts 55, 4; 10.21273/HORTSCI14750-19
Normal litchi quality parameters change with storage time
Changes in the quality parameters of normal litchi during cold storage are shown in Fig. 3. Similar changes occurred in injury-free litchis stored at room temperature, and the BI of cold-stored litchis increased constantly with storage (Fig. 3A); however, it increased slower than that of litchis stored at room temperature. The TSS (Fig. 3B) and TA (Fig. 3C) of cold-stored litchis changed barely during the entire storage period (days 0–24), similar to the changes in injury-free litchis stored at room temperature.
Normal litchi quality parameters change during storage. (A) Browning index change. (B) Total soluble solid change. (C) Titratable acidity change.
Citation: HortScience horts 55, 4; 10.21273/HORTSCI14750-19
E-nose for normal litchi quality parameter detection during storage
Elimination of singular sensor.
An example of the sensor response of the E-nose to normal ‘Guiwei’ litchi on day 0 is shown in Fig. 4. The sensor (R7) reached the maximum value during 10 to 55 s of the sampling period, which may use the interference information for the entire sampling period of R7, even though the response value supported the response range after 55 s.
Response of the electronic nose to litchi samples at day 0.
Citation: HortScience horts 55, 4; 10.21273/HORTSCI14750-19
The LDA classifications of the storage time of litchi (Fig. 5) have further proven the hypothesis. The response data of E-nose sensors at 115 s were selected as the feature value for analysis. LDA classification results of the storage time of litchi based on the responses of all sensors of the E-nose are shown in Fig. 5A. All storage days overlapped with each other and cannot be classified. LDA classifications of the storage time of litchi based on the responses of all sensors except R7 are shown in Fig. 5B. The independent characteristic of each storage day is more evident than the LDA results in Fig. 5A, and on days 0 and 3 they can be well classified with other storage days. We can further confirm that R7 contains more inferential classification information than helpful information; therefore, R7 should be removed from the next data analysis.
Linear discriminant analysis (LDA) classification results of litchi storage time based on all E-nose sensors (A). A singular sensor (R7) was removed from the electronic nose (E-nose) sensors (B).
Citation: HortScience horts 55, 4; 10.21273/HORTSCI14750-19
Comprehensive feature extraction and storage time detection.
The raw data of the E-nose sensor response to litchi on storage days 0, 6, 15, and 24 are shown in Fig. 6A, B, C, and D, respectively. The response value increased with storage time. The increasing rate of the response value (from 0 to 10 s) of sensors R2, R6, R8, and R9 increased with storage time. At 50 s, the response values of the sensors showed a different cross-status. Therefore, only using the 115-s response value of each sensor cannot cover the entire changed state of volatile litchi with storage. According to Fig. 6, the average differential value of 0 to 10 s, the maximum value, the50 s value, and the 115 s value were selected as the comprehensive feature value for the next analysis.
Raw data of the electronic nose (E-nose) response to litchi samples on different storage days. (A) Day 0. (B) Day 6. (C) Day 15. (D) Day 24.
Citation: HortScience horts 55, 4; 10.21273/HORTSCI14750-19
After comprehensive feature value extraction, LDA classification results of the storage time of litchi are shown in Fig. 7. All storage days can be classified. Therefore, the comprehensive feature value indicates better information regarding volatile litchi feature changes than the single 115-s value during storage.
Linear discriminant analysis (LDA) classification results of litchi storage time after feature extraction.
Citation: HortScience horts 55, 4; 10.21273/HORTSCI14750-19
All litchis on different storage days can be classified by LDA; however, some storage days, such as days 10 and 15 and days 0 and 3 are too close to each other. To further test the storage time detection effect of litchi based on the E-nose, the KNN was used for this research. Fifteen samples of each storage day were selected randomly as the calibration set, and the remaining five samples were used as the validation set. Therefore, the validation set had a total of 105 samples and the calibration set had a total of 35 samples. For KNN detection, the neighbor number (K) would affect the detection effect, and an optimal K value should be chosen by repeated attempts. After modeling, the optimal K value was 3, the detection accuracy of the calibration set was 100%, and that of the validation set was 91.43%.
PLSR for the litchi pericarp browning degree, TSS, and TA determination.
The results of PLSR used to detect the degree of litchi pericarp browning are shown in Fig. 8A and B. There were 140 samples in total, and 36, 38, 23, 20, and 23 litchi samples had browning degrees of 1, 2, 3, 4, and 5, respectively. Furthermore, 27, 29, 18, 15, and 18 samples with browning degrees from 1 to 5, respectively, were randomly selected as the calibration set. Then, 9, 9, 5, 5, and 5 samples with browning degrees of 1 to 5, respectively, were selected as the validation set. When predicting PLSR, to judge the correlation between the predicted and actual values, it is necessary to fit the coefficient (R2), as reported previously (Zhou and Wang, 2011). The range of R2 is 0 to 1; the larger the R2, the better the prediction effect. Therefore, the E-nose can effectively detect the degree of litchi browning. The R2 values of calibration (Fig. 8A) and validation (Fig. 8B) for detecting the degree of browning with PLSR are larger than 0.8.
Partial least-squares regression (PLSR) detection results of litchi pericarp browning degree (A and B), total soluble solid (TSS) (C and D), and titratable acidity (TA) (E and F) detection. (A, C, E) Detection results of calibration sets. (B, D, F) Detection results of validation sets.
Citation: HortScience horts 55, 4; 10.21273/HORTSCI14750-19
The results of PLSR used to determine litchi TSS and TA were are shown in Fig. 8 C to F. Because TSS and TA barely changed during storage, the TSS and TA values should be uniformly distributed during the entire storage period. However, considering the comprehensiveness of the data that the detection model should possess, the detection model should include sampling data of each storage day. There were a total of 140 samples; 15 samples of each storage day were selected randomly as the calibration set and 5 samples were chosen as the validation set. Therefore, the validation set had a total of 105 samples and the calibration set had a total of 35 samples. The results of using PLSR to detect the TSS are shown in Fig. 8C, and those of the validation set are shown in Fig. 8D. The results of using PLSR to detect TA are shown in Fig. 8E, and those of the validation set are shown in Fig. 8F. The detected effects of TSS and TA based on the E-nose were both unsatisfied (all R2 < 0.4).
Discussion
Changes in quality parameters of litchi due to mechanical injury and storage.
Color fading and browning of litchi are mainly caused by degradation of anthocyanin (Rivera-López et al., 1999) and water loss in the pericarp (Jiang and Fu, 1999). Decreases in TSS and TA of litchi are mainly due to respiration that consumes the nutrient substances of fresh litchi (Feng et al., 2011). The pericarp color fades significantly during storage; however, previous research found that the flavor of litchi is less changed (Jiang et al., 2006), which is in agreement with the results of this research. Our results also found that the quality of litchi fruit with mechanical injury decreases faster than that of injury-free ones, and that litchis with heavier mechanical injury experience faster degrees of browning and decreased rates of TSS or TA compared to those with lighter mechanical injury. This could be explained by previous research that indicated that mechanical injury increases water loss and respiration of litchi fruit (Chen et al., 2013a).
E-nose for mechanical injury detection of litchi fruit.
Mechanical injury can break fruit tissues and increase the release of ethylene, carbon dioxide, and secondary metabolites (Wang et al., 2007). Therefore, mechanical injury of litchi can be detected by the E-nose. This research indicated that mechanical injury of litchi can be well-classified during days 1 to 4 after injury, cannot be classified on day 5, and is best detected on day 4. This may be because the variety and concentration of injury of volatiles increased with the storage time after injury; however, volatiles with mild mechanical injury were similar to those with severe mechanical injury on storage day 5. Gas chromatography-mass spectrometer combined with E-nose should be applied in further studies to explore the details of volatile changes after mechanical injury.
Sensor optimization and feature extraction.
The successful use of the E-nose for food quality detection has been proven by many studies (Natale et al., 1997). During E-nose detection, not all sensors are useful for classification. Some sensors contain more interference information than useful information, such as the R7 for ‘Yuhebao’ litchi sampling during this experiment, and should be removed before data analysis (Shi et al., 2013). The E-nose response value of ‘Guiwei’ litchi did not reach the maximum value, but the response data of all sensors have been kept for future data analysis. In addition, previous research proved that multiple features can discover more comprehensive information of samples than a single feature, thereby providing better detection effects (Wei et al., 2015). This research suggested the use of an average differential value of 0 to 10 s, the maximum value, the 50-s value, and the 115-s value to construct the comprehensive feature value for litchi storage quality detection based on the E-nose. Therefore, E-nose data preprocessing differs according to the litchi variety and detection target.
Litchi storage time, browning degree, TSS, and TA determination.
Litchi volatiles are changed during storage, as proven by our previous research, thereby providing the theoretical basis for using the E-nose to detect quality parameters of litchi during storage (Xu et al., 2016). Our previous research results showed that alkene, alcohols, and ketones exist in litchi volatiles. The concentration of alkene was the highest and increased constantly during storage. Other aromatic components decreased during storage. This research further indicated that the E-nose is able to predict the storage time and browning degree of litchi but cannot accurately predict the TSS and TA of litchi during storage. This might be because both litchi volatiles and degrees of browning change with increased storage time; that is, there are strong relationships between E-nose data and storage time and browning degrees. However, with increased storage time, litchi volatiles changed a lot, as did TSS and TA, indicating that the relationships between E-nose data and TSS and TA are very weak.
Litchi fruit supervision during the circulation process.
During the circulation process, litchi fruit can easily sustain mechanical injury during the harvest and transportation stages and quickly decay during the storage stage. Litchis with mechanical injury should be removed at the transportation stage or early storage stage. Stored litchis should be sold before reaching a certain degree of browning. Litchis with significant browning should be discarded. Because the TSS and TA of litchi fruit do not change much during the circulation process, their supervision is unnecessary. According to our experimental results, mechanical injury of litchi fruit could be detected by the E-nose with the LDA method, the storage time of litchi fruit could be detected by the E-nose with the KNN method, and the browning degree of litchi fruit could be detected by the E-nose with the PLSR method. Therefore, the E-nose could be a feasible tool for supervising litchi fruit during the circulation process.
Conclusions
This study used the E-nose (PEN 3) to detect the quality of postharvest litchi fruits, detect mechanical injury, and detect storage quality. The E-nose may be an effective tool for litchi fruit supervision during the circulation process. The experimental results were as follows:
(1) The BI, TSS, and TA of both mildly and severely mechanically injured litchis decreased with storage time. The quality of litchis with severe mechanical injury decreased the fastest, followed by litchis with mild mechanical injury. The BI quickly increased during storage; however, the TSS and TA of injury-free and normal litchis changed slightly during room temperature and cold storage.
(2) To detect mechanical injury of ‘Guiwei’ litchi fruit, E-nose sensor response values at 115 s could be selected as feature values for analysis. However, to detect storage quality of ‘Yuhebao’ litchi fruit, sensor R7 should be removed to avoid interference. The average differential value of 0 to 10 s, the maximum value, the 50-s value, and the 115-s value are also recommended for constructing a comprehensive feature value for data analysis. E-nose data preprocessing differs according to the litchi variety and detection target.
(3) Mechanical injury of litchis can be well-classified by the E-nose with LDA during day 1 to day 4 after injury. The best time to detect mechanical injury is at day 4 after injury during room temperature storage.
(4) Litchi storage time can be classified by the E-nose with LDA; however, some storage times are too close to each other, which may lead to misclassification during practical storage time. KNN further proved the feasibility of using the E-nose to detect litchi storage time with detection accuracy rates of 100% and 91.43%, respectively, for the calibration set and validation set. The degree of browning can be detected by the E-nose with PLSR, with R2 values of 0.8582 and 0.08015, respectively, for the calibration set and validation set. However, the TSS and TA cannot be satisfactorily detected by the E-nose (R2 value <0.4 for the calibration set and the validation set).
(5) Supervision should be performed during the circulation process using the E-nose. Litchis with mechanical injury should be removed at the transportation stage or early storage stage. Litchis in storage should be sold before reaching a certain degree of browning. Litchis with significant browning should be discarded. Because the TSS and TA of litchi fruit barely change during the circulation process, their supervision is unnecessary.
Literature Cited
Ali, S., Khan, A.S. & Malik, A.U. 2016 Postharvest l-cysteine application delayed pericarp browning, suppressed lipid peroxidation and maintained antioxidative activities of litchi fruit Postharvest Biol. Technol. 121 135 142
Alves, J.A., Lima, D.O.L., Nunes, C.A., Dias, D.R. & Schwan, R.F. 2011 Chemical, physical-chemical, and sensory characteristics of lychee (Litchi chinensis Sonn) wines J. Food Sci. 76 S330 S336
AOAC Official Method 942.15 1965 Acidity (titratable) of fruit product. <http://www.eoma.aoac.org/methods/result.asp?string=b>
Cardozo, C.J.M. & Londoño, G.A.C. 2013 Determination of Soursop (Annona muricata L. cv. Elita) fruit volatiles during ripening by electronic nose and gas chromatography coupled to mass spectroscopy Rev. Fac. Nac. Agron. 66 7117 7128
Chen, Y., Cai, W.A., Xiang, H.A. & Zou, X.A. 2013a Research on the respiration and peel shape of litchi in mechanical damage Journal of Agricultural Mechanization Research 35 138 141
Chen, Y., Tan, J.H., Xiang, H.P., Zou, X.J., Bo, L.I. & Jiang, Z.L. 2014 Effect of mechanical damage on dielectric properties of litchi Food Ferment. Ind. 40 47 50
Chen, Y., Xiang, H.P., Tan, J.H., Zou, X.J., Huang, G.G. & Bo, L.I. 2013b Effects of extrusion on mechanical damage and mechanical parameters of litchi J. Hunan Agr. Univ. 39 688 692
Dharini, S., Eva, A. & Lise, K. 2008 Postharvest decay control and quality retention in litchi (cv. McLean’s Red) by combined application of modified atmosphere packaging and antimicrobial agents Crop Prot 27 1208 1214
Elshiekh, A.F. & Habiba, R.A. 1996 Effect of storage time on the quality of peach fruit held in cold storage in different types of packaging Gartenbauwissenschaf 1 8 10
Feng, C., Liu, B. & Xiao, Q. 2011 Comparative study on postharvest quality of different litchi cultivars in hainan Chinese Journal of Tropical Crops 32 1046 1050
Gong, X.J., Yu, S.Y., Yu, Y.J., Wu, J.J., Xiao, G.S. & Chen, W.D. 2014 Analysis on nutritional components of litchi juice and stability of its pulp sediments Guangdong Agr. Sci. 19 90 93
Gorjichakespari, A., Nikbakht, A.M., Sefidkon, F., Varnamkhasti, M.G., Brezmes, J. & Llobet, E. 2016 Performance comparison of fuzzy ARTMAP and LDA in qualitative classification of iranian rosa damascena essential oils by an electronic nose Sensors 16 636
Hu, W., Zhang, Z., Ji, Z., Liu, S. & Zhang, A. 2004 Changes of pericarp color and the content of anthocyanin and flavonoids in litchi pericarp during chilling-injured temperature storage Acta Hort. Sin. 31 723 726
Huang, F., Guo, Y., Zhang, R., Zhang, M., Liu, Y. & Bai, Y. 2016 Comparison of physicochemical properties and antioxidant activity of polysaccharides from litchi pulp dried by different methods J. Chin. Inst. Food Sci. Technol. 16 212 218
Jiang, Y.M. & Fu, J.R. 1999 Biochemical and physiological changes involved in browning of litchi fruit caused by water loss J. Pomol. Hort. Sci. 74 43 46
Jiang, Y. 2000 Role of anthocyanins, polyphenol oxidase and phenols in lychee pericarp browning J. Sci. Food Agr. 80 305 310
Jiang, Y., Li, Y. & Li, J. 2004 Browning control, shelf life extension and quality maintenance of frozen litchi fruit by hydrochloric acid J. Food Eng. 63 147 151
Jiang, Y.M., Wang, Y., Song, L., Liu, H., Lichter, A., Kerdchoechuen, O., Joyce, D.C. & Shi, J. 2006 Postharvest characteristics and handling of litchi fruit-an overview Aust. J. Exp. Agr. 46 12 476 482
Khan, A.S., Ahmad, N., Malik, A.U. & Amjad, M. 2012 Cold storage influences the postharvest pericarp browning and quality of litchi Intl. J. Agr. Biol. 14 389 394
Natale, C.D., Macagnano, A., Davide, F., D’Amico, A., Paolesse, R., Boschi, T., Faccio, M. & Ferrio, G. 1997 An electronic nose for food analysis Sens. Actuators B Chem. 44 521 526
Pearce, T.C., Schiffman, S.S., Nagle, H.T. & Gardner, J.W. 2006 Handbook of machine olfaction: Electronic nose technology. Wiley-VCH Verlag GmbH & KGaA, Weinheim, Germany
Rivera-López, J., Ordorica-Falomir, C. & Wesche-Ebeling, P. 1999 Changes in anthocyanin concentration in Lychee (Litchi chinensis Sonn.) pericarp during maturation Food Chem. 65 195 200
Röck, F., Barsan, N. & Weimar, U. 2008 Electronic nose: Current status and future trends Chem. Rev. 108 705 725
Ruan, W., Liu, B. & Song, X. 2012 Comparison of cooling method for litchi fruit Sci. Technol. Food Ind. 11 352 362
Shi, B., Zhao, L., Zhi, R. & Xi, X. 2013 Optimization of electronic nose sensor array by genetic algorithms in Xihu-Longjing Tea quality analysis Math. Comput. Model. 58 752 758
Tian, H., Li, F., Qin, L., Yu, H. & Ma, X. 2015 Quality evaluation of beef seasonings using gas chromatography-mass spectrometry and electronic nose: Correlation with sensory attributes and classification according to grade level Food Anal. Methods 8 6 476 482
Wang, Y.Y., Hu, W.Z., Pang, K., Zhu, P.W. & Fan, S.D. 2007 Research progress of browning mechanism of fruit and vegetables responding to mechanical stress Sci. Technol. Food Ind. 11 230 233
Wang, Z., Sun, X., Miao, J., Wang, Y., Luo, Z. & Li, G. 2017 Conformal prediction based on K-Nearest neighbors for discrimination of ginsengs by a home-made electronic nose Sensors 17 1869
Wei, Z., Wang, J. & Zhang, W. 2015 Detecting internal quality of peanuts during storage using electronic nose responses combined with physicochemical methods Food Chem. 177 89 96
Xiong, J., Lin, R., Bu, R., Liu, Z., Yang, Z. & Yu, L. 2018 A micro-damage detection method of litchi fruit using hyperspectral imaging technology Sensors 18 700
Xiong, J., Zou, X., Chen, L. & Guo, A. 2011 Recognition of mature litchi in natural environment based on machine vision Trans. Chin. Soc. Agr. Mach. 42 162 166
Xu, S., Lü, E.L., Lu, H.Z., Wang, Y.J., Yang, J. & Lin, X.J. 2016 Volatile comparison of different environment stored litchi based on SPME-GC-MS Sci. Technol. Food Ind. 37 72 77
Zhang, D. & Quantick, P.C. 1997 Effects of chitosan coating on enzymatic browning and decay during postharvest storage of litchi (Litchi chinensis Sonn.) fruit Postharvest Biol. Technol. 12 195 202
Zhang, Z. & Tong, J. 2005 Research and application of electronic nose and electronic tongue in food inspection J. Huazhong Agr. Univ. S1 25 30
Zhou, B. & Wang, J. 2011 Use of electronic nose technology for identifying rice infestation by Nilaparvata lugens Sens. Actuators B Chem. 160 15 21