May 8, 2023 at 3:39 pm | Updated May 11, 2023 at 10:34 pm | 9 min read
- Non-destructive, accurate, and real-time collection and analysis of fruit quality data continue to grow more critical.
- Their applications are being expanded to new fruits and to improve scientific research.
- Another trend is the search for non-toxic postharvest treatments to keep fruits safe for consumption and augment fruit nutritional value to meet consumer demands.
Fruit quality is vital for human health and nutrition due to the compounds they contain. The quality of the fruit will depend on the plant’s physiological processes, from fruit setting to ripening and postharvest handling. Fruit quality is built in the pre-harvest stages, and post-harvest handling is about maintaining the quality. If plants have optimal nutrition, stress-free field conditions, and the correct postharvest handling, fruit development, and quality will also be optimal. This article tracks the five most important developments in 2022 that helped food producers improve fruit quality.
1. Machine Vision
Figure 1: Schematic representation of the image classification process using Machine vision-based AI application, Bird et al. 2022. (Image credits: https://doi.org/10.1016/j.scienta.2021.110684)
Subscribe to the Felix instruments Weekly article series.
In 2019, 20 million tonnes of lemons and limes were harvested. It is also an export commodity earning USD 3.3 billion globally. International trade in fruit keeps increasing, and reducing waste can improve economic returns.
Prompt culling of spoilt fruits, starting from the farm and at various supply chain stages, can prevent the rest of the fruits from getting spoilt by limiting the defusion of ethylene, the phytohormone responsible for ripening and decay, and triggering ethylene production in other fruits.
In 2022, Bird et al. reported how they developed a Fruit Quality Recognition system involving AI (Artificial Intelligence) technologies – computer vision and machine learning – to discern damaged or moldy lemons and reject spoilt fruits. Production efficiency can be improved if spoilt fruits can be recognized and sorted rapidly, automatically, and non-destructively.
Fruit Quality Recognition programs use algorithms to score or classify fruits when trained with pictures. To train a computer program to recognize a fruit, it is necessary to have enough images that represent the large population of lemons harvested globally so that the program can be used worldwide. However, increasing sample collection is challenging. So, Bird et al. attempted to solve the data scarcity problem with a Conditional Generative Adversarial Network (Conditional GAN).
The steps they used for their project (depicted in Figure 1) were as follows:
- Using Convolutional Neural Network for fruit quality recognition with an original set of 2690 images.
- Use training data to train the Conditional GAN system to generate synthetic pictures of healthy fruits and lemons spoilt due to mold or decay. The program learned to produce synthetic images of healthy and spoilt lemons successfully.
- Augment the original data set with synthetic images to improve Convolutional Neural Network accuracy.
- Running the models.
Tests run without Conditional GAN had a classification accuracy of only 59.4%. Tests showed that incorporating Conditional GAN, and 4096 interpretation neurons, improved the efficiency of fruit recognition to give an average classification accuracy of 88.75%. Even when the Conditional GAN-augmented classification was compressed to 50% of its original size, it still could maintain 81.16% of classification accuracy.
The scientific team also suggested a method by which future fruit quality image classification could be improved by transfer learning so that the findings in this study can have wider applications in the fruit industry.
2. Effects of Selenium Fertilizers on Tomato Quality
Figure 2: Effect of selenium on tomato fruit quality, Xu et al. 2022. (Image credits: https://doi.org/10.1016/j.scienta.2022.111242)
Tomatoes are one of the most popular vegetables consumed and can be a source of essential nutraceuticals like lycopene, β-carotene, minerals, and phenolic compounds. The nutritional value of tomatoes depends on fruit quality. Biofortification to enhance fruit quality is therefore increasingly used in agriculture.
Selenium is not only an essential nutrient for plants but is also beneficial for people as it has antioxidant properties and can prevent cancer. Tomatoes biofortified by selenium can improve people’s health and reduce nutrient deficiencies.
There is no agreement on the best method of applying selenium and the optimal type of fertilizer. Soil applications are affected by organic matter, oxidation, soil, pH, competitive ion content, etc. In comparison, foliar applications don’t improve vitamin C and carotenoids level.
In 2022, Xu et al. decided to use meta-analysis to synthesize statistical results from previous studies to find the best method of applying selenium, and its effects on tomato quality. They used 696 pairs of data and found that among the fertilizer types tested, tomato quality improvement by adding Na2SeO3 was more than by adding Na2SeO4.
Selenium had a positive effect on most of the quality variables investigated. However, there was no improvement in fruit firmness, nitrate content, and other nutrient (iron and calcium) levels. However, the mode of application can moderate effects. Soil fertilization with 1∼5 mg·L−1 improved levels of vitamin C, total soluble solids, and nitrate content. Soil fertilization with more than 5 mg·L−1 selenium improved soluble sugar, lycopene, and titratable acid.
Also, adding selenium improved the quality of small-sized tomatoes more than big-sized fruits.
By quantitatively analyzing several previous studies, Zu et al. could resolve conflicting information and provide guidelines on selenium fertilizer application modes, methods, and correlated effects. As a result, tomato quality biofortification will be easier moving forward.
3. Impact of Preharvest Factors on Peach Fruit Quality
Figure 3: “Development of novel non-destructive techniques using Vis-NIRS to estimate tree-fruit quality and maturity with a single scan effectively can support growers on decisions regarding the proper harvest time and researchers on a fair evaluation of cultural techniques, new cultivars, and rootstocks towards increasing orchard quality potential,” Anthony et al. 2022. (Image credits: https://doi.org/10.1016/j.scienta.2022.110919).
Peach production and consumption have fallen in the last decades because of inferior quality, poor texture, and flavorless profile. Improving peach quality to meet consumer satisfaction can reinvigorate the industry struggling with low demands and high production costs.
Peach fruit quality is evaluated based on size, shape, color, firmness, and internal characteristics. Good color and correct texture do not indicate taste, but internal characteristics do; these are soluble solids concentration (SSC), titratable acidity (TA), and the SSC: TA ratio. Dry matter (DM), which is directly correlated to consumer satisfaction, must also be measured.
The factors that influence fruit quality will affect maturity. Therefore, scientists must control for maturity, which can otherwise become a confounding variable to find the preharvest factors that can improve fruit quality.
Historically, it was difficult to control for maturity control as methods to estimate firmness were destructive, labor- and time-intensive, and color assessments were subjective and limited to particular cultivars.
Anthony et al. used Vis- and Near Infrared Spectroscopy (NIRs) analyzed by chemometrics for non-destructive, rapid, and objective estimation of maturity and quality. The device used is Felix Instruments’ F-750 Quality Meter, as seen in Figure 3. They used the index of absorbance difference (IAD) to measure color change due to chlorophyll degradation to simultaneously track maturity and peach quality. They estimated DM, TA, DM, and overcolor to measure peach quality, see Figure 4.
Controlling for maturity helped the scientists investigate preharvest factors using proteomic, metabolomic, and transcriptomic approaches.
Figure 4: “Peach fruit quality changes during growth, development, and maturation on-tree,” Anthony et al. 2022. (Image credits: https://doi.org/10.1016/j.scienta.2022.110919).
In studies conducted on canopy position and rootstock, the fruit’s light environment significantly altered quality for two cultivars than position in the canopy or genotypic vigor.
The crop load in a tree determines the carbon supply to fruits. Thinning improved carbon supply to existing fruits and enhanced fruit quality- more DM, SSC, size, and weight. Where the carbon supply was sufficient, the quality of fruits was better than those starved of carbon. The difference in carbon supply produces a difference in metabolite profiles in initial fruit development, leading to different phenotypes at harvest. High-quality fruits with higher DM and SSC had catechin and sorbitol, while low-quality fruits had more amino acids and citric acid.
The environment also plays a crucial role, and fruits from orchards from high altitudes have higher overcolor, more SSC, and lower TA. These peaches also produced more ethylene and revealed more postharvest softening. On the other hand, low-altitude peaches lacked juiciness and were prone to internal bleeding, mealiness, and chilling postharvest injury.
Without maturity control, concluding the effect of various preharvest factors would be difficult.
4. Melatonin Maintains Papaya Fruit Quality
Figure 5: “Impact of melatonin application on fruit disease incidence. (A) Pictorial view of papaya fruits during various melatonin applications. (B) Fruit disease index. (C) Fruit disease incidence. (D) Fruit commodity rate. Papaya fruit were stored at room temperature (25 ± 1 °C) after treatment. The data are shown as mean ± SE (n = 3) and different letters represent the significant differences at the 5% level. CK: control group, “100”, “400” and “800” indicated the “100 μmolL−1”, “400 μmolL−1,” and “800 μmolL−1″ melatonin treatments,” Fan et al. 2022. (Image credits: https://doi.org/10.3390/antiox11050804)
Papaya, a fruit from the tropics and subtropics, is a nutritious fruit that is also economically and medically important. It ripens quickly after harvest and softens during storage and transport. It is also prone to a common postharvest disease, Anthracnose, caused by Colletotrichum spp., a fungus that leads to loss of quality.
Melatonin is safe and non-toxic and is widely used to delay ripening during the storage and transport of many fruits. Melatonin is a crucial signaling compound in many plants and would be better than other alternative chemical postharvest treatments for papaya.
Fan et al. decided to test if it would effectively preserve papaya. They tested melatonin’s effects on papaya quality during postharvest storage, transportation, and marketing.
All the melatonin treatments reduced ethylene production and respiration during later storage. The best treatment resulted in delayed ripening and fruit softening. It also reduced anthracnose incidence and product loss on shelves. By reducing disease incidence, melatonin decreases ethylene production and respiration rate during later storage stages. Scientists think melatonin’s most effective mode of action is reducing disease than fruit ripening. The optimal dose of melatonin was 400 μmol L−1.
Ripening is usually associated with increased reactive oxygen species (ROS) activity that damages lipid proteins, nucleic acids, and cell activity, leading to ripening. Exogenous melatonin application minimized ROS production by increasing antioxidant levels in papaya.
Melatonin suppresses disease by enhancing the activity of enzymes involved in defense and inhibiting anthracnose development.
The study was able to show that safe and non-toxic treatments can be effective in improving papaya quality without using chemicals. Such fruits will be more attractive to consumers.
5. Prediction of Date Fruit Quality Using ANN Models
Figure 6: “Artificial neural network architecture,” Mohammed et al. 2022. (Image credits: https://doi.org/10.3390/foods11111666)
Date palm is a crucial source of income and food in the Middle East and North Africa. They are also exported to European countries and the USA, requiring long-distance transport and storage. Maintaining suitable low temperatures (0–5 °C) and high relative humidity allows storing dates for six to twelve months. Color, flavor, texture, nutrients, and weight are maintained, and diseases are prevented if the conditions are correct. It is necessary to be able to monitor date fruit quality before, during, and after storage.
Tradition slow and destructive quality testing methods are being replaced by digital data collection of external and internal quality parameters and analyzed by AI technologies.
Mohammed et al. aimed to develop and test an Artificial Neural Networks (ANNs) model, the most popular machine learning AI technique, to predict date fruit physicochemical properties. The parameters they tried were total soluble solids, pH, moisture content, and water activity based on their electrical properties. They collected 800 samples from ten cultivars stored for 0, 2, 4, and 6 months. A high-precision LCR meter was used to measure the electrical properties of the 800 samples at frequencies 10 Hz to 100 kHz. Traditional laboratory tests also determined the values of the parameters.
The ANN models were compared with a Multiple Linear Regression (MLR) model for all the frequencies.
The MLR proved less accurate than ANN in predicting pH, and its performance was also low in predicting sugar and water content and water activity.
The best ANN model had an input layer with 14 neurons and a hidden layer with 15 neurons, as shown in Figure 6. This ANN model would accurately predict all the quality parameters tested at 10 kHz with high R2 and low RMSE and could be a new solution for non-destructive real-time prediction of date fruit quality.
Non-Destructive Measurement and AI Analytics
Felix Instruments Applied Food Science, an industry leader, has been producing portable quality meters that are NIRs based. Growers and other stakeholders use it globally for fruit quality control and monitoring as it is easy to use, rapid, and accurate. Even scientists have been using these devices in experiments as the tools’ accuracy is reliable and compares well with standard laboratory techniques.
Anthony, B. M., & Minas, I. S. (2022). Redefining the Impact of preharvest factors on peach fruit quality development and metabolism: A review. Scientia Horticulturae, 297, 110919. https://doi.org/10.1016/j.scienta.2022.110919
Bird, J. J., Barnes, C. M., Manso, L. J., Ekárt, A., & Faria, D. R. (2022). Fruit quality and defect image classification with conditional GAN data augmentation. Scientia Horticulturae, 293, 110684. https://doi.org/10.1016/j.scienta.2021.110684
Fan, S., Li, Q., Feng, S., Lei, Q., Abbas, F., Yao, Y., Chen, W., Li, X., & Zhu, X. (2022). Melatonin Maintains Fruit Quality and Reduces Anthracnose in Postharvest Papaya via Enhancement of Antioxidants and Inhibition of Pathogen Development. Antioxidants, 11(5), 804. https://doi.org/10.3390/antiox11050804
Mohammed, M., Munir, M., & Aljabr, A. (2022). Prediction of Date Fruit Quality Attributes during Cold Storage Based on Their Electrical Properties Using Artificial Neural Networks Models. Foods, 11(11), 1666. https://doi.org/10.3390/foods11111666
Xu, X., Wang, J., Wu, H., Yuan, Q., Wang, J., Cui, J., & Lin, A. (2022). Effects of selenium fertilizer application and tomato varieties on tomato fruit quality: A meta-analysis. Scientia Horticulturae, 304, 111242. https://doi.org/10.1016/j.scienta.2022.111242
- Spectrophotometry in 2023
- The Importance of Food Quality Testing
- NIR Applications in Agriculture – Everything…
- The 5 Most Important Parameters in Produce Quality Control
- Active Packaging: What it is and why it’s important
- Liquid Spectrophotometry & Food Industry Applications
- Ethylene (C2H4) – Ripening, Crops & Agriculture
- Guide to Fresh Fruit Quality Control
- Melon Fruit: Quality, Production & Physiology
- Understanding Chemometrics for NIR Spectroscopy