Non-destructive sensing methods for quality assessment of on-tree fruits: a review

A Specular Highlight Removal Algorithm for Quality Inspection of Fresh Fruits

Nondestructive inspection technology based on machine vision can effectively improve the efficiency of fresh fruit quality inspection. However, fruits with smooth skin and less texture are easily affected by specular highlights during the image acquisition, resulting in light spots appearing on the surface of fruits, which severely affects the subsequent quality inspection. Aiming at this issue, we propose a new specular highlight removal algorithm based on multi-band polarization imaging. First of all, we realize real-time image acquisition by designing a new multi-band polarization imager, which can acquire all the spectral and polarization information through single image capture. Then we propose a joint multi-band-polarization characteristic vector constraint to realize the detection of specular highlight, and next we put forward a Max-Min multi-band-polarization differencing scheme combined with an ergodic least-squares separation for specular highlight removal, and finally, the chromaticity consistency regularization is used to compensate the missing details. Experimental results demonstrate that the proposed algorithm can effectively and stably remove the specular highlight and provide more accurate information for subsequent fruit quality inspection. Besides, the comparison of algorithm speed further shows that our proposed algorithm has a good tradeoff between accuracy and complexity.

Predicting Grape Sugar Content under Quality Attributes Using Normalized Difference Vegetation Index Data and Automated Machine Learning

Wine grapes need frequent monitoring to achieve high yields and quality. Non-destructive methods, such as proximal and remote sensing, are commonly used to estimate crop yield and quality characteristics, and spectral vegetation indices (VIs) are often used to present site-specific information. Analysis of laboratory samples is the most popular method for determining the quality characteristics of grapes, although it is time-consuming and expensive. In recent years, several machine learning-based methods have been developed to predict crop quality. Although these techniques require the extensive involvement of experts, automated machine learning (AutoML) offers the possibility to improve this task, saving time and resources. In this paper, we propose an innovative approach for robust prediction of grape quality attributes by combining open-source AutoML techniques and Normalized Difference Vegetation Index (NDVI) data for vineyards obtained from four different platforms-two proximal vehicle-mounted canopy reflectance sensors, orthomosaics from UAV images and Sentinel-2 remote sensing imagery-during the 2019 and 2020 growing seasons. We investigated AutoML, extending our earlier work on manually fine-tuned machine learning methods. Results of the two approaches using Ordinary Least Square (OLS), Theil-Sen and Huber regression models and tree-based methods were compared. Support Vector Machines (SVMs) and Automatic Relevance Determination (ARD) were included in the analysis and different combinations of sensors and data collected over two growing seasons were investigated. Results showed promising performance of Unmanned Aerial Vehicle (UAV) and Spectrosense+ GPS data in predicting grape sugars, especially in mid to late season with full canopy growth. Regression models with both manually fine-tuned ML (R² = 0.61) and AutoML (R² = 0.65) provided similar results, with the latter slightly improved for both 2019 and 2020. When combining multiple sensors and growth stages per year, the coefficient of determination R² improved even more averaging 0.66 for the best-fitting regressions. Also, when considering combinations of sensors and growth stages across both cropping seasons, UAV and Spectrosense+ GPS, as well as Véraison and Flowering, each had the highest average R² values. These performances are consistent with previous work on machine learning algorithms that were manually fine-tuned. These results suggest that AutoML has greater long-term performance potential. To increase the efficiency of crop quality prediction, a balance must be struck between manual expert work and AutoML.

Making Sense of Light: The Use of Optical Spectroscopy Techniques in Plant Sciences and Agriculture

As a result of the development of non-invasive optical spectroscopy, the number of prospective technologies of plant monitoring is growing. Being implemented in devices with different functions and hardware, these technologies are increasingly using the most advanced data processing algorithms, including machine learning and more available computing power each time. Optical spectroscopy is widely used to evaluate plant tissues, diagnose crops, and study the response of plants to biotic and abiotic stress. Spectral methods can also assist in remote and non-invasive assessment of the physiology of photosynthetic biofilms and the impact of plant species on biodiversity and ecosystem stability. The emergence of high-throughput technologies for plant phenotyping and the accompanying need for methods for rapid and non-contact assessment of plant productivity has generated renewed interest in the application of optical spectroscopy in fundamental plant sciences and agriculture. In this perspective paper, starting with a brief overview of the scientific and technological backgrounds of optical spectroscopy and current mainstream techniques and applications, we foresee the future development of this family of optical spectroscopic methodologies.

Application of hue spectra fingerprinting during cold storage and shelf-life of packaged sweet cherry

Presented work investigated the application of a new color analysis technique in post-harvest life of sweet cherry (Prunus avium L. ‘Hudson’). The hue spectra fingerprinting creates a histogram of image colors by summarizing the saturation. The advantage of this calculation method is that vivid colors make peaks while neutral background color is eliminated without object segmentation. Partial Least Squares (PLS) regression was used to estimate reference parameters during 9 d cold storage at 10 ± 0.5 °C (RH = 90 ± 1%) and following 2 d shelf-life at 20 ± 0.5 °C. The reference parameters of respiration, weight loss, fruit firmness and total soluble solid (TSS) content were measured. Samples were split into seven groups according to the number of perforations of polypropylene film and fructose concentration of moisture absorber. It was observed that parameters TSS and fruit firmness were the most sensitive to the length of storage. Weight loss was affected significantly by packaging. All reference parameters were estimated by PLS model with R2 > 0.917, but weight loss and respiration obtained high estimation error of RMSE% = 48.02% and 11.76%, respectively. TSS and fruit firmness prediction were successful with RMSE% = 0.84% and 1.85%, respectively. Desiccation and color change of peduncle became visible in the green range of hue spectra. Color change of red fruit was observed with decreasing saturation in the red range of hue spectra. Our findings suggest that hue spectra fingerprinting can be a useful nondestructive method for monitoring quality change of sweet cherry during post-harvest handling and shelf-life.

Recent advances in ultrasound application as a novel technique in analysis, processing and quality control of fruits, juices and dairy products industries: A review

Recently, ultrasound has been widely used in the researches in the food science and technology. Among the food materials, fruits, juices and dairy products are strongly sensitive to ultrasound-based techniques. In this review paper, applications of ultrasound in terms of high and low-power modes in processing, instrumentation and control of the industrial operations are studied in the fruit, juice and dairy sectors of the food science and technology by emphasizing the principles of these techniques and their physicochemical effects on the products. Also, a comprehensive analysis of these methods and important factors influencing their performance are presented, along with the advantages and the drawbacks of each ultrasound-assisted techniques. The solutions for better productivity of this technique would be presented and future trend of this technology would be outlined. By studying the latest researches and advances in the field of ultrasound applications, it has been confirmed that this technique can be helpful in accelerating processes, reducing energy requirements, increasing productivity, and producing better quality food materials in the fields of fruits, juices and dairy products. Notwithstanding, in order to solve the challenges ahead, and for potential applications of ultrasound technology, further researches need to be expanded in the areas mentioned. Also, with the advancement of technology, more advanced equipment, ultrasound transducers and instrumentations would be introduced that needs to be applied in this regard, to solve the current challenges.

Evaluation Performance of Ultrasonic Testing on Fruit Quality Determination

There are several destructive and nondestructive methods for quality evaluation of agricultural products. Most of the employed traditional techniques are time-consuming and involve considerable degree of manual works. Destructive methods provide reasonable success rate of quality determination of fruits; however, they practically have many concerns about effectiveness, time, and cost. Therefore, developing portable, fast, and cost-effective techniques without harming fruits are desired for fruit quality evaluation. This work aims to develop a complete nondestructive quality evaluation system with (a) ultrasonic testing and (b) volume estimation by automatic machine vision techniques. The ultrasonic system consisted of a programmable bipolar remote pulser unit, a couple of piezoelectric probes for ultrasonic signal acquisitions as a transmitter and a receiver, an oscilloscope, and a computer. Visual appearance (size/volume) was determined using a machine vision system based on image processing techniques. Five different images of a fruit from different angles were captured by high-resolution digital cameras. Volume of the fruit was computed after horizontal and vertical distance of the fruit’s images captured. The calculated volume values by the computer vision system are validated with the theoretical values. Although nondestructive ultrasonic estimation and volume estimation by image processing methods are cheap, fast, and practical, the results obtained in our experiments concluded that these methods are not as reliable as claimed in the literature.

1 H-NMR Based Metabolomics Profiling of Citrus Juices Produced in Veracruz, México

This study describes the ¹ H-NMR-based metabolomics profiling of juices from citrus fruits harvested in the state of Veracruz, México. The hydrophilic profile of commercial lemons (Agrio and Persian), tangerines (Fremont and Mónica), oranges (Valencia and Washington Navel), and grapefruits (Red Ruby and Rio Red) was determined. According to our results, 35 metabolites were identified in the ¹ H-NMR profiling. The statistical differences obtained by PCA and OPLS-DA revealed that specific amino acids, sugars, and organic acids were differential metabolites in the species and cultivars studied. High endogenous levels of sucrose (10-190 mM), α-glucose, β-glucose (α- and β-isomers, 40-205 mm), and fructose (36-170 mm) were detected in the juices of grapefruits, oranges, and tangerines, whereas citric acid (40-530 mm) was the principal organic acid in the juices of lemons. To calculate the specific amounts of metabolites from these species and their cultivars, the results were finely analyzed using the qNMR method. According to these calculations, Valencia oranges had the highest concentration of ascorbic acid (>2 mm). The described ¹ H-NMR method is highly reproducible, inexpensive, and highly robust in comparison to other analytical methods used to determine the hydrophilic profile of citrus juices.