Next Article in Journal
Efficient Room-Temperature Storage of Toona sinensis with Salicylic Acid Treatment: Enhancing Postharvest Quality Through Antioxidant and Nitrogen Metabolism Modulation
Previous Article in Journal
Agronomic and Environmental Suitability of ‘Tabtim Siam’ and ‘Xishi’ Pomelo in Dabu County (China)
Previous Article in Special Issue
Non-Destructive Monitoring of Postharvest Hydration in Cucumber Fruit Using Visible-Light Color Analysis and Machine-Learning Models
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Horticulturae
Volume 11
Issue 11
10.3390/horticulturae11111291
horticulturae-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Optical and Chemical Profiling of Japanese Strawberries: Fluorescence Fingerprints, Imaging Features, and Quality Attributes Prediction

by
Maulidia Hilaili
1,†,
Ayoub Fathi-Najafabadi
2,†,
Nurwahyuningsih
3,4,
Noelia Castillejo
5,
Lucia Russo
2,
Naoshi Kondo
1 and
Danial Fatchurrahman
2,*
1
Laboratory of Bio-Sensing Engineering, Graduate School of Agriculture, Kyoto University, Kyoto 606-8502, Japan
2
Dipartimento di Scienze Agrarie, Alimenti, Risorse Naturali e Ingegneria (DAFNE), Università di Foggia, Via Napoli 25, 71122 Foggia, Italy
3
Department of Agricultural and Environmental Engineering, United Graduate School of Agricultural Science, Tokyo University of Agricultural and Technology, Tokyo 183-0057, Japan
4
Department of Agricultural Technology, Politeknik Negeri Jember, Jember 68121, Indonesia
5
Postharvest and Refrigeration Group, Department of Agricultural Engineering, Institute of Plant Biotechnology, Universidad Politécnica de Cartagena, 30203 Cartagena, Spain
*
Author to whom correspondence should be addressed.
†
These authors contributed equally to this work.
Horticulturae 2025, 11(11), 1291; https://doi.org/10.3390/horticulturae11111291
Submission received: 8 September 2025 / Revised: 13 October 2025 / Accepted: 18 October 2025 / Published: 27 October 2025
(This article belongs to the Special Issue Application of Non-Destructive Detection Techniques in Horticultural Plants)
Browse Figures
Versions Notes

Abstract

New strawberry cultivars with unusual peel colors, such as white and peach, require thorough characterization to understand their quality traits. In this study, we examined three Japanese cultivars, ‘Kotoka’ (red), ‘Awayuki’ (peach), and ‘Pearl White’ (white), to investigate their differences in chemistry and optical properties. We measured the sugar content, acidity, and maturity index, and combined these with fluorescence spectroscopy and imaging under three LED lights (365 nm, 420 nm, and white). The fluorescence data showed clear differences between cultivars, in which the ‘Pearl White’ gave a strong near-UV peak around 290/325 nm, ‘Awayuki’ had a high far-red signal in 490/745 nm, and ‘Kotoka’ showed lower fluorescence overall. Imaging backed up these findings, with ‘Pearl White’ and ‘Awayuki’ looking brighter under UV while Kotoka appeared darker and more uniform. Texture analysis showed ‘Pearl White’ had a more uneven surface, while ‘Kotoka’ was smoother. The basic chemistry also matched these trends, as ‘Kotoka’ had the most sugar and acid, giving it a sharper taste, while ‘Pearl White’ had the highest maturity index and a milder flavor. These results demonstrate how fluorescence fingerprints and imaging features, when combined, can rapidly characterize strawberry types and assess their quality without damaging the fruit.

1. Introduction

The cultivated strawberry (Fragaria × ananassa) is a globally significant fruit crop, highly valued for its distinct sensory attributes, nutritional content, and abundance of bioactive compounds [1,2,3]. This valuation is reflected in global production, which surged by over 102% in the last two decades, reaching 9.5 million tons in 2022 [4]. While its domestic production has declined, the strawberry remains one of Japan’s most important horticultural commodities, with a focus on premium varieties for the luxury market. Among them are ‘Kotoka’ (red), ‘Awayuki’ (peach-colored), and ‘Pearl White’ (white), now more widely distributed worldwide.
Strawberries are widely recognized as a functional food due to their high concentration of phenolic compounds, particularly anthocyanins, which impart their vibrant color and are associated with numerous health benefits [5,6,7]. These compounds are credited with antioxidant and anti-inflammatory properties, potentially reducing the risk of chronic conditions such as cardiovascular disease, type-2 diabetes, and neurodegeneration [8,9]. In parallel, consumer perception of quality is heavily influenced by flavor, a complex trait governed by the balance of primary metabolites, such as soluble solids and organic acids [10].
The economic and nutritional importance of strawberries has driven extensive research aimed at improving quality traits. Consequently, modern breeding programs and postharvest management prioritize fruits with superior visual appeal, enhanced flavor, extended shelf life, and high antioxidant capacity [11]. Conventional quality assessment relies on a suite of physicochemical analyses, including total anthocyanin content, surface color, firmness, total phenolic compounds, total soluble solids content (SSC), and titratable acidity (TA) [12,13]. However, these methods are typically destructive, labor-intensive, and not amenable to high-throughput screening. This limitation is particularly critical for strawberries, a highly perishable commodity that is prone to rapid postharvest deterioration, compromising the fruit’s firmness, color, and flavor.
To overcome these challenges, various non-destructive techniques have been investigated for inspecting strawberry quality. For instance, near-infrared (NIR) spectroscopy has been used to estimate firmness, SSC, and TA [14], while hyperspectral imaging has been applied to predict moisture content, SSC, and pH [15]. Despite their potential, techniques like hyperspectral imaging can be hindered by the complexity and computational cost of processing large datasets [16].
Fluorescence spectroscopy has emerged as a sensitive, rapid, and cost-effective alternative that targets intrinsic fluorophores linked to pigments and other key phytochemicals [17]. Specifically, excitation–emission matrix (EEM) spectroscopy provides a comprehensive fluorescence ‘fingerprint’ by mapping emission intensity as a function of both excitation and emission wavelengths (I(λex, λem)). Although complex food matrices can produce overlapping bands and inner-filter effects, EEMs offer a powerful, non-destructive snapshot of the composite fluorophore profile underlying pigment and phenolic composition [18,19].
This fingerprinting strategy has proven effective in various agri-food applications, including the detection of hazelnut oil adulteration in olive oil [20], pre-harvest detection of bitter pit in apples [21], monitoring the quality of green bell peppers [22,23], quality monitoring of green bell peppers using both fluorescence for damage detection, and RGB imaging for freshness assessment [24,25]. Similar approaches have also been used for the quality characterizations of Ruby and Hayward kiwifruit [26]. Nevertheless, the application of EEM spectroscopy to differentiate strawberry cultivars with contrasting pigmentation, such as the Japanese white and peach varieties, remains largely unexplored. Because EEMs probe fluorophore families (e.g., phenolics, chlorophyll-related compounds) that are precursors or contributors to visible color, we hypothesize that combining fluorescence fingerprints with image-derived color and texture metrics is a promising approach for non-destructive varietal characterization. This investigation sought to establish a correlative model for assessing the quality of three Japanese strawberry cultivars (‘Kotoka’, ‘Awayuki’, and ‘Pearl White’) as both intact and fresh-cut fruit. To achieve this, non-destructive optical measurements from excitation–emission matrix (EEM) fluorescence spectroscopy and digital imaging were integrated with traditional physicochemical analyses. The study evaluated the predictive capacity of the derived optical and image-based features for key quality attributes, including firmness, SSC, TA, and the maturity index (MI).

2. Materials and Methods

2.1. Fruit Source

The experiment used 60 fruit samples from each of three different strawberry varieties: ‘Kotoka’ (red strawberries), ‘Awayuki’ (peach-colored strawberries), and ‘Pearl White’ (white strawberries), as illustrated in Figure 1. The strawberries were purchased from a supermarket, and they were chosen based on uniform size, healthy external appearance, absence of visible defects or mechanical damage, and skin color typical of commercial maturity for each cultivar, as indicated on the packaging. According to the label, all fruits were harvested on 4 May 2024, at the Nara Strawberry Lab in Nara Prefecture, Japan. Measurements were taken one day after harvesting. The experiments included fluorescence imaging and EEM measurement. Additionally, all strawberries were divided into two groups for destructive analysis.

2.2. RGB Image Acquisition

The image acquisition process was designed to ensure high reproducibility and comparability across all samples (Figure 2), utilizing a Canon EOS R10 mirrorless camera paired with an RF-S18-150 IS STM lens. To achieve this, the camera was operated in full manual mode, with exposure settings and a custom white balance locked prior to data collection. The white balance was calibrated using a certified 18% gray card under the experimental lighting. A comprehensive calibration workflow was implemented, which included capturing images of an X-Rite ColorChecker card to generate a color correction profile, a black reference image (with the lens cap on) to correct for sensor noise, and a white reference image (of a 99% diffuse reflectance standard) to correct for non-uniform illumination. For capturing true-color images, illumination was provided by two white LED lamps (LDL2-80X16SW2, CCS Inc., Kyoto, Japan) positioned at 45° angles to create a diffuse field (8700 lx, 4500 K), with camera settings of ISO 320, a 1/4 s shutter speed, and f/11. For UV-A fluorescence imaging, two 365 nm LED lamps (LDL-71X12UV2-365-N, CCS Inc., Kyoto, Japan) were used for excitation, and the resulting emission was captured at ISO 4000 with a 1 s shutter speed, utilizing polarizing filters to minimize specular reflection. Similarly, for blue-light fluorescence imaging, excitation was provided by two 420 nm LED lamps (HLV2-22VL420-NR, CCS Inc., Kyoto, Japan), with the camera configured to a 2 s shutter speed and fitted with both a polarizing filter and a Y-49 long-pass filter (Toshiba, Tokyo, Japan) to isolate the emission signal. All subsequent image processing was conducted using a self-developed algorithm in Python (v. 3.11.5).

2.3. Image Processing

Images of the strawberry samples were processed in Python (v. 3.11.5) using a multi-step workflow to extract the region of interest (ROI) corresponding to the fruit. The pre-processing step involved applying a Gaussian blur to reduce image noise. Subsequently, the image was converted to the CIELAB color space, where automated thresholding (Otsu’s method) on the a* channel was used to segment the fruit from the background. The resulting binary mask was then refined using morphological operations, specifically an ‘opening’ to eliminate background noise and a ‘hole-filling’ procedure to ensure a solid ROI. This automated segmentation was visually assessed and required no manual corrections due to the high-contrast imaging setup. From this final ROI, quantitative features were extracted. The average CIELAB L*, a*, and b* values were computed from the masked color pixels to ensure a consistent color measurement. The ROI was also converted to grayscale, and six gray-level co-occurrence matrix (GLCM) features (contrast, dissimilarity, homogeneity, energy, and correlation) were calculated to describe the uniformity and patterns of the peel’s surface texture.

2.4. Fluorescence Excitation-Emission (EEM) Measurement

A fluorescence spectrophotometer (Spectrofluorometer FP-8300, Jasco, Tokyo, Japan) was employed to measure fluorescence EEM of intact skin samples of whole and flesh from sliced strawberries using front-face geometry for sample handling. The scan speed was set to 5000 nm/min, and the response time was 50 msec. Spectra were recorded with excitation ranging from 200 nm to 730 nm and emission from 215 nm to 745 nm. The excitation and emission bandwidths were both set at 10 nm, with data acquisition intervals of 10 nm for both wavelengths. A low-sensitivity setting was used throughout the measurements. The experiment involved scanning both the skin and flesh parts of the strawberries, with three replicates performed for each measurement. Prior to analysis, all data were corrected to remove instrumental biases using the Spectra Manager™ software (Version 20.15, JASCO Corp). Raman peak measurements were performed on pure water under 350 nm excitation to calibrate the system to Raman units. Following the methodology described by [27], the acquired EEM data were initially corrected and then converted from arbitrary units to Raman units.

2.5. Physicochemical Properties Assessment

To measure the TSS content of the strawberries, the fruits were first carefully crushed using a mortar to extract the juice. The extracted juice was then filtered through a piece of cheesecloth to remove any pulp, seeds, or solid particles, ensuring that only clear liquid was used for analysis. The filtered juice was analyzed using a digital refractometer (model Rx-7000cx; Atago Co., Ltd., Tokyo, Japan), which is specifically designed to measure the concentration of soluble solids in a liquid. The refractometer provided readings expressed as a percentage, representing the proportion of soluble solids present in the juice. This value served as an indicator of the sweetness of the strawberries [28,29].
The TA of the strawberries was determined using a standard titration method. A specific amount of the filtered juice was placed into a titration flask, and 0.1 N NaOH was added while the sample was gently stirred. Titration continued until the solution reached an endpoint pH of approximately 8.1–8.2, indicating the neutralization of the organic acids present, primarily citric acid. The amount of NaOH used was recorded and used to calculate the total acidity of the juice. Results were expressed as citric acid equivalents, providing a standardized measure of the sample’s acid content. Measuring titratable acidity is valuable because it influences flavor balance, fruit ripeness, and overall quality.
The maturity index (MI) of the strawberries was determined by calculating the ratio between the TSS and the TA in the juice [30]. This ratio provides a simple way to describe the balance between sweetness and acidity in the fruit, which is a crucial factor in determining overall eating quality.
Fruit firmness was quantified using a digital force gauge (Model FGE-X, Nidec-Shimpo, Kyoto, Japan) equipped with a 3 mm-diameter cylindrical stainless-steel probe. Individual fruits were positioned on a stable, horizontal surface, and the probe was applied perpendicularly to the equatorial region at a constant velocity until a penetration depth of 5 mm was achieved. Sixty independent measurements were conducted for each cultivar, and the resulting force values were expressed in newtons (N).

2.6. Statistical Analysis

Statistical analysis was performed to evaluate the effect of the cultivar on the measured physicochemical and optical properties. A one-way analysis of variance (ANOVA) was conducted, and mean comparisons were performed using Fisher’s Least Significant Difference (LSD) test at a significance level of p ≤ 0.05. This analysis was carried out using Statgraphics Plus software (version XV, Statpoint Technologies Inc., Warrenton, VA, USA). Subsequently, multiple linear regression (MLR) models were developed in RStudio (version 2025.09.1) to predict the physicochemical properties of the strawberries from their fluorescence variables. For this purpose, the dataset was partitioned into a calibration set (70% of the data) and a prediction set (30%), ensuring that data from all cultivars were proportionally represented in both subsets. The predictive performance of the models was evaluated using the coefficient of determination (R2), adjusted R2, root mean square error (RMSE), and mean absolute error (MAE).

3. Results

3.1. Physicochemical Parameters

The three Japanese strawberry varieties showed distinct patterns in their physicochemical composition (Table 1). Regarding TSS, ‘Kotoka’ had the highest value (15.53 °Brix) at harvest, followed by ‘Pearl White’ (13.11 °Brix) and ‘Awayuki’ (11.66 °Brix). ‘Kotoka’ also had the highest titratable acidity (0.96%), while ‘Awayuki’ and ‘Pearl White’ had significantly lower (and similar) TA (0.61%). The maturity index (TSS/TA) values reflected these flavor differences: ‘Pearl White’ had the highest MI (22.28), followed by ‘Awayuki’ (19.62) and ‘Kotoka’ (16.37). These differences indicate distinct sugar–acid balances and maturation patterns for each cultivar. Regarding firmness, among the three cultivars, Pearl White exhibits the highest texture value at 4.85 N, which means it has the firmest flesh, followed closely by Kotoka at 4.74 N. These two varieties are statistically similar in firmness, as indicated by the same grouping letter. In contrast, Awayuki has a significantly lower texture value of 4.01 N, reflecting a softer, more tender flesh. This comparison shows that Pearl White and Kotoka are more resistant to mechanical deformation, which can enhance their shelf life and handling, while Awayuki’s lower texture value suggests a “melt-in-your-mouth” quality preferred for immediate consumption or luxury desserts.

3.2. Colorimetric Attributes and Physical Properties of GLCM Based on Fluorescence Image

Table 2 presents the skin (whole) and internal flesh (slice) characteristics of the three strawberry varieties under 365 nm, 420 nm, and white LED illumination, while Figure 3 provides the corresponding fluorescence images for visual comparison. Differences in RGB channel values were evident across all lighting conditions, showing distinct color features associated with each cultivar.
Under 365 nm UV illumination, the two paler cultivars had much higher blue and green channel intensities than the red cultivar. ‘Awayuki’ skin showed the highest blue value (160.17), followed by ‘Pearl White’ (139.49), whereas ‘Kotoka’ was minimal (36.87); a similar trend was observed for the green channel (‘Pearl White’ 125.67, ‘Awayuki’ 111.28, ‘Kotoka’ 18.95). The red channel intensity was greatest in ‘Awayuki’ (64.85), moderate in ‘Pearl White’ (48.21), and lowest in ‘Kotoka’ (32.85). In the CIELAB color space, ‘Kotoka’ presented the highest a* value (138.82), consistent with its deep red hue, while ‘Pearl White’ showed the highest lightness (L* = 123.14), showing a lighter appearance. ‘Awayuki’ had the intermediate values for L* (115.91) and a* (128.09). GLCM indicated that ‘Pearl White’ had the highest surface contrast (115.46) and dissimilarity (5.41), reflecting a more irregular peel pattern, whereas ‘Kotoka’ had the lowest contrast (32.01), corresponding to a more uniform red surface. As illustrated in Figure 3, fluorescence was visually most pronounced in the red cultivar ‘Kotoka’, moderate in ‘Awayuki’, and minimal in ‘Pearl White’, consistent with pigment distribution and the numerical RGB/CIELAB data.
Once the fruits were sliced to expose the flesh, brightness increased markedly for all cultivars. ‘Pearl White’ slices were the brightest (B = 228.85, G = 191.22), followed by ‘Awayuki’ (B = 216.01, G = 166.3), while ‘Kotoka’ flesh remained much darker (B = 98.04, G = 87.98). L* also increased in sliced fruit, with ‘Pearl White’ showing 185.93, ‘Awayuki’ 166.92, and ‘Kotoka’ 94.12. Texture became smoother and more uniform in all sliced samples, evidenced by much lower contrast and dissimilarity and higher homogeneity compared to the intact whole fruits.
Under 420 nm illumination, the color response patterns shifted: ‘Pearl White’ skin had the highest blue (10.85) and green (105.66) intensities, whereas ‘Awayuki’ (B = 1.43; G = 60.79) and ‘Kotoka’ (B = 2.57; G = 26.19) were much lower. ‘Awayuki’ showed the highest red channel under this light (R = 49.34), compared to ‘Pearl White’ (36.15) and ‘Kotoka’ (15.97). Regarding the CIELAB space, ‘Kotoka’ retained the highest a* value (120.27), whereas ‘Pearl White’ had the highest b* (169.05). GLCM analysis similarly indicated that ‘Pearl White’ had a more variable surface (higher contrast = 47.96; dissimilarity = 4.03) than the other cultivars at 420 nm. As seen in Figure 3, these trends were visually confirmed: under 420 nm light, ‘Pearl White’ appeared bright green-blue, ‘Kotoka’ retained a reddish tone, and ‘Awayuki’ was intermediate in appearance.
When fruits were sliced and viewed under 420 nm, all cultivars appeared brighter internally than on the peel. ‘Pearl White’ flesh had the highest blue (14.05) and green (117.62) levels, while ‘Kotoka’ flesh had the lowest (B = 2.7, G = 56.34). Notably, ‘Kotoka’ slices showed the highest red intensity (55.93) among the varieties under this lighting. L* also increased in sliced fruit compared to the fruit’s skin with values of 111.01 for ‘Pearl White’, 77.19 for ‘Awayuki’, and 59.18 for ‘Kotoka’. Texture metrics indicated smoother, more homogeneous surfaces in all sliced samples (lower contrast and higher homogeneity) compared to the outer skin.
Under white LED illumination, the varietal color differences were very clear. ‘Pearl White’ had the brightest appearance (B = 125.27, G = 173.57) and highest lightness (L* = 185.01), ‘Awayuki’ was intermediate (B = 76.45, G = 139.27, L* = 162.41), and ‘Kotoka’ was darkest (B = 1.36, G = 9.25, L* = 60.54) but had the most intense redness. Indeed, ‘Kotoka’ showed the highest a* (170.23), whereas ‘Pearl White’ had the lowest a* (133.41) among the three. The red channel values were similarly high in ‘Awayuki’ (208.11) and ‘Pearl White’ (204.96), while ‘Kotoka’ was lower (116.83). GLCM texture analysis showed that ‘Pearl White’ peels had the highest contrast (187.47) and dissimilarity (5.79), whereas ‘Kotoka’ had the lowest values (contrast = 59.15; dissimilarity = 3.13), indicating a smoother, more uniform surface. Visual observations from Figure 3 corroborated these findings: ‘Kotoka’ appeared vivid red, ‘Awayuki’ pinkish, and ‘Pearl White’ pale and uniform.
Slicing under white light further increased the brightness of all cultivars’ flesh and reduced their color differences. ‘Pearl White’ slices were the brightest (B = 173.08, G = 212.66), followed by ‘Awayuki’ (B = 139.46, G = 195.49), while ‘Kotoka’ slices were the least bright (B = 79.32, G = 141.95). Unlike the whole fruit, the red channel intensities became similar across the sliced samples. L* values increased significantly, especially in ‘Pearl White’ (216.43) and ‘Awayuki’ (207.71), while ‘Kotoka’ slice remained darker (173.59). Texture parameters showed lower contrast and higher homogeneity in all sliced samples, with ‘Pearl White’ slices showing the most uniform internal surface with a homogeneity value of 0.47.
Overall, these visual observations confirm the cultivar-dependent differences in color, pigment distribution, and fluorescence properties. The qualitative patterns observed (Figure 3) aligned closely with the quantitative RGB and texture data (Table 2), underscoring that each cultivar possesses a unique optical signature.

3.3. EEM of the Three Studied Varieties of Strawberry

EEM fluorescence spectra revealed clear differences among the strawberry varieties and between peel and flesh tissues (Figure 4). In the peel, ‘Pearl White’ showed the highest intensity at the 290/325 nm peak (112.29 R.U.), followed by ‘Awayuki’ (74.27 R.U.), and ‘Kotoka’ (54.07 R.U.) (Table 3). ‘Awayuki’ showed the strongest fluorescence at longer wavelengths, recording the highest intensity at both 340/435 (1184.78 R.U.) and 490/745 nm (1493.76 R.U.), compared with ‘Kotoka’ with values of 507.78 and 764.05 R.U. and ‘Pearl White’ with values of 73.78 and 86.56 R.U., respectively. In the flesh, all varieties had similar intensities at 280/235 nm (274.5 R.U.). However, ‘Kotoka’ had the highest intensity at 440/685 (158.72 R.U.), whereas ‘Pearl White’ and ‘Awayuki’ were lower with values of 134.74 and 97.22 RU, respectively. At 440/745 nm, ‘Pearl White’ exhibited the strongest response (442.85 R.U.), followed by ‘Awayuki’ (357.41 R.U.) and ‘Kotoka’ (218.08 R.U.). These results indicate that ‘Awayuki’ peel strongly fluoresces at longer wavelengths, ‘Pearl White’ peel shows high response at shorter wavelengths, ‘Kotoka’ flesh dominates at 440/685 nm, and ‘Pearl White’ flesh fluoresces most at 440/745 nm.

3.4. Prediction of Firmness, SSC, and TA Using EEM Fluorescence

The performance of the multiple linear regression (MLR) models, developed to predict strawberry quality attributes from 3 EEM fluorescence features found, is detailed in Table 4. The results indicate that TA and SSC were predicted with high accuracy for both skin and flesh tissues. Specifically, the models for TA demonstrated exceptional predictive power, yielding prediction coefficients of determination (R2pred) of 88.77% for skin and 86.99% for flesh, with very low corresponding root mean square errors (RMSEpred) of 0.05 and 0.07, respectively. Similarly, the SSC models performed strongly, explaining 84.84% of the variance in the skin and 81.56% in the flesh, with RMSEpred values of 0.59 and 0.76. The strong correlation between measured and predicted values for TA and SSC, as illustrated by data points clustering near the 1:1 line (Figure 5), confirms that fluorescence fingerprints are robust indicators of these chemical attributes. The prediction of firmness showed a notable dependency on tissue type. The model for skin firmness was highly effective, with an R2pred of 82.23% and a low RMSEpred of 0.33. In stark contrast, the model for flesh firmness was considerably weaker, accounting for only 43.77% of the variance (RMSEpred = 0.62), suggesting that fluorescence features from the flesh are less correlated with its textural properties. The maturity index (MI) was predicted with moderate success, with R2pred values of 70.73% for the skin and 67.61% for the flesh. Overall, these findings demonstrate that EEM fluorescence spectroscopy is a powerful tool for the non-destructive prediction of TA, SSC, and skin firmness in Japanese strawberries, with TA being the most accurately predicted attribute.

4. Discussion

The findings of this study unveil clear varietal distinctions among ‘Kotoka’, ‘Awayuki’, and ‘Pearl White’ strawberries through an integrative approach combining conventional physicochemical assessments and advanced non-destructive optical analyses. Measurements of SSC and TA were generally in line with the expected taste of each cultivar, but they did not perfectly match the impression given by color alone. ‘Kotoka’ contained the highest levels of both SSC and TA, producing a balanced but slightly sharper flavor. ‘Pearl White’ and ‘Awayuki’, on the other hand, were lower in sweetness and much lower in acidity, which led to considerably higher MI values. Among them, ‘Pearl White’ showed the highest MI, giving it a mild and sweet taste despite only moderate SSC, a profile consistent with its reputation as a delicately sweet, low-acid berry. ‘Awayuki’ showed intermediate values, while ‘Kotoka’, with higher acidity, gave it a noticeably tangier character despite its ample sugars. Although this aspect of the study was secondary, these flavor profiles support the optical findings. Cultivars with less pigmentation, like ‘Pearl White’ and ‘Awayuki’, tend to be perceived as sweeter and less acidic, whereas darker red varieties, such as ‘Kotoka’, often have higher acidity. Importantly, all three cultivars exhibited SSC/TA ratios above 7–8, a threshold frequently associated with high consumer acceptability, since it provides a desirable balance between sweetness and acidity [31,32]. However, consumer preference studies confirm that sweetness and flavor intensity are not determined solely by sugar-to-acid ratios, but also by the interaction of volatile compounds with sugars [10]. These findings are consistent with prior reports emphasizing the strong genetic regulation of sugar–acid balance in strawberries, with cultivar identity being the main driver of SSC and TA variability, even under different growing conditions [31,33]. For instance, [31] demonstrated that although the environmental factors studied (open field and high tunnel) modulate anthocyanin levels and acidity, with the exception of soluble solids, these continue to depend largely on genotype.
Image texture analysis using GLCM features offered additional insight into surface variation among the cultivars. In our images, ‘Pearl White’ had the highest contrast and dissimilarity values, indicating a more irregular surface pattern. Visually, this is understandable, as its white background and scattered red seed coats create strong pixel intensity differences. ‘Kotoka’, in contrast, showed the lowest contrast and more homogeneous texture metrics, reflecting its uniform red pigmentation that gives the peel an even appearance with fewer intensity changes. ‘Awayuki’ displayed intermediate values, in line with its lighter pinkish hue and moderate seed contrast. When the fruits were sliced, surface variability largely disappeared across all cultivars. The internal tissue of ‘Pearl White’ was particularly light (L* > 210) and uniform, while the flesh of ‘Kotoka’, though lighter than its peel, remained relatively darker with L* close to 174 and showed some residual heterogeneity. Overall, the reduction in contrast and increase in homogeneity after slicing confirm that most texture differences are associated with the peel, while the interior is more uniform, especially in pale-fruited varieties where the flesh remains white or light pink [34]. Importantly, similar GLCM descriptors have proven effective for fruit classification and grading in other commodities with subtly varying peels. These patterns are consistent with those reported by [15], who demonstrated that co-occurrence texture provides valuable information in conjunction with color and spectra, and when combined with them, substantially improves the prediction of internal quality attributes [35], which further demonstrated that textural descriptors sensitively track surface changes during storage, reinforcing our observation that texture differences diminish once the peel is removed and highlighting GLCM’s value for monitoring early deterioration. Complementarily, Ref. [36] reported that GLCM features derived from standard digital images reliably discriminate maturity classes in strawberries, underscoring their generalizability to practical imaging setups.
Color analysis using both RGB channel values and CIELAB parameters further supported the visual distinctions among cultivars. Under white LED illumination, the color differences among the cultivars appeared as expected and reflected their pigment composition. ‘Pearl White’ showed the highest lightness (L* ≈ 185) and the lowest a* value (≈ 133), giving it a pale appearance with only a faint blush of color. ‘Kotoka’, by contrast, was much darker, with the lowest lightness (L* ≈ 60) and the highest a* value (≈170), consistent with its rich anthocyanin content producing a deep red hue. ‘Awayuki’ fell between these extremes, with moderate L* and a* values reflecting its peach-pink coloration. These differences directly relate to anthocyanin content, as these pigments are primarily responsible for the characteristic red coloration of strawberries [37], which is why their abundance in ‘Kotoka’ results in a rich, intense red color, while their relative absence in ‘Pearl White’ gives the fruit its white to pale pink appearance.
RGB analysis under controlled LED illumination revealed similar cultivar-specific signatures. Under 365 nm light, both ‘Pearl White’ and ‘Awayuki’ displayed a bright blue-green fluorescence across their surfaces, standing out sharply against the much darker ‘Kotoka’ fruit. Image analysis confirmed this visual contrast: the blue channel intensity for ‘Awayuki’ reached about 160, while ‘Pearl White’ measured around 139, compared to only about 37 for ‘Kotoka’. This strong UV-induced glow in the paler varieties is consistent with their higher content of fluorescent phenolic compounds. Previous studies have shown that this blue-green autofluorescence in strawberries is closely linked to the presence of hydroxycinnamic acid derivatives in the peel [38]. Kotoka’s deep red peel, on the other hand, appeared much darker under UV illumination. This is likely due to its high anthocyanin content, which absorbs a significant portion of both the UV excitation and the resulting emitted light. Anthocyanins are well known to act as internal light filters in pigmented fruits, limiting the excitation of chlorophyll and other naturally fluorescent compounds [39,40]. Similar trends were observed under 420 nm LED lighting. ‘Pearl White’ continued to produce the strongest blue-green reflectance and fluorescence, with the highest blue channel values, while ‘Kotoka’ remained much dimmer overall but still showed a pronounced red tone, reflected in its higher R channel and a* values. ‘Awayuki’ displayed intermediate characteristics, combining moderate blue-green brightness with a noticeable red component. These patterns suggest that even at 420 nm excitation, the absence of strong red pigmentation in ‘Pearl White’ allows more light scattering and blue-green fluorescence, whereas the anthocyanins in ‘Kotoka’ continue to dominate its appearance, enhancing redness but suppressing fluorescence. Importantly, these imaging observations are consistent with the EEM findings that ‘Pearl White’ and ‘Awayuki’, which contain more non-anthocyanin phenolics, produce stronger UV-induced fluorescence, while Kotoka’s anthocyanin richness results in reduced fluorescence, a well-known effect of pigment-induced fluorescence quenching [41].
EEM fluorescence spectroscopy provided biochemical specificity to these observations by revealing distinct excitation/emission peaks associated with different pigment families and secondary metabolites. In the case of ‘Pearl White’, it stood out, showing the strongest fluorescence response in the near-UV region. Its peel exhibited a distinct peak at 290 nm excitation and 325 nm emission with an intensity of about 112 R.U., nearly twice that measured in the red fruit ‘Kotoka’. This pronounced signal suggests that Pearl White’s peel contains higher concentrations of colorless phenolic compounds, particularly hydroxycinnamic acid derivatives, which are well known to produce blue-green fluorescence when excited by UV light [42,43]. In contrast, the peach-colored ‘Awayuki’ cultivar displayed remarkably strong fluorescence at longer wavelengths. When excited at 490 nm, its peel produced a prominent emission peak at 745 nm, reaching an intensity of about 1494 R.U., nearly ten times greater than the 87 R.U. observed in ‘Pearl White’. This far-red fluorescence is typically associated with residual chlorophyll or related pigment compounds that remain present in the fruit tissue [44]. By comparison, the red ‘Kotoka’ cultivar showed much weaker fluorescence at the 490/745 nm peak. This reduction is consistent with its high anthocyanin content, which can act as an internal light filter and suppress chlorophyll fluorescence [40]. The fluorescence signatures of the inner flesh varied noticeably among the three cultivars. ‘Kotoka’ displayed a strong peak at 440/685 nm, whereas ‘Pearl White’ produced its highest response at 440/745 nm. This difference suggests that Kotoka’s anthocyanin pigments extend into the flesh, giving rise to stronger emissions in the mid-red region, while ‘Pearl White’ and ‘Awayuki’ retain greater far-red signals, likely linked to traces of chlorophyll or other fluorescent compounds. Overall, these findings indicate that paler varieties, which lack high anthocyanin levels, contain more UV-responsive phenolics and chlorophyll-related substances. In contrast, the deeper pigmentation of ‘Kotoka’ produces a distinct fluorescence pattern, shaped by its anthocyanin content and the resulting reduction in certain emissions [43].
In addition to cultivar differences, fluorescence characteristics can also change with fruit storage (Table 5). Previous studies demonstrated that fresh strawberries exhibit a single intrinsic fluorescence peak (Ex 250–300 nm, Em 300–400 nm), whereas after storage, an additional peak emerges (Ex 310–395 nm, Em 370–565 nm), indicating compositional changes during storage [29,30]. The fluorescence signal in the 250–300 nm excitation/300–400 nm emission range is most likely linked to amino acids, while the appearance of a second blue-green fluorescence band (Ex 310–395 nm, Em 370–565 nm) has been attributed primarily to p-coumaric acid and its glycosides [31,32]. Compounds such as certain B vitamins [33], flavonoids [45], and ellagic acid [35] may also influence fluorescence in this region. These literature findings underscore that changes in specific metabolites due to storage or genetic differences can alter the fluorescence fingerprint, complementing our observation of distinct cultivar-specific EEM profiles.
This study demonstrates that integrating physicochemical measurements with non-destructive fluorescence spectroscopy provides a powerful, multidimensional tool for the quality assessment of Japanese strawberries. The predictive models for TA and SSC were particularly robust, achieving high prediction accuracies (R2pred > 81%) for both skin and flesh tissues. The exceptional performance for TA (R2pred ≈ 87–89%), which surpassed all other parameters, confirms that EEM fluorescence is highly sensitive to the biochemical compounds governing acidity. A key finding was the differential predictive accuracy for firmness. While the model for firmness by using EEMs of the skin part was highly effective (R2pred = 82.23%), the prediction for firmness by using Flesh’s EEM was notably poor (R2pred = 43.77%). This disparity suggests that the fluorescence signal primarily captures surface-level textural properties and is less indicative of the internal flesh structure. The maturity index (MI) was predicted with moderate success (R2pred ≈ 68–71%), indicating a more complex relationship with the measured spectral features. Our results are in accordance with previous studies on the prediction of freshness level from firmness attribute of kiwifruit, green bell pepper, with the R2 prediction of above 80% [23,25,26]. These validated, high-performing models for TA, SSC, and firmness from skin’s EEMs fingerprint have significant potential for practical applications, such as in breeding programs for rapid phenotyping, for post-harvest quality monitoring, and in the development of optical systems for sorting premium cultivars. However, the study’s primary limitation is its focus on only three Japanese cultivars. Genetic background and cultivation conditions can significantly alter fruit chemistry and spectral profiles, likely requiring model recalibration for other varieties. Furthermore, the influence of maturity stage and storage duration was not assessed, and these factors could alter the fluorescence patterns. Future research should therefore expand to a broader range of cultivars and incorporate environmental and post-harvest variables to build more universally applicable and robust prediction models.

Author Contributions

Conceptualization, D.F. and N.K.; methodology, D.F., N.K., A.F.-N., M.H. and L.R.; software, D.F., N.C., A.F.-N., M.H., N. and L.R.; validation, D.F., N.K., A.F.-N., M.H., N.C. and L.R.; formal analysis, D.F., N.C., A.F.-N., M.H. and L.R.; investigation, M.H. and N.; resources, D.F., N.K., M.H., N. and L.R.; data curation, D.F., N.C., A.F.-N., M.H., N. and L.R.; writing—original draft preparation, D.F., N.C., A.F.-N., M.H. and L.R.; writing—review and editing, D.F., N.C., A.F.-N.; visualization, D.F., N.C., A.F.-N., M.H. and L.R.; supervision, D.F. and N.K.; project administration, M.H. and D.F. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Data Availability Statement

The raw data supporting the conclusions of this article will be made available by the authors on request.

Acknowledgments

Maulidia Hilaili extends her most profound gratitude to the Asian Development Bank-Japan Scholarship Program (ADB-JSP) for their generous financial support, enabling her to pursue a master’s degree. Nurwahyuningsih sincerely thanks Japan’s Ministry of Education, Culture, Sports, Science and Technology (MEXT) for their generous financial support through the MEXT Scholarship program. This research would not have been possible without their funding. She profoundly appreciates the opportunity ADB-JSP provided to advance her education and contribute to her field of study. Noelia Castillejo is grateful for her research contract, which financed by Beatriz Galindo Program (BG23-00121) from the Spanish Ministry of Science and Innovation.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Parvez, S.; Wani, I.A. Postharvest Biology and Technology of Strawberry. In Postharvest Biology and Technology of Temperate Fruits; Springer International Publishing: Berlin/Heidelberg, Germany, 2018; pp. 331–348. [Google Scholar] [CrossRef]
  2. De Simone, N.; Scauro, A.; Fatchurrahman, D.; Russo, P.; Capozzi, V.; Spano, G.; Fragasso, M. Inclusion of Antifungal and Probiotic Lactiplantibacillus plantarum Strains in Edible Alginate Coating as a Promising Strategy to Produce Probiotic Table Grapes and Exploit Biocontrol Activity. Horticulturae 2024, 10, 4. [Google Scholar] [CrossRef]
  3. De Simone, N.; Scauro, A.; Fatchurrahman, D.; Amodio, M.L.; Capozzi, V.; Colelli, G.; Spano, G.; Fragasso, M.; Russo, P. Probiotic Lactiplantibacillus plantarum strains showing anti-Botrytis activity: A food-grade approach to improve the overall quality of strawberry in post-harvest. Postharvest Biol. Technol. 2024, 218, 113125. [Google Scholar] [CrossRef]
  4. FAOSTAT. Available online: https://www.fao.org/faostat/en/#data/QCL/visualize (accessed on 5 September 2025).
  5. Crecente-Campo, J.; Nunes-Damaceno, M.; Romero-Rodríguez, M.A.; Vázquez-Odériz, M.L. Color, anthocyanin pigment, ascorbic acid and total phenolic compound determination in organic versus conventional strawberries (Fragaria × ananassa Duch, cv Selva). J. Food Compos. Anal. 2012, 28, 23–30. [Google Scholar] [CrossRef]
  6. Palma, C.; Morales-Quintana, L.; Ramos, P. Phenolic Content, Color Development, and Pigment−Related Gene Expression: A Comparative Analysis in Different Cultivars of Strawberry During the Ripening Process. Agronomy 2020, 10, 588. [Google Scholar] [CrossRef]
  7. Chen, Y.; Belwal, T.; Xu, Y.; Ma, Q.; Li, D.; Li, L.; Xiao, H.; Luo, Z. Updated insights into anthocyanin stability behavior from bases to cases: Why and why not anthocyanins lose during food processing. Crit. Rev. Food Sci. Nutr. 2023, 63, 8639–8671. [Google Scholar] [CrossRef]
  8. Basu, A.; Nguyen, A.; Betts, N.; Lyons, T. Strawberry As a Functional Food: An Evidence-Based Review. Crit. Rev. Food Sci. Nutr. 2014, 54, 790–806. [Google Scholar] [CrossRef]
  9. Giampieri, F.; Forbes-Hernandez, T.Y.; Gasparrini, M.; Alvarez-Suarez, J.M.; Afrin, S.; Bompadre, S.; Quiles, J.L.; Mezzetti, B.; Battino, M. Strawberry as a health promoter: An evidence based review. Food Funct. 2015, 6, 1386–1398. [Google Scholar] [CrossRef]
  10. Schwieterman, M.; Colquhoun, T.A.; Jaworski, E.A.; Bartoshuk, L.M.; Gilbert, J.L.; Tieman, D.M.; Odabasi, A.Z.; Moskowitz, H.R.; Folta, K.M.; Klee, H.J.; et al. Strawberry Flavor: Diverse Chemical Compositions, a Seasonal Influence, and Effects on Sensory Perception. PLoS ONE 2014, 9, e88446. [Google Scholar] [CrossRef]
  11. Šamec, D.; Maretić, M.; Lugarić, I.; Mešić, A.; Salopek-Sondi, B.; Duralija, B. Assessment of the differences in the physical, chemical and phytochemical properties of four strawberry cultivars using principal component analysis. Food Chem. 2016, 194, 828–834. [Google Scholar] [CrossRef]
  12. Alvarez-Suarez, J.; Mazzoni, L.; Forbes, T.; Gasparrini, M.; Sabbadini, S.; Giampieri, F. The effects of pre-harvest and post-harvest factors on the nutritional quality of strawberry fruits: A review. J. Berry Res. 2014, 4, 10. [Google Scholar] [CrossRef]
  13. Błaszczyk, J.; Bieniasz, M.; Nawrocki, J.; Kopeć, M.; Mierzwa-Hersztek, M.; Gondek, K.; Zaleski, T.; Knaga, J.; Bogdał, S. The Effect of Harvest Date and Storage Conditions on the Quality of Remontant Strawberry Cultivars Grown in a Gutter System under Covers. Agriculture 2022, 12, 1193. [Google Scholar] [CrossRef]
  14. Mancini, M.; Mazzoni, L.; Gagliardi, F.; Balducci, F.; Duca, D.; Toscano, G.; Mezzetti, B.; Capocasa, F. Application of the Non-Destructive NIR Technique for the Evaluation of Strawberry Fruits Quality Parameters. Foods 2020, 9, 441. [Google Scholar] [CrossRef] [PubMed]
  15. Weng, S.; Yu, S.; Guo, B.; Tang, P.; Liang, D. Non-Destructive Detection of Strawberry Quality Using Multi-Features of Hyperspectral Imaging and Multivariate Methods. Sensors 2020, 20, 3074. [Google Scholar] [CrossRef] [PubMed]
  16. Qin, J.; Chao, K.; Kim, M.S.; Lu, R.; Burks, T.F. Hyperspectral and multispectral imaging for evaluating food safety and quality. J. Food Eng. 2013, 118, 157–171. [Google Scholar] [CrossRef]
  17. Fathi-Najafabadi, A.; Besada, C.; Gil, R.; Calatayud, M.A.; Salvador, A. Chlorophyll fluorescence imaging as a tool to evaluate calyx senescence during the ripening of persimmon fruit treated with gibberellic acid. Postharvest Biol. Technol. 2021, 179, 111582. [Google Scholar] [CrossRef]
  18. Momin, A.; Kondo, N.; Al Riza, D.; Ogawa, Y.; Obenland, D. A Methodological Review of Fluorescence Imaging for Quality Assessment of Agricultural Products. Agriculture 2023, 13, 1433. [Google Scholar] [CrossRef]
  19. Sikorska, E.; Nowak, P.; Pawlak-Lemanska, K.; Sikorski, M. Characterization and Classification of Direct and Commercial Strawberry Beverages Using Absorbance–Transmission and Fluorescence Excitation–Emission Matrix Technique. Foods 2022, 11, 2143. [Google Scholar] [CrossRef]
  20. Aparicio, R.; Sayago, A.; Morales, M. Detection of hazelnut oil in virgin olive oil by a spectrofluorimetric method. Eur. Food Res. Technol. 2004, 218, 480–483. [Google Scholar] [CrossRef]
  21. Lötze, E.; Huybrechts, C.; Sadie, A.; Theron, K.; Valcke, R. Fluorescence imaging as a non-destructive method for pre-harvest detection of bitter pit in apple fruit (Malus domestica Borkh.). Postharvest Biol. Technol. 2006, 40, 287–294. [Google Scholar] [CrossRef]
  22. Fatchurrahman, D. Identification of UV-Fluorescence Components Associated with and Detection of Surface Damage in Green Pepper (Capsicum annum L). 2015. Available online: https://www.researchgate.net/publication/301285323 (accessed on 19 May 2025).
  23. Fatchurrahman, D.; Kuramoto, M.; Al Riza, D.F.; Ogawa, Y.; Suzuki, T.; Kondo, N. Fluorescence time series monitoring of different parts of green pepper (Capsicum annuum L.) under different storage temperatures. Comput. Electron. Agric. 2020, 179, 105850. [Google Scholar] [CrossRef]
  24. Fatchurrahman, D.; Castillejo, N.; Hilaili, M.; Russo, L.; Fathi-Najafabadi, A.; Rahman, A. A Novel Damage Inspection Method Using Fluorescence Imaging Combined with Machine Learning Algorithms Applied to Green Bell Pepper. Horticulturae 2024, 10, 1336. [Google Scholar] [CrossRef]
  25. Fatchurrahman, D.; Hilaili, M.; Nurwahyuningsih; Russo, L.; Jahari, M.B.; Fathi-Najafabadi, A. Utilizing RGB imaging and machine learning for freshness level determination of green bell pepper (Capsicum annuum L.) throughout its shelf-life. Postharvest Biol. Technol. 2025, 222, 113359. [Google Scholar] [CrossRef]
  26. Nurwahyuningsih; Fatchurrahman, D.; Hilaili, M.; Chosa, T. Assessment of fluorescence fingerprint in Hayward and Ruby-Red kiwi fruit as a potential approach to monitor quality changes during storage. Food Control 2025, 178, 111528. [Google Scholar] [CrossRef]
  27. Lawaetz, A.; Stedmon, C. Fluorescence Intensity Calibration Using the Raman Scatter Peak of Water. Appl. Spectrosc. 2009, 63, 936–940. [Google Scholar] [CrossRef]
  28. Fathi-Najafabadi, A.; Besada, C.; Gil, R.; Salvador, A. Harvest time and postharvest behavior of six japanese nonastringent persimmon cultivars grown under mediterranean conditions. HortScience 2020, 55, 1766–1771. [Google Scholar] [CrossRef]
  29. Fatchurrahman, D.; Amodio, M.L.; de Chiara, M.L.V.; Chaudhry, M.M.A.; Colelli, G. Early discrimination of mature-and immature-green tomatoes (Solanum lycopersicum L.) using fluorescence imaging method. Postharvest Biol. Technol. 2020, 169, 111287. [Google Scholar] [CrossRef]
  30. Corvino, A.; Romaniello, R.; Palumbo, M.; Ricci, I.; Cefola, M.; Pelosi, S.; Pace, B. Image analysis to predict the maturity index of strawberries. Adv. Hortic. Sci. 2023, 37, 83–87. [Google Scholar] [CrossRef]
  31. Patel, H.; Taghavi, T.; Samtani, J.B. Fruit Quality of Several Strawberry Cultivars during the Harvest Season under High Tunnel and Open Field Environments. Horticulturae 2023, 9, 1084. [Google Scholar] [CrossRef]
  32. Wozniak, W.; Radajewka, B.; Reszelska-Sieciechowicz, A.; Dejwor, I. Sugar and acid content influence organoleptic evaluation of fruits of six strawberry cultivars from controlled cultivation. Acta Hortic. 1997, 439, 333–336. [Google Scholar] [CrossRef]
  33. Ikegaya, A. Composition of free sugars and organic acids in Japanese strawberry cultivars and their influence on the perception of sweetness and sourness. J. Food Sci. 2024, 89, 614–624. [Google Scholar] [CrossRef]
  34. Zubair, A.R.; Alo, O.A. Grey Level Co-occurrence Matrix (GLCM) Based Second Order Statistics for Image Texture Analysis. Int. J. Sci. Eng. Investig. 2019, 8, 93. [Google Scholar]
  35. Ladika, G.; Strati, I.; Tsiaka, T.; Cavouras, D.; Sinanoglou, V. On the Assessment of Strawberries’ Shelf-Life and Quality, Based on Image Analysis, Physicochemical Methods, and Chemometrics. Foods 2024, 13, 234. [Google Scholar] [CrossRef] [PubMed]
  36. Takemoto, R.; Watanabe, T.; Nakamura, N.; Koseki, S.; Koyama, K. Investigating the relationship between strawberry maturity and glossiness on the surface during storage through digital image analysis. J. Food Meas. Charact. 2023, 18, 1776–1785. [Google Scholar] [CrossRef]
  37. Voća, S.; Žlabur, J.Š.; Dobričević, N.; Jakobek, L.; Šeruga, M.; Galić, A.; Pliestić, S. Variation in the Bioactive Compound Content at Three Ripening Stages of Strawberry Fruit. Molecules 2014, 19, 10370–10385. [Google Scholar] [CrossRef]
  38. Razgonova, M.; Zinchenko, Y.; Pikula, K.; Tekutyeva, L.; Son, O.; Zakharenko, A.; Kalenik, T.; Golokhvast, K. Spatial Distribution of Polyphenolic Compounds in Corn Grains (Zea mays L. var. Pioneer) Studied by Laser Confocal Microscopy and High-Resolution Mass Spectrometry. Plants 2022, 11, 630. [Google Scholar] [CrossRef]
  39. Kapoor, L.; Simkin, A.J.; Doss, C.G.P.; Siva, R. Fruit ripening: Dynamics and integrated analysis of carotenoids and anthocyanins. BMC Plant Biol. 2022, 22, 27. [Google Scholar] [CrossRef]
  40. Merzlyak, M.N.; Melø, T.B.; Naqvi, K.R. Effect of anthocyanins, carotenoids, and flavonols on chlorophyll fluorescence excitation spectra in apple fruit: Signature analysis, assessment, modelling, and relevance to photoprotection. J. Exp. Bot. 2008, 59, 349–359. [Google Scholar] [CrossRef]
  41. Collings, D.A. Anthocyanin in the Vacuole of Red Onion Epidermal Cells Quenches Other Fluorescent Molecules. Plants 2019, 8, 596. [Google Scholar] [CrossRef]
  42. Morales, F.; Cartelat, A.; Álvarez-Fernández, A.; Moya, I.; Cerovic, Z.G. Time-Resolved Spectral Studies of Blue−Green Fluorescence of Artichoke (Cynara cardunculus L. Var. Scolymus) Leaves:  Identification of Chlorogenic Acid as One of the Major Fluorophores and Age-Mediated Changes. J. Agric. Food Chem. 2005, 53, 9668–9678. [Google Scholar] [CrossRef]
  43. Yoshioka, Y.; Nakayama, M.; Noguchi, Y.; Horie, H. Use of image analysis to estimate anthocyanin and UV-excited fluorescent phenolic compound levels in strawberry fruit. Breed. Sci. 2013, 63, 211–217. [Google Scholar] [CrossRef]
  44. Ma, T.; Inagaki, T.; Tsuchikawa, S. Development of a sensitivity-enhanced chlorophyll fluorescence lifetime spectroscopic method for nondestructive monitoring of fruit ripening and postharvest decay. Postharvest Biol. Technol. 2023, 198, 112231. [Google Scholar] [CrossRef]
  45. Zhang, Y.; Seeram, N.P.; Lee, R.; Feng, L.; Heber, D. Isolation and Identification of Strawberry Phenolics with Antioxidant and Human Cancer Cell Antiproliferative Properties. J. Agric. Food Chem. 2008, 56, 670–675. [Google Scholar] [CrossRef]
  46. Sikorska, E.; Włodarska, K.; Khmelinskii, I. Application of multidimensional and conventional fluorescence techniques for classification of beverages originating from various berry fruit. Methods Appl. Fluoresc. 2020, 8, 015006. [Google Scholar] [CrossRef]
  47. Zhang, J.; Wang, X.; Yu, O.; Tang, J.; Gu, X.; Wan, X.; Fang, C. Metabolic profiling of strawberry (Fragaria × ananassa Duch.) during fruit development and maturation. J. Exp. Bot. 2011, 62, 1103–1118. [Google Scholar] [CrossRef]
  48. Mattila, P.; Hellström, J.; Törrönen, R. Phenolic Acids in Berries, Fruits, and Beverages. J. Agric. Food Chem. 2006, 54, 7193–7199. [Google Scholar] [CrossRef]
  49. Sano, H.; Kawaguchi, S.; Iimori, T.; Kuragano, M.; Tokuraku, K.; Uwai, K. Using fluorescence spectra to quantitatively evaluate the constituent content and functionality of hydroponically grown Fragaria × Ananassa. Eur. Food Res. Technol. 2025, 251, 3087–3103. [Google Scholar] [CrossRef]
Horticulturae 11 01291 g001
Figure 1. The visual appearance of the three studied varieties of strawberry.
Figure 1. The visual appearance of the three studied varieties of strawberry.
Horticulturae 11 01291 g001
Horticulturae 11 01291 g002
Figure 2. Imaging system (left: white LED, middle: 365 nm LED, and right: 420 nm LED). The gray icon represents the camera, while the black and blue icons indicate the illumination sources. The double-headed arrows show the illumination paths from the LED light sources, sample and the camera, and the single-headed arrow indicates the position of the long pass filter. The green ellipse represents the fruit sample, and the light blue shaded regions illustrate the illuminated area. The distances between the sample and the illumination sources (25 cm) are shown by dashed lines.
Figure 2. Imaging system (left: white LED, middle: 365 nm LED, and right: 420 nm LED). The gray icon represents the camera, while the black and blue icons indicate the illumination sources. The double-headed arrows show the illumination paths from the LED light sources, sample and the camera, and the single-headed arrow indicates the position of the long pass filter. The green ellipse represents the fruit sample, and the light blue shaded regions illustrate the illuminated area. The distances between the sample and the illumination sources (25 cm) are shown by dashed lines.
Horticulturae 11 01291 g002
Horticulturae 11 01291 g003
Figure 3. Fluorescence image of skin (A) and flesh (B) part of the three strawberry varieties under three different light sources, 365 nm, 420 nm, and white LED lights.
Figure 3. Fluorescence image of skin (A) and flesh (B) part of the three strawberry varieties under three different light sources, 365 nm, 420 nm, and white LED lights.
Horticulturae 11 01291 g003
Horticulturae 11 01291 g004
Figure 4. EEM fluorescence spectra of three varieties (‘Kotoka’, ‘Awayuki’, and ‘Pearl White’) and two parts (peel and flesh) of the strawberry. The scale ranges from 0 to 1000 R.U.
Figure 4. EEM fluorescence spectra of three varieties (‘Kotoka’, ‘Awayuki’, and ‘Pearl White’) and two parts (peel and flesh) of the strawberry. The scale ranges from 0 to 1000 R.U.
Horticulturae 11 01291 g004
Horticulturae 11 01291 g005
Figure 5. Representative prediction plot of firmness (left) and SSC (right) obtained from multiple regression models using selected EEM fluorescence peaks of the skin part.
Figure 5. Representative prediction plot of firmness (left) and SSC (right) obtained from multiple regression models using selected EEM fluorescence peaks of the skin part.
Horticulturae 11 01291 g005
Table 1. Chemical parameters of the three studied varieties of strawberry: Kotoka (red cultivar), Awayuki (peach-colored cultivar), and Pearl White (white cultivar). Values are reported as average ± standard error (SE) from 180 fruit.
Table 1. Chemical parameters of the three studied varieties of strawberry: Kotoka (red cultivar), Awayuki (peach-colored cultivar), and Pearl White (white cultivar). Values are reported as average ± standard error (SE) from 180 fruit.
CultivarTSS
(°Brix)		TA
(%)		MIFirmness
(N)
Kotoka15.53 ± 0.23 a0.96 ± 0.01 a16.37 ± 0.51 c4.74 ± 0.4 a
Awayuki11.66 ± 0.31 c0.61 ± 0.01 b19.62 ± 1.11 b4.01 ± 0.2 b
Pearl White13.11 ± 0.25 b0.61 ± 0.01 b22.28 ± 0.96 a4.85 ± 0.3 a
Different letters indicate statistically significant differences in each parameter among the three varieties, as determined by the LSD test (p < 0.05).
Table 2. Features values for the RGB channel, CIELAB parameters (L*, a*, b*), and GLCM of three varieties of strawberries (‘Kotoka’, ‘Awayuki’, and ‘Pearl White’) under three different light sources (365 nm, 420 nm, and white LEDs). R, G, and B indicate red, green, and blue colors in digital imaging, respectively, and L*, a*, and b* show lightness, red-green, and yellow-blue components in the CIELAB color system, respectively. Texture GLCM measures contrast, showing intensity contrast between a pixel and its neighbor; dissimilarity measures the degree of difference between pairs of gray; homogeneity measures the uniformity of distribution of elements in the GLCM; angular second moment (ASM) measures the uniformity or concentration of the distribution of values in the GLCM; energy is the square root of ASM; correlation measures the linear dependency of gray levels in neighboring pixels. Lowercase letters denote significant differences among varieties and grades (p ≤ 0.05).
Table 2. Features values for the RGB channel, CIELAB parameters (L*, a*, b*), and GLCM of three varieties of strawberries (‘Kotoka’, ‘Awayuki’, and ‘Pearl White’) under three different light sources (365 nm, 420 nm, and white LEDs). R, G, and B indicate red, green, and blue colors in digital imaging, respectively, and L*, a*, and b* show lightness, red-green, and yellow-blue components in the CIELAB color system, respectively. Texture GLCM measures contrast, showing intensity contrast between a pixel and its neighbor; dissimilarity measures the degree of difference between pairs of gray; homogeneity measures the uniformity of distribution of elements in the GLCM; angular second moment (ASM) measures the uniformity or concentration of the distribution of values in the GLCM; energy is the square root of ASM; correlation measures the linear dependency of gray levels in neighboring pixels. Lowercase letters denote significant differences among varieties and grades (p ≤ 0.05).
CultivarB ChannelG ChannelR ChannelLabContrastDissimilarityHomogeneityEnergyCorrelationASM
Skin365 LED‘Kotoka’36.87 ± 17.17 c18.95 ± 4.72 c32.85 ± 5.28 c22.32 ± 7.19 c138.82 ± 4.14 a119.41 ± 7.63 a32.01 ± 9.48 c3.66 ± 0.45 c0.26 ± 0.01 a0.05 ± 0.01 a0.84 ± 0.09 b0.01 ± 0.01 a
‘Awayuki’160.17 ± 21.22 a111.28 ± 15.75 b64.85 ± 6.83 a115.91 ± 15.07 b128.09 ± 2.06 b97.11 ± 3.78 b64.86 ± 23.6 b4.42 ± 0.57 b0.26 ± 0.01 a0.03 ± 0.01 b0.98 ± 0.01 a0.01 ± 0.01 a
‘Pearl White’139.49 ± 16.95 b125.67 ± 13.25 a48.21 ± 5.01 b123.14 ± 12.33 a110.12 ± 2.54 c113.97 ± 4.21 a115.46 ± 28.35 a5.41 ± 0.37 a0.25 ± 0.01 a0.03 ± 0.01 b0.97 ± 0.01 a0.01 ± 0.01 a
420 LED‘Kotoka’2.57 ± 0.24 b26.19 ± 9.37 c15.97 ± 1.83 c22.46 ± 8.13 c120.27 ± 5.19 a138.88 ± 3.77 c45.81 ± 14.41 a3.57 ± 0.77 a0.33 ± 0.04 a0.08 ± 0.01 a0.82 ± 0.05 b0.01 ± 0.01 a
‘Awayuki’1.43 ± 0.19 b60.79 ± 7.12 b49.34 ± 2.73 a59.87 ± 6.83 b115.16 ± 3.19 b157.86 ± 2.75 b27.55 ± 7.91 a3.23 ± 0.21 a0.31 ± 0.01 a0.04 ± 0.01 b0.97 ± 0.01 a0.01 ± 0.01 b
‘Pearl White’10.85 ± 1.73 a105.66 ± 8.75 a36.15 ± 3.97 b98.26 ± 8.21 a90.16 ± 2.28 b169.05 ± 2.54 a47.96 ± 5.27 a4.03 ± 0.16 a0.27 ± 0.01 a0.04 ± 0.01 b0.98 ± 0.01 a0.01 ± 0.01 b
White LED‘Kotoka’1.36 ± 0.1 c9.25 ± 1.23 c116.83 ± 4.21 b60.54 ± 2.65 c170.23 ± 1.11 c162.04 ± 1.28 b59.15 ± 5.23 c3.13 ± 0.16 c0.47 ± 0.02 a0.06 ± 0.01 a0.91 ± 0.01 b0.01 ± 0.01 a
‘Awayuki’76.45 ± 3.47 b139.27 ± 3.18 b208.11 ± 2.01 a162.41 ± 2.46 b148.49 ± 1.04 b171.47 ± 1.08 a108.28 ± 6.34 b5.04 ± 0.17 b0.33 ± 0.01 b0.04 ± 0.01 b0.98 ± 0.01 a0.01 ± 0.01 b
‘Pearl White’125.268 ± 4.54 a173.57 ± 3.55 a204.96 ± 2.52 a185.01 ± 3.01 a133.41 ± 0.68 c157.15 ± 1.46 c187.47 ± 11.55 a5.79 ± 0.19 a0.36 ± 0.01 b0.04 ± 0.01 b0.98 ± 0.01 a0.01 ± 0.01 b
Flesh365 LED‘Kotoka’98.04 ± 32.38 c87.98 ± 14.63 c73.18 ± 7.59 a94.12 ± 15.22 c130.79 ± 3.29 a122.22 ± 10.95 a66.09 ± 25.25 a3.86 ± 0.34 a0.30 ± 0.01 b0.04 ± 0.01 a0.95 ± 0.01 a0.01 ± 0.01 a
‘Awayuki’216.013 ± 6.76 b166.30 ± 7.48 b94.38 ± 5.45 a166.92 ± 6.58 b120.55 ± 1.29 b96.18 ± 1.20 a42.58 ± 3.75 a3.16 ± 0.11 a0.34 ± 0.01 a0.05 ± 0.01 a0.99 ± 0.01 a0.01 ± 0.01 a
‘Pearl White’228.95 ± 5.21 a191.22 ± 4.60 a87.00 ± 4.44 a185.93 ± 4.24 a109.31 ± 0.45 c100.35 ± 0.54 a62.09 ± 2.45 a3,12 ± 0,04 a0.35 ± 0.01 a0.04 ± 0.01 a0,99± 0,01 a0,01 ± 0.01 a
420 LED‘Kotoka’2.7 ± 1.51 c56.34 ± 1.49 c55.93 ± 7.3859.18 ± 7.95 c121.56 ± 8.32 a157.57 ± 2.75 c109.43 ± 12.00 a3.95 ± 1.61 a0.32 ± 0.03 a0.05 ± 0.01 a0.92 ± 0.07 a0.01 ± 0.01 a
‘Awayuki’5.01 ± 2.23 b79.71 ± 6.69 b44.58 ± 9.2477.19 ± 7.05 b104.26 ± 3.36 b163.58 ± 2.01 b66.89 ± 97.05 a3.35 ± 1.44 a0.34 ± 0.04 a0.07 ± 0.02 a0.96 ± 0.03 a0.01 ± 0.01 a
‘Pearl White’14.05 ± 0.91 a117.62 ± 2.71 a34.17 ± 0.98111.01 ± 2.19 a86.81 ± 1.21 c172.01 ± 0.56 a43.06 ± 3.29 a3.56 ± 0.05 a0.31 ± 0.03 a0.05 ± 0.01 a0.98 ± 0.01 a0.01 ± 0.01 a
White LED‘Kotoka’79.32 ± 12.97 c141.95 ± 11.99 b229.60 ± 5.78 a173.59 ± 8.54 b156.10 ± 2.74 a174.51 ± 2.21 a83.01 ± 13.25 b4.82 ± 0.32 a0.33 ± 0.01 b0.05 ± 0.01 b0.99 ± 0.01 ab0.01 ± 0.01 b
‘Awayuki’139.46 ± 5.34 b195.49 ± 4.35 a236.41 ± 1.66 a207.71 ± 3.28 a135.94 ± 1.17 b161.77 ± 1.20 b89.61 ± 8.13 b3.22 ± 0.14 b0.47 ± 0.01 a0.07 ± 0.01 a0.99 ± 0.01 a0.01 ± 0.01 a
‘Pearl White’173.08 ± 7.91 a211.66 ± 5.91 a223.32 ± 5.39 a216.43 ± 5.34 a126.28 ± 0.29 c148.39 ± 1.42 c159.92 ± 11.54 a3.84 ± 0.24 b0.47 ± 0.01 a0.07 ± 0.01 a0.99 ± 0.01 b0.01 ± 0.01 ab
Table 3. EEM peaks and their intensity (R.U.) of three varieties (‘Kotoka’, ‘Awayuki’, and ‘Pearl White’) and two parts (peel and flesh) of the strawberry. Lowercase letters denote significant differences among varieties and grades (p ≤ 0.05).
Table 3. EEM peaks and their intensity (R.U.) of three varieties (‘Kotoka’, ‘Awayuki’, and ‘Pearl White’) and two parts (peel and flesh) of the strawberry. Lowercase letters denote significant differences among varieties and grades (p ≤ 0.05).
SkinFlesh
Ex/EmCultivarIntensity (R.U.)Ex/EmCultivarIntensity (R.U.)
290/325‘Kotoka’54.0691 c280/235‘Kotoka’274.673 a
‘Awayuki’74.235 b‘Awayuki’274.463 a
‘Pearl White’112.287 a‘Pearl White’274.653 a
340/435‘Kotoka’507.675 b440/685‘Kotoka’158.718 a
‘Awayuki’1184.78 a‘Awayuki’97.2218 c
‘Pearl White’73.7794 c‘Pearl White’134.736 b
490/745‘Kotoka’764.048 b440/745‘Kotoka’218.083 c
‘Awayuki’1493.76 a‘Awayuki’357.413 b
‘Pearl White’86.5641 c‘Pearl White’442.849 a
Table 4. Performance of multiple regression models for predicting total soluble solids SSC, TA, MI, and firmness in the skin and flesh part of the strawberry using selected EEM fluorescence peaks. R2: coefficient of determination; adjusted R2: adjusted coefficient of determination; RMSE: root mean square error; MAE: mean absolute error.
Table 4. Performance of multiple regression models for predicting total soluble solids SSC, TA, MI, and firmness in the skin and flesh part of the strawberry using selected EEM fluorescence peaks. R2: coefficient of determination; adjusted R2: adjusted coefficient of determination; RMSE: root mean square error; MAE: mean absolute error.
ParametersTissueR2 (%)
Cal		Adjusted R2 (%)
Cal		RMSE
Cal		MAE
Cal		R2 (%)
Pred		Adjusted R2 (%)
Pred		RMSE
Pred		MAE
Pred
SSCSkin83.4883.070.610.4984.8483.930.590.45
SSCFlesh81.7981.340.750.5881.5680.460.760.61
TASkin90.5490.310.050.0488.7788.100.050.04
TAFlesh88.8488.570.060.0486.9986.210.070.05
MISkin61.2560.301.911.3770.7368.591.851.42
MIFlesh65.4864.631.611.3167.6165.671.511.17
FirmnessSkin83.5184.850.410.2882.2381.160.330.21
FirmnessFlesh35.0333.430.560.4243.7740.390.620.51
Table 5. Fluorescence emission peaks found in the EEM of the strawberries.
Table 5. Fluorescence emission peaks found in the EEM of the strawberries.
NoExcitation (nm)Emission (nm)Suggested CompoundPartVarietyReferences
1250–300300–400Amino acidsFleshTroyonoka[46,47]
2310–395370–565Coumaric acid and glycosides (p-coumaric acid, p-hydroxybenzoic, gallic acid)FleshPolka and Jonsok[43,45,48]
3<350260–450
550–650		polyphenolsLeavesYotsuboshi[49]
4>350650–850chlorophyllsLeavesYotsuboshi[49]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Hilaili, M.; Fathi-Najafabadi, A.; Nurwahyuningsih; Castillejo, N.; Russo, L.; Kondo, N.; Fatchurrahman, D. Optical and Chemical Profiling of Japanese Strawberries: Fluorescence Fingerprints, Imaging Features, and Quality Attributes Prediction. Horticulturae 2025, 11, 1291. https://doi.org/10.3390/horticulturae11111291

AMA Style

Hilaili M, Fathi-Najafabadi A, Nurwahyuningsih, Castillejo N, Russo L, Kondo N, Fatchurrahman D. Optical and Chemical Profiling of Japanese Strawberries: Fluorescence Fingerprints, Imaging Features, and Quality Attributes Prediction. Horticulturae. 2025; 11(11):1291. https://doi.org/10.3390/horticulturae11111291

Chicago/Turabian Style

Hilaili, Maulidia, Ayoub Fathi-Najafabadi, Nurwahyuningsih, Noelia Castillejo, Lucia Russo, Naoshi Kondo, and Danial Fatchurrahman. 2025. "Optical and Chemical Profiling of Japanese Strawberries: Fluorescence Fingerprints, Imaging Features, and Quality Attributes Prediction" Horticulturae 11, no. 11: 1291. https://doi.org/10.3390/horticulturae11111291

APA Style

Hilaili, M., Fathi-Najafabadi, A., Nurwahyuningsih, Castillejo, N., Russo, L., Kondo, N., & Fatchurrahman, D. (2025). Optical and Chemical Profiling of Japanese Strawberries: Fluorescence Fingerprints, Imaging Features, and Quality Attributes Prediction. Horticulturae, 11(11), 1291. https://doi.org/10.3390/horticulturae11111291

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop