Ripeness Level Research Articles

Manipulator with Integrated Flexible Tactile Sensing Arrays for Kiwifruit Ripeness and Size Classification.

Fruit grading for ripeness and size is an essential process in the supply chain. Incorrect grading can easily lead to spoiled and degraded fruits entering the market, reducing consumers' confidence in purchasing. At the same time, it is easy to cause the fruit supply chain to reduce profits, unreasonable resource allocation, and related practitioners' income. The current mainstream machine vision grading and manual grading in the production line have dilemmas such as susceptibility to environmental interference, inconsistent grading standards, high cost, and labor shortage. To overcome these problems, this study proposes an integrated flexible tactile sensing array (3 × 4) manipulator for efficient, stable, low-cost, and accurate ripeness and size grading of kiwifruit. The flexible sensing manipulator grasps the kiwifruit, detects the hardness of the kiwifruit by relying on tactile sensing, and determines the ripeness level based on the hardness. The size of the kiwifruit is also differentiated according to whether there is a significant change in the resistance of the topmost sensing unit of the flexible pressure sensor array. The 0, 1, 2, 3, 4, and 5 anomalies that may occur in actual production were tested and combined with machine learning KNN, SVM, and RF algorithms for data modeling and grading. The results show that the lowest accuracy of 0, 1, 2, 3, 4, and 5 possible outliers is 86.67% (KNN), 95.83% (SVM), and 92.5% (RF), respectively. KNN has the lowest classification effect, and SVM has the best. This study overcomes the drawbacks of inefficient destructive detection and unstable manual detection and makes up for the vulnerability of single machine vision to interference from environmental factors. This study can alleviate the challenges caused by fruit wastage and promote the sustainable production and consumption of the fruit industry chain.

Olfactory Visualization Sensing Array Made with CelluMOFs to Predict Fruit Ripeness Using Deep Learning.

Developing a colorimetry-based artificial scent screening system (i.e., an olfactory visual sensing system) with high sensitivity and accurate pattern recognition for detecting fruit ripeness remains challenging. In this work, we construct a flexible dye/CelluMOFs-based sensor array with improved sensitivity for on-site detection of characteristic gases of fruits and integrate a densely connected convolutional network (DenseNet) into the sensor array, enabling it to recognize unique scent fingerprints and categorize the ripeness of fruits. In the system, CelluMOFs are synthesized through in situ growth of γ-cyclodextrin metal-organic frameworks (γ-CD-MOFs) on flexible fiber filter paper to fabricate a uniform, flexible and porous dye/CelluMOFs sensitive membrane. Compared to the pristine filter paper, the CelluMOFs exhibit increased porosity with a 62 times higher specific surface area and a 3-fold increase in dye loading capacity after 12 h of adsorption. The prepared dye/CelluMOFs sensing film shows outstanding mechanical and detection stability with negligible deviation after 100 cycles of rubbing. The colorimetric visualization arrays with multiple colorimetric dye/CelluMOFs chips, enable the sensitive recognition and detection of nine kinds of characteristic fruit odors and achieve a high response at 8-1500 ppm of trans-2-hexenal, showcasing remarkably low gas detection thresholds. On the basis of the ppm-level limit of detection with high sensitivity, the fabricated colorimetric sensor arrays are typically used for in situ assessment of fruit ripeness by integrating DenseNet. This approach achieves a satisfactory classification accuracy of 99.09% on the validation set, enabling high-precision prediction of fruit ripeness levels.

Identifikasi Kematangan Buah Pisang Berdasarkan Variasi Jarak Menggunakan Metode K-Nearest Neighbor

This research aims to identify the level of ripeness of kepok bananas based on the color of their skin using the K-Nearest Neighbor (K-NN) method. Bananas are an important commodity in Indonesia, and various ripeness levels need to be identified. The current process of identifying banana ripeness is still done manually, which requires a lot of labor and tends to be subjective. The K-NN method is used to classify bananas based on their skin color. This research involves the collection of banana images with three ripeness levels (raw, ripe, and overripe) and the extraction of RGB color features from these images. Three distance methods, namely Euclidean, Minkowski, and Manhattan, are also employed to compare accuracy results. The evaluation results of this research show that the accuracy value for the Euclidean distance method is 84%, the Minkowski distance method is 82%, and the Manhattan distance method is 80%. Thus, the findings indicate that the K-NN method and the Euclidean distance method provide good results in identifying the ripeness level of bananas. By implementing the K-NN algorithm, this research attempts to address the weaknesses of the time-consuming and subjective manual identification process, with the hope of providing a more accurate and efficient solution for the banana industry. The results of this research can be used to automate the identification process of banana ripeness levels and improve efficiency in banana sorting. It is expected that this research can provide practical benefits to the community and serve as a basis for further research in this field.

Perbandingan Algoritma CART dan C.4 5 Pada Citra Tandan Buah Sawit Untuk Mengetahui Tingkat Kematangan Dalam Penentuan Harga

Information technology is a means and infrastructure of a system method to organize, send, interpret, use, process, obtain, and store data in a meaningful and useful way. Oil palm is a tropical plant originating from West Africa. The advantage of this plant is that it can also be planted outside its place of origin, including Indonesia. This plant has been widely cultivated in the form of plantations and factories in various regions in Indonesia. Oil palm is an industrial plant that is used as a raw material for vegetable oil, industrial oil, and fuel. Oil palm is important for Indonesia because it creates jobs for local people and is a source of foreign currency for the country. Oil palm plants begin to flower and form fruit after 2-3 years. The fruit will ripen about 5-6 months after pollination. The ripening process of oil palm fruit can be seen from the change in color of the fruit's skin. The fruit will turn orange-red when ripe. When the fruit is ripe, the oil content in the fruit flesh is at its maximum. If it is too ripe, the oil palm fruit will fall from the stalk of the bunch. This study aims to assess the maturity of a bunch of oil palm fruit. The methods used in this study are CART and C.4 5. Each method has several stages that will produce entropy and gain values ​​that will later form a decision tree. The dataset consists of 37 data consisting of 10 criteria originating from Ramp 789 Batang Peranap. Based on the implementation of the C4.5 algorithm and the CART algorithm in determining the level of ripeness of oil palm fruit bunches on RAMP 789 Batang Peranap which produces an accuracy of 98.00%. These results are obtained based on Process data with testing using the RapidMiner application, which produces a Decision tree that is useful as a reference for decisions in determining whether or not oil palm fruit bunches are ripe, which so far have only been predicted.

Detection of new oil palm fruit bunches based on deep learning

Abstract The Indonesian oil palm plantation experiences a production rate of fresh fruit bunches and crude palm oil (CPO) influenced by the quality of fruits with excellent and uniform maturity accepted by the palm oil processing factory. The uniformity of fruit quality produced from the harvesting process through fruit maturity criteria, namely fractions 00, fraction 1, fraction 3, and fraction 4, necessitates the need for maturity detection based on deep learning with a pre-trained YOLOv5 model capable of detecting the maturity of new oil palm fruit bunches. The aim is to develop a deep learning-based maturity detection model based on visual images and analyse the deep learning algorithm’s performance by classifying the maturity level of new oil palm fruit bunches. The method collects a dataset of new oil palm fruit bunches with 4180 images. The collected dataset will be annotated with Roboflow, so the annotation results will undergo training, validation, and testing processes using the deep learning script. The result obtained for the data found to detect the ripeness levels of Fresh Fruit Bunches (FFB), shows that the YOLOv5 model demonstrates strong performance with an accuracy reaching 72.5% in testing, which falls into the category of fair annotation. These results indicate that the system created and used is functioning well.

Classification of Avocado Ripeness Levels Using CNN Method

This research aims to develop a model for classifying the ripeness level of avocados using the Convolutional Neural Network (CNN) algorithm. The dataset comprises images of avocados categorized into three classes: unripe, ripe, and overripe. The CNN model is trained to classify the images into one of these three categories. The results indicate that the developed model can classify avocado images with high accuracy. The primary tool used for developing and implementing this method is MATLAB R2022a. The CNN algorithm is utilized to recognize and classify the ripeness level of avocados. This process involves several image processing steps, starting with preprocessing, image enhancement, and segmentation to isolate the avocado area. The dataset used in this research consists of 452 images distributed in 3 classes (unripe with 142, ripe with 66, and rotten with 244), with 80% used for training and 20% for testing. After 10 accuracy tests, the results indicate an accuracy rate of 90%. Additionally, features extracted from the images include color, shape, size, and texture characteristics, such as Mean, Standard Deviation, Kurtosis, Skewness, Variance, Entropy Value, Maximum Pixel, and Minimum Pixel. This research contributes to the field of agricultural technology by providing a robust method for the automatic classification of avocado ripeness. The findings are expected to facilitate accurate and efficient recognition of avocado ripeness, thereby supporting agricultural practices and market operations. Future research could explore the use of data augmentation techniques to further improve the accuracy and generalization of this model.

Determination of the Effects of Some Post-Harvest Treatments on the Quality of Banana Fruits During Storage and After Ripening

In this study, it was aimed to impact the effect of different post-harvest treatments on the shelf life of banana fruits of 'Dwarf Cavendish' banana cultivar after storage and ripening. 1) Modified atmosphere packaging (MAP), 2) 1-methylcyclopropene (1-MCP) + MAP, 3) Three different treatments (MAP + ethylene absorbent (EA)) were applied and only those using macro-aperture PE packaging were considered as control. Banana fruits were stored at 13°C and 90% RH for 30 days. Quality analyses were performed on samples taken at 10-day intervals. After 30 days of storage, the banana fruits were ripened to the ripeness level of No.3 and quality changes were determined after 4 days of shelf life. The treatments significantly limited the color change, softening of the peel and flesh, increased total soluble solids (TSS), ethylene secretion and respiration rate, especially in the last period of storage compared to the control. Green bananas treated with 1-MCP + MAP showed the best results during storage, but ripening problems were experienced. MAP + EA treatment delayed the coloration of banana fruits, softening of the flesh and increases in the TSS content during shelf life. MAP + EA treatment gave the best results both in terms of storage of banana fruits at green maturity and their resistance after ripening.

Assessing chemical composition variations on vintage, vineyard location and grape ripeness: A regional study of red wines produced in the colorado river middle basin, La Pampa (Argentina)

This work aimed to characterize chemical composition of Malbec and Cabernet Franc red wines from two different regions of La Pampa’s province, which stands out as one of the youngest wine-producing regions in Argentina and has remained unexplored until now. These wines were produced from grapes of two consecutive vintages (2021 and 2022), and were harvested at two different ripeness levels. Global phenolic parameters, colour characterization and multi-elemental profile were evaluated. Multivariate statistical approaches were successfully implemented to investigate the vintage and ripeness effects on chemical profile of red wines and characterize them based on regional variations. Our results showed that vintage and ripeness affected both phenolic and multi-elemental profiles, and such chemical variables allowed the wines differentiation according to region. The results of this first study could form the basis for future research into optimizing the region’s winemaking potential.

Vine spacing of <i>Vitis vinifera</i> cv. Shiraz/101-14 Mgt. III. Grape and wine quality

Vine spacing effects of Shiraz/101-14 Mgt on a relatively high potential soil in the Breede River Valley, Robertson, South Africa, on grape composition and wine quality, were investigated. Canopies were VSP trellised and orientated approximately NNE-SSW. Spacing between rows was fixed at 2.2 m. In-row vine spacing changed from 0.3 – 4.5 m with increments of 30 cm (from 15151 – 1010 vines/ha). Grape composition and wine quality were monitored over six seasons and two grape ripeness levels. Decreasing trends of soluble solids (°B) and pH and increasing titratable acidity from narrow to wide spacing were observed. Sugar accumulation of wider spacings seemed delayed. Berry skin total anthocyanin index and phenol content showed virtually no change. Malvidin mono-glucoside and its p-coumaroyl and acetyl derivatives were present in the highest concentrations, followed by peonidin-, petunidin-, delphinidin- and cyanidin mono-glucosides and derivatives, each vine spacing treatment displaying a unique grape anthocyanin profile. The narrower spacing of vines seemed to favour skin anthocyanin accumulation compared to wider spacing. Decreasing trends in wine total anthocyanin intensity, anthocyanin density, and phenols were observed from narrow to wide spacing; all had higher levels in wines from riper grapes. Individual wine anthocyanin concentrations also showed qualitative profile changes. The first regression join points for all anthocyanins occurred around 1.8 – 2.4 m vine spacing; those for malvidin and derivatives consistently appeared at 1.8 m spacing. Although volatile compounds only seemed to respond to ripeness level and not to vine spacing, wine sensory quality generally decreased with wider vine spacing at both grape ripeness levels. Given the changes in biotic and abiotic growth conditions with wider vine spacing, the results likely displayed adaptation to the combined impact of increasing light exposure, berry temperature, physiological functioning, vegetative and reproductive growth ratios, and plant stress. The study showed the prominent role of plant spacing in sustainable grape growing and wine outcomes and thus the importance of judicious decision-making during the establishment and management of vines under any growth conditions and in any terroir.

Vine spacing of <i>Vitis vinifera</i> cv. Shiraz/101-14 Mgt II. Establishment, and vegetative and reproductive growth

Knowledge of vine growth reaction to plant spacing under relatively high potential soil conditions is limited. This study aimed to determine the effects of vine spacing (with fixed row spacing) of Shiraz (clone SH 9C)/101-14 Mgt on a relatively high potential soil in the Breede River Valley, Robertson, South Africa on establishment and training labour, aboveground growth, and yield. The vineyard was planted in 2008 to a Vertical Shoot Positioning trellis, a fixed row spacing of 2.2 m and row orientation of approximately NNE-SSW (30 °). The 15 in-row vine spacing treatments changed from 0.3 - 4.5 m with increments of 30 cm (from 15151 – 1010 vines/ha). Results were generated over six seasons after cordon development was completed. Grapes were harvested at two ripeness levels. Clear groupings were evident for establishment and further training management. Total costs showed trends with an optimal of around 1.8 m vine spacing, but were markedly higher for closer spacing. Canopies developed uniformly with cordon extension. General vegetative growth (over treatments) varied according to seasonal conditions. Individual leaf size was higher for more closely spaced vines. Total vegetative growth parameters mostly showed increasing and decreasing trends on a /vine and /ha basis, respectively, with an increase in spacing. Yields showed a general increase with wider spacing on a /vine and /ha basis. Yield:cane mass ratios displayed an increasing trend from narrow to wide vine spacing, as did fertility and bunch mass. The bunches from the narrowly spaced treatments seemed more compact. Growth and yield parameters showed very clear trends reaching an optimal at around 1.8 m vine spacing. The results of this study can be applied in vine spacing selection and sustainability strategies for different terroirs. Comprehensive guidelines for extrapolation to different growth conditions and terroirs are provided.

Improving 3D reconstruction for accurate measurement of appearance characteristics in shiny fruits using post-harvest particle film: A case study on tomatoes

Besides the importance of internal quality, the textural quality properties of the product have a remarkable effect on consumer satisfaction and market price. Mainly, the three-dimensional (3D) modeling technique is applied to evaluate the quality of agricultural products. This study aimed to employ the structure from motion (SFM) method along with the particle film technique to predict the physical features of tomatoes. Reconstructing 3D models of tomatoes is challenging due to their shiny skin. In this study, a new method based on particle films was developed to address this issue. Fifty samples of tomatoes were randomly selected with various sizes and four ripeness levels with no-external damage. The SFM technique applied with kaolin particle film was a suitable approach for assessing the surface area, volume, and color index, with a correlation coefficient of 0.76, 0.88, and 0.67, respectively. Also, 3D model predictions achieved these textural properties by obtaining RMSEs of 2.44, 2.30, and 0.35, respectively. Finally, the proposed SFM method’s results showed an error of less than 5%. This technique has proven to be a cable technique for the 3D reconstruction of products with a glossy surface.

DEEP LEARNING FOR AUTOMATIC CLASSIFICATION OF AVOCADO FRUIT MATURITY

Avocado (Persea Americana), a fleshy fruit with a single seed, has increased in popularity globally, especially in tropical and Mediterranean climates, thanks to its commercial and nutritional value. Rich in bioactive compounds, avocados contribute to the prevention and treatment of various diseases, including cardiovascular problems and cancer. Avocado production in Indonesia, for example, is showing a significant increase, reflecting the growing demand. Avocado ripeness affects shelf life and quality, making the determination of ripeness level a critical aspect of postharvest management. Skin color and pulp firmness change during storage, affecting quality and nutritional value. Proper classification of ripeness is important to reduce post-harvest losses, improve quality and optimize export costs. Recent research shows the use of technologies such as machine learning and YOLO (You Only Look Once) version 9 in real-time detection of avocado ripeness, offering innovative solutions to reduce post-harvest losses and improve distribution efficiency. This approach not only benefits farmers and consumers but also ensures consumer satisfaction and reduces economic losses. This study highlights the importance of real-time detection in monitoring avocado ripeness, where the training process was conducted for 89,280 iterations resulting in a new model for avocado ripeness detection. The final model has a mean Average Precision (mAP) validation value of 84.3%, mAP 84.3% signifies the optimal level of accuracy in object recognition in avocado fruit maturity images using the YOLO model that has undergone an intensive training process.

Analysis of the effect of oil palm partnership pattern on the selling price and quality of Fresh Fruit Bunches (FFB) at the level of independent farmers in Sumber Rejeki village Sungai Lilin District Musi Banyuasin Regency

The objectives of this research are: (1) To examine the partnership patterns among independent oil palm farmers in Sumber Rejeki Village. (2) To identify factors that influence the price and quality of FFB within the partnership patterns of independent oil palm farmers in Sumber Rejeki Village. (3) To assess the impact of partnership patterns on the price and quality of FFB among independent oil palm farmers in Sumber Rejeki Village. (4) To provide recommendations for a relevant partnership management model to be implemented in the partnership patterns of oil palm farmers in Sumber Rejeki Village. This research was conducted in Sumber Rejeki Village. The research methods used are survey method and purposive sampling method. The number of samples taken is 31 independent oil palm farmers from Sumber Rejeki Village. The sample was determined using the Slovin’s method. The research findings indicate that: The partnership model existing in Sumber Rejeki Village is the inti-plasma partnership. The factor that most influences the price of FFB is the quality of the seeds. The factor that most affects the quality of the FFB is the level of ripeness. The analysis has shown that there is a significant impact of the village cooperative unit, core companies, and independent farmers on the price and quality of FFB in Sumber Rejeki Village. The recommended partnership model to be implemented, based on the first priority alternative results, is the core-plasma partnership with a focus on the criteria of capital support.

Analysis of Cross Validation on Classification of Mangosteen Maturity Stages using Support Vector Machine

This study explores the efficacy of the Support Vector Machine (SVM) method in classifying mangosteen fruit images based on six ripeness levels. Employing SVM enables nonlinear data classification and simultaneous utilization of multiple feature extractions, resulting in enhanced accuracy. Analysis reveals that models integrating three feature extractions outperform those with only two. With ample training data and optimized parameters, SVM achieves detection accuracy exceeding 90%. However, algorithmic enhancements are necessary to compute RGB color index values for all pixels on mangosteen skin surfaces, possibly through circular-shaped windows approximating the fruit's contour. Moreover, comparative assessments of RGB color system calculations against alternative systems such as HSI are crucial for selecting the most suitable color model in alignment with human perception.

IOT Based Monitoring Model to Identify and Classify the Grading of Fruits and Vegetables

Thirty to forty percent of the fruits and vegetables harvested are discarded at this time due to the large number of employees with no training. A growing problem throughout reaping is fruit and vegetable waste because human judgment is subjective when it comes to crop recognition, categorization, and assessment. The fruit and vegetable business is in serious require of the launch and deployment of a robotic system in order to classify and grade fruits and vegetables based on their level of ripeness. Machine learning techniques can produce a sophisticated machine learning framework that can differentiate among fruits and vegetables based upon their category, variation, development, and preservation using relevant and suitable computational imaging ideas. We analyzed, contrasted, and proposed an Internet of Things-based monitoring approach in this work for identifying, classifying, and grading fruits and vegetables.

AnarMitra: Enhancing Pomegranate Farming with CNN Technology

Abstract: This study introduces AnarMitra, a mobile application designed to revolutionize pomegranate cultivation practices. AnarMitra employs machine learning algorithms to swiftly assess ripeness levels and detect prevalent diseases in pomegranates, such as bacterial blight and fungal infections. Utilizing smartphone cameras, farmers can effortlessly capture images for immediate feedback on fruit quality and health. The paper details AnarMitra's development process, including model training with diverse datasets and algorithmic methodologies for classification and disease detection. Field trials validate AnarMitra's efficacy in enhancing productivity and fostering sustainable agricultural practices, empowering farmers with timely insights for informed decision-making and crop management.

Detection of Gannan Navel Orange Ripeness in Natural Environment Based on YOLOv5-NMM

In order to achieve fast and accurate detection of Gannan navel orange fruits with different ripeness levels in a natural environment under all-weather scenarios and then to realise automated harvesting of Gannan navel oranges, this paper proposes a YOLOv5-NMM (YOLOv5 with Navel orange Measure Model) object detection model based on the improvement in the original YOLOv5 model. Based on the changes in the phenotypic characteristics of navel oranges and the Chinese national standard GB/T 21488-2008, the maturity of Gannan navel oranges is tested. And it addresses and improves the problems of occlusion, dense distribution, small target size, rainy days, and light changes in the detection of navel orange fruits. Firstly, a new detection head of 160 × 160 feature maps is constructed in the detection layer to improve the multi-scale target detection layer of YOLOv5 and to increase the detection accuracy of the different maturity levels of Gannan navel oranges of small sizes. Secondly, a convolutional block attention module is incorporated in its backbone layer to capture the correlations between features in different dimensions to improve the perceptual ability of the model. Then, the weighted bidirectional feature pyramid network structure is integrated into the Neck layer to improve the fusion efficiency of the network on the feature maps and reduce the amount of computation. Lastly, in order to reduce the loss of the target of the Gannan Navel Orange due to occlusion and overlapping, the detection frame is used to remove redundancy using the Soft-NMS algorithm to remove redundant candidate frames. The results show that the accuracy rate, recall rate, and average accuracy of the improved YOLOv5-NMM model are 93.2%, 89.6%, and 94.2%, respectively, and the number of parameters is only 7.2 M. Compared with the mainstream network models, such as Faster R-CNN, YOLOv3, the original model of YOLOv5, and YOLOv7-tiny, it is superior in terms of the accuracy rate, recall rate, and average accuracy mean, and also performs well in terms of the detection rate and memory occupation. This study shows that the YOLOv5-NMM model can effectively identify and detect the ripeness of Gannan navel oranges in natural environments, which provides an effective exploration of the automated harvesting of Gannan navel orange fruits.

Understanding cheese ripeness: An artificial intelligence-based approach for hierarchical classification

Within the contemporary dairy industry, the effective monitoring of cheese ripeness constitutes a critical yet challenging task. This paper proposes the first public dataset encompassing images of cheese wheels that depict various products at distinct stages of ripening and introduces an innovative hybrid approach, integrating machine learning and computer vision techniques to automate the detection of cheese ripeness. By leveraging deep learning and shallow learning techniques, the proposed method endeavors to overcome the limitations associated with conventional assessment methodologies. It aims to provide automation, precision, and consistency in the evaluation of cheese ripeness, delving into a hierarchical classification for the simultaneous classification of distinct cheese types and ripeness levels and presenting a comprehensive solution to enhance the efficiency of the cheese production process. By employing a lightweight hierarchical feature aggregation methodology, this investigation navigates the intricate landscape of preprocessing steps, feature selection, and diverse classifiers. We report a noteworthy achievement, attaining a best F-measure score of 0.991 through the merging of features extracted from EfficientNet and DarkNet-53, opening the field to concretely address the complexity inherent in cheese quality assessment.

Multiple metabolite profiles uncover remarkable bioactive compounds and metabolic characteristics of noni fruit (Morinda citrifolia L.) at various stages of ripeness

This study aimed to investigate the changes in physicochemical properties, bioactive compounds, and metabolic characteristics of noni fruit at different ripeness levels. The results showed that there were significant differences in physicochemical properties. HPLC analysis was conducted, revealing succinic acid, scopoletin, deacetylasperulosidic acid, and asperulosidic acid were key bioactive compounds as the fruit ripened. Additionally, 4 differentbiomarkers (isocitric acid, 4,4-thiodiphenol, lobaric acid, and octocrylene), identified using 1HNMR and LC-IT-TOF-MS, were found to have a VIP value over 1. The results from HS-GC-IMS demonstrated noteworthy that 14 volatile compounds were identified as highly discriminative features during fruit ripening. Furthermore, correlation analysis showed that different ripeness had significant effects on bioactive components and functional activities, e.g., the inhibition rate of enzyme and E. coli of noni fruit with different ripeness exceeded 90% at the last stage. This study contributes new insights into the effective utilization of bioactive ingredients in noni fruit.

A Lightweight Deep Learning Semantic Segmentation Model for Optical-Image-Based Post-Harvest Fruit Ripeness Analysis of Sugar Apples (Annona squamosa)

The sugar apple (Annona squamosa) is valued for its taste, nutritional richness, and versatility, making it suitable for fresh consumption and medicinal use with significant commercial potential. Widely found in the tropical Americas and Asia’s tropical or subtropical regions, it faces challenges in post-harvest ripeness assessment, predominantly reliant on manual inspection, leading to inefficiency and high labor costs. This paper explores the application of computer vision techniques in detecting ripeness levels of harvested sugar apples and proposes an improved deep learning model (ECD-DeepLabv3+) specifically designed for ripeness detection tasks. Firstly, the proposed model adopts a lightweight backbone (MobileNetV2), reducing complexity while maintaining performance through MobileNetV2′s unique design. Secondly, it incorporates the efficient channel attention (ECA) module to enhance focus on the input image and capture crucial feature information. Additionally, a Dense ASPP module is introduced, which enhances the model’s perceptual ability and expands the receptive field by stacking feature maps processed with different dilation rates. Lastly, the proposed model emphasizes the spatial information of sugar apples at different ripeness levels by the coordinate attention (CA) module. Model performance is validated using a self-made dataset of harvested optical images categorized into three ripeness levels. The proposed model (ECD-DeepLabv3+) achieves values of 89.95% for MIoU, 94.58% for MPA, 96.60% for PA, and 94.61% for MF1, respectively. Compared to the original DeepLabv3+, it greatly reduces the number of model parameters (Params) and floating-point operations (Flops) by 89.20% and 69.09%, respectively. Moreover, the proposed method could be directly applied to optical images obtained from the surface of the sugar apple, which provides a potential solution for the detection of post-harvest fruit ripeness.