Fruit type classification using deep learning and feature fusion

Gill, Harmandeep Singh; Murugesan, G.; Mehbodniya, Abolfazl; Sajja, Guna Sekhar; Gupta, Gaurav; Bhatt, Abhishek

doi:10.1016/j.compag.2023.107990

Cited by 10 publications

(2 citation statements)

References 30 publications

(19 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In recent years, large mechanical devices have also been used to classify fruits based on size, weight, or appearance characteristics (Shi et al, 2020). However, this approach may not be able to accurately identify quality aspects such as internal ripeness or flesh texture (Gill et al, 2023). Fruits lose water gradually due to their own metabolic activities, and therefore weight loss is also an important indicator for determining fruit quality (Huang, Wang, et al, 2023).…”

Section: Introductionmentioning

confidence: 99%

Non‐destructive grading technique for mangoes using a flexible impedance sensing system and YOLOv5s_CBAM

Huang,

Wang,

Wang

et al. 2024

J Food Process Engineering

View full text Add to dashboard Cite

Flexible sensors for food quality control are experiencing rapid development. The purpose of this study is to address the time‐consuming issues associated with traditional fruit grading methods by utilizing a homemade flexible impedance sensing system (FISS). A customized spiral slide system was innovatively designed in the study to simulate the grading process on a fruit assembly line with multimodal features for low power consumption and non‐destructive evaluation. FISS integrated components such as a camera, spiral slider, 3D printed slider, latex ball, flexible impedance electrodes, and back‐end measurement circuits and successfully achieved accurate food quality assessment by using a visual classification model for primary grading based on YOLOv5s‐CBAM and an impedance feature classification model for secondary grading of mangoes based on the coefficient of variation method and threshold coefficient method. With an impressive assessment accuracy of 97.07% and a classification speed of up to 3 s/piece, the system successfully classified mangoes into seven grades covering overripe, fully ripe, ripe, unevenly ripe, unripe, underripe, and rotten states. By eliminating the reliance on complex instruments and expensive equipment, FISS provides a cost‐effective alternative for food quality control, significantly reducing operational costs.Practical applicationsThe purpose of this study is to address the time‐consuming issues associated with traditional fruit grading methods by utilizing a homemade flexible impedance sensing system (FISS). A customized spiral slide system was innovatively designed in the study to simulate the grading process on a fruit assembly line with multimodal features for low power consumption and non‐destructive evaluation. FISS integrated components such as a camera, spiral slider, 3D printed slider, latex ball, flexible impedance electrodes, and back‐end measurement circuits and successfully achieved accurate food quality assessment by using a visual classification model for primary grading based on YOLOv5s‐CBAM and an impedance feature classification model for secondary grading of mangoes based on the coefficient of variation method and threshold coefficient method. With an impressive assessment accuracy of 97.07% and a classification speed of up to 3 s/piece, the system successfully classified mangoes into seven grades covering overripe, fully ripe, ripe, unevenly ripe, unripe, underripe, and rotten states. By eliminating the reliance on complex instruments and expensive equipment, FISS provides a cost‐effective alternative for food quality control, significantly reducing operational costs.

show abstract

Section: Introductionmentioning

confidence: 99%

Non‐destructive grading technique for mangoes using a flexible impedance sensing system and YOLOv5s_CBAM

Huang,

Wang,

Wang

et al. 2024

J Food Process Engineering

View full text Add to dashboard Cite

show abstract

“…Several methods can be found in the literature to classify fruits according to their qualitative characteristics [6][7][8], spectroscopy [9][10][11][12][13] and hyperspectral imaging [14]. Qualitative characteristics include ripeness, freshness, color, nutritional value and texture.…”

Section: Introductionmentioning

confidence: 99%

Classification of Healthy and Frozen Pomegranates Using Hyperspectral Imaging and Deep Learning

Mousavi,

Pourdarbani,

Sabzi

et al. 2024

Horticulturae

View full text Add to dashboard Cite

Pomegranate is a temperature-sensitive fruit during postharvest storage. If exposed to cold temperatures above its freezing point for a long time, it will suffer from cold stress. Failure to pay attention to the symptoms that may occur during storage will result in significant damage. Identifying pomegranates susceptible to cold damage in a timely manner requires considerable skill, time and cost. Therefore, non-destructive and real-time methods offer great benefits for commercial producers. To this end, the purpose of this study is the non-destructive identification of healthy frozen pomegranates. First, healthy pomegranates were collected, and hyperspectral images were acquired using a hyperspectral camera. Then, to ensure that enough frozen pomegranates were collected for model training, all samples were kept in cold storage at 0 °C for two months. They were then transferred to the laboratory and hyperspectral images were taken from all of them again. The dataset consisted of frozen and healthy images of pomegranates in a ratio of 4:6. The data was divided into three categories, training, validation and test, each containing 1/3 of the data. Since there is a class imbalance in the training data, it was necessary to increase the data of the frozen class by the amount of its difference with the healthy class. Deep learning networks with ResNeXt, RegNetX, RegNetY, EfficientNetV2, VisionTransformer and SwinTransformer architectures were used for data analysis. The results showed that the accuracies of all models were above 99%. In addition, the accuracy values of RegNetX and EfficientNetV2 models are close to one, which means that the number of false positives is very small. In general, due to the higher accuracy of EfficientNetV2 model, as well as its relatively high precision and recall compared to other models, the F1 score of this model is also higher than the others with a value of 0.9995.

show abstract

Deep learning models for digital image processing: a review

Archana,

Jeevaraj

2024

Artif Intell Rev

View full text Add to dashboard Cite

Fruit type classification using deep learning and feature fusion

Cited by 10 publications

References 30 publications

Non‐destructive grading technique for mangoes using a flexible impedance sensing system and YOLOv5s_CBAM

Non‐destructive grading technique for mangoes using a flexible impedance sensing system and YOLOv5s_CBAM

Classification of Healthy and Frozen Pomegranates Using Hyperspectral Imaging and Deep Learning

Deep learning models for digital image processing: a review

Contact Info

Product

Resources

About