A hybrid computational framework that fuses deep learning with handcrafted image analysis to achieve accurate, interpretable, and real-time mango grading aligned with established AGMARK quality standards is presented.
Abstract
Mango grading in post-harvest supply chains has traditionally relied on manual inspection, a process prone to subjectivity, inconsistency, and economic inefficiency. Automated quality assessment offers a compelling alternative, yet existing systems often lack the multi-dimensional feature integration necessary to replicate the nuanced judgements of experienced graders. This study presents a hybrid computational framework that fuses deep learning with handcrafted image analysis to achieve accurate, interpretable, and real-time mango grading aligned with established AGMARK quality standards. The framework simultaneously evaluates shape regularity, surface texture, and chromatic properties extracted from multiple colour spaces, combining a Hierarchical Shape Morphology Convolutional Neural Network model with a Grey-Level Co-occurrence Matrix Support Vector Machine classifier through a calibrated weighted fusion mechanism. Trained and evaluated on 1,800 mango images spanning four quality grades, the system achieved an overall classification accuracy of 96.8%, a macro F1-score of 0.967, and an area under the receiver operating characteristic curve of 0.991, surpassing standalone deep learning and traditional machine learning baselines by meaningful margins. Each image was processed in 42 ms on an embedded GPU platform, demonstrating suitability for deployment on industrial conveyor systems without sacrificing throughput. These results indicate that integrating morphological and textural handcrafted features with learned deep representations produces a more robust and generalisable grading system than either approach alone. Beyond mango assessment, the proposed fusion methodology offers a transferable blueprint for automated quality inspection across a broader range of horticultural produce, with the potential to reduce post-harvest losses, standardise grading practices, and support more efficient agricultural supply chains globally.
Manual grading of arecanut is slow, subjective and dependent on the availability of experienced graders. The novel contribution of this work is a controlled, like-for-like comparison of three convolutional neural networks (ResNet50, DenseNet-121 and EfficientNetV2) and a hybrid vision transformer (MaxViT) for grading a...
A. Pinto, Adarsh Rai· 2026 Forth International Con...· 0 citations
In recent years, automated fruit classification systems have become a more critical component in precision agriculture for efficient yield estimation, quality control, and supply chain optimization. However, they still face significant challenges under varying quality, fruit maturity, illumination, texture degradation,...
Khush Kataruka, Mohammad Muzammila, Z. Akhtar et al.· 0 citations
Deep learning has strengthened automated image analysis by enabling hierarchical feature extraction and classification without extensive handcrafted feature engineering. This study developed an integrated image processing framework combining deep feature learning, multiclass classification, explainability, and confiden...
H. C. Shashikala, M. G. Anuradha, C. Srushti et al.· Computer and Decision Making...· 0 citations
Detecting and classifying insect pests is a critical challenge in agricultural pest management, as infestations can reduce crop yield and quality. This study introduces JPNet, a convolutional neural network (CNN) architecture that uses multi-layer feature fusion to detect and classify insect pests affecting jute crops....
Mejbah Ahammad, Md. Ashraful Babu, M. Hossain et al.· Mathematical Modelling and N...· 0 citations
Batik, an internationally recognized cultural heritage of Indonesia, requires rigorous quality control to preserve both its artistic integrity and commercial value. Conventional manual inspection of Batik klowong defects is labor-intensive, time-consuming, and prone to human error, so an automated, reliable alternative...
Denny Sukma Eka Atmaja, Muhammad Kusumawan Herliansyah, A. Sudiarso· JOIV: International Journal...· 0 citations
This study evaluates whether multi-scale feature extraction provides an advantage over single-scale CNN feature extraction under limited-data conditions. Using the Indonesian Spices Dataset as a controlled benchmarking testbed, which contains 31 distinct categories with only 210 images per class, we propose a multi-sca...