Machine learning-based image classification and estimation of hazelnuts with potential implications for marketing efficiency and producer income

Hazelnut classification and quality assessment are critical stages of post-harvest management that directly influence product quality, market value, and commercialization. Conventional grading methods are labor-intensive, time-consuming, and highly dependent on expert evaluation, limiting their efficiency and consistency in large-scale applications. This study aimed to develop an image-based artificial intelligence framework for automated hazelnut variety classification and weight estimation to improve post-harvest management, marketing efficiency, and producer income. A dataset comprising 700 images of five commercially important hazelnut varieties (Çakıldak, Koca Kulak, Palaz, Yağlı, and Yomra) was acquired under controlled imaging conditions. Due to their superior performance, the Artificial Neural Network (ANN) model was selected for hazelnut variety classification, while the Gaussian Process Regression (GPR) model was preferred for weight estimation. The dataset was divided into three subsets: 70% for training, 15% for validation, and 15% for testing. The ANN model achieved 100% classification accuracy, while the GPR model achieved an R 2 value of 0.72. These findings suggest that image-based machine learning approaches may offer reliable and efficient solutions for automated hazelnut characterization. The proposed framework has the potential to reduce labor requirements, minimize subjective grading errors, and support quality-based classification and pricing. It may also facilitate weight estimation for logistics and inventory management and potentially contribute to improved producer profitability through more efficient resource allocation. The proposed system may support the digital transformation of agricultural production and post-harvest operations by providing a potentially practical and scalable approach for precision agriculture. In the longer term, such applications could contribute to value chain optimization, sustainable agricultural economics, improved marketing efficiency, and enhanced competitiveness in domestic and international hazelnut markets. However, these potential benefits require further validation under commercial and large-scale field conditions.

Authors

Institutions

Publication Details

Journal
Scientific Reports
Published
2026-11-13
DOI
https://doi.org/10.1038/s41598-026-74354-1
Primary Topic
Smart Agriculture and AI
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Machine learning-based image classification and estimation of hazelnuts with potential implications for marketing efficiency and producer income

İbrahim Hakkı Kadirhanoğulları, Alper GÜLBE, Emrah KUŞ, Ersin Gülsoy
1 links
Scientific Reports
Smart Agriculture and AI
article

Machine learning-based image classification and estimation of hazelnuts with potential implications for marketing efficiency and producer income

İbrahim Hakkı Kadirhanoğulları, Alper GÜLBE, Emrah KUŞ, Ersin Gülsoy
article en

Abstract

Hazelnut classification and quality assessment are critical stages of post-harvest management that directly influence product quality, market value, and commercialization. Conventional grading methods are labor-intensive, time-consuming, and highly dependent on expert evaluation, limiting their efficiency and consistency in large-scale applications. This study aimed to develop an image-based artificial intelligence framework for automated hazelnut variety classification and weight estimation to improve post-harvest management, marketing efficiency, and producer income. A dataset comprising 700 images of five commercially important hazelnut varieties (Çakıldak, Koca Kulak, Palaz, Yağlı, and Yomra) was acquired under controlled imaging conditions. Due to their superior performance, the Artificial Neural Network (ANN) model was selected for hazelnut variety classification, while the Gaussian Process Regression (GPR) model was preferred for weight estimation. The dataset was divided into three subsets: 70% for training, 15% for validation, and 15% for testing. The ANN model achieved 100% classification accuracy, while the GPR model achieved an R 2 value of 0.72. These findings suggest that image-based machine learning approaches may offer reliable and efficient solutions for automated hazelnut characterization. The proposed framework has the potential to reduce labor requirements, minimize subjective grading errors, and support quality-based classification and pricing. It may also facilitate weight estimation for logistics and inventory management and potentially contribute to improved producer profitability through more efficient resource allocation. The proposed system may support the digital transformation of agricultural production and post-harvest operations by providing a potentially practical and scalable approach for precision agriculture. In the longer term, such applications could contribute to value chain optimization, sustainable agricultural economics, improved marketing efficiency, and enhanced competitiveness in domestic and international hazelnut markets. However, these potential benefits require further validation under commercial and large-scale field conditions.

Scientific Reports
Iğdır Üniversitesi (TR)
Zero hunger
Smart Agriculture and AI
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon