ระบบจัดเก็บและค้นหางานวิจัยอัจฉริยะ

AUTOMATIC CORN FIELD DETECTION USING UNMANNED AERIAL VEHICLE AND TRANSFER LEARNING BASED APPROACH

ผศ.ดร.นิกร กรรณิกากลาง | Agricultural monitoring; Computer vision; Corn detection; Deep learning; Transfer learning; UAV imaging

จำนวนผู้เข้าชม

Creative Commons License

AUTOMATIC CORN FIELD DETECTION USING UNMANNED AERIAL VEHICLE AND TRANSFER LEARNING BASED APPROACH is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 Thailand License.

ดาวน์โหลดเอกสาร

ดาวน์โหลด: 0 ครั้ง

DOI Link EndNote OpenURL

Collection	RMUTK Research Repository (RMUTK IR) วารสารวิชาการ — e-Journal Articles
ID RMUTK Digital	RMUTK000114
Title	AUTOMATIC CORN FIELD DETECTION USING UNMANNED AERIAL VEHICLE AND TRANSFER LEARNING BASED APPROACH
Alternative title	AUTOMATIC CORN FIELD DETECTION USING UNMANNED AERIAL VEHICLE AND TRANSFER LEARNING BASED APPROACH
Authors	ผศ.ดร.นิกร กรรณิกากลาง ผู้แต่งหลัก
Faculty	คณะบริหารธุรกิจ
Journal Title	ICIC Express Letters
ISSN	1881803X
Volume / Issue / Pages	Vol.20 \| No.2 \| pp.215 - 223
Published	2026-02-01
Year	2569
Level	ระดับนานาชาติ (SCOPUS)
Quartile	Q4
DOI	10.24507/icicel.20.02.215
Funding Source	Nakhon Phanom University
Abstract	This study presents an automated system for corn field detection and measurement that combines Unmanned Aerial Vehicle (UAV) technology with deep transfer learning to address critical inefficiencies in agricultural data verification. Current verification methods rely on time-consuming on-site inspections that strain resources and delay operations. Our methodology integrates aerial data acquisition with automated analysis by leveraging transfer learning techniques to adapt pre-trained convolutional neural networks for corn field detection. Through a comparative evaluation of three architectures – MobileNetV1, MobileNetV2, and InceptionV3 – we demonstrate how transfer learning enables efficient model training with limited agricultural datasets while maintaining high accuracy. Testing revealed MobileNetV2’s superior performance with a Mean Absolute Error of 0.147, enabling field measurements accurate to within ±50 square meters of actual dimensions. Field validation demonstrates the system’s capability to automatically detect and delineate corn cultivation areas with high precision, offering substantial improvements over traditional manual verification methods. This transfer learning approach significantly reduces both human intervention requirements and operational costs while improving measurement accuracy and minimizing the need for extensive training data. The system’s demonstrated effectiveness suggests broad applications in agricultural monitoring and resource management, particularly in regions where manual verification poses logistical challenges.
Abstract (EN)	This study presents an automated system for corn field detection and measurement that combines Unmanned Aerial Vehicle (UAV) technology with deep transfer learning to address critical inefficiencies in agricultural data verification. Current verification methods rely on time-consuming on-site inspections that strain resources and delay operations. Our methodology integrates aerial data acquisition with automated analysis by leveraging transfer learning techniques to adapt pre-trained convolutional neural networks for corn field detection. Through a comparative evaluation of three architectures – MobileNetV1, MobileNetV2, and InceptionV3 – we demonstrate how transfer learning enables efficient model training with limited agricultural datasets while maintaining high accuracy. Testing revealed MobileNetV2’s superior performance with a Mean Absolute Error of 0.147, enabling field measurements accurate to within ±50 square meters of actual dimensions. Field validation demonstrates the system’s capability to automatically detect and delineate corn cultivation areas with high precision, offering substantial improvements over traditional manual verification methods. This transfer learning approach significantly reduces both human intervention requirements and operational costs while improving measurement accuracy and minimizing the need for extensive training data. The system’s demonstrated effectiveness suggests broad applications in agricultural monitoring and resource management, particularly in regions where manual verification poses logistical challenges.
Keywords (TH)	Agricultural monitoring; Computer vision; Corn detection; Deep learning; Transfer learning; UAV imaging
Keywords (EN)	Agricultural monitoring; Computer vision; Corn detection; Deep learning; Transfer learning; UAV imaging
Access Level	Open Access (เปิดสาธารณะ)

เอกสารดิจิทัล

RMUTK000114_full.pdf

จำนวนดาวน์โหลด: 0 ครั้ง

Open Access

ดาวน์โหลด