TY - JOUR

T1 - Методы генерации синтетических данных для обучения нейросетей в задаче сегментации уровня азотного режима растений на снимках БПЛА с/х поля

AU - Молин, Александр Евгеньевич

AU - Блеканов, Иван Станиславович

AU - Митрофанов, Евгений Павлович

AU - Митрофанова, Ольга Александровна

PY - 2024

Y1 - 2024

N2 - This study is devoted to the automatization of the image masks’ construction of large-sized agricultural objects in precision farming tasks for training neural network methods for crop’s nitrogen status analysis using georeferenced images. The scientific direction is extremely relevant because it allows to automate and replace the manual process of data labeling, significantly reducing the cost of preparing training samples. In the paper, four new synthetic data generation methods are proposed for training neural networks aimed at UAV image segmentation by the level of crop nitrogen supply on an agricultural field. In particular, the paper gives a description of synthetic data generation algorithms based on nitrogen covering with lines, parabolas, and areas. Experiments were carried out to test and evaluate the quality of these algorithms using eight modern image segmentation methods: two classical machine learning methods (Random Forest and XGBoost), four convolutional neural network methods based on U-Net architecture, and two transformers (TransUnet and UnetR). The results showed that two algorithms based on areas gave the best accuracy for convolutional neural networks and transformers — 98–100 %. Classical machine learning methods showed very low values for all quality metrics — 27–44 %.

AB - This study is devoted to the automatization of the image masks’ construction of large-sized agricultural objects in precision farming tasks for training neural network methods for crop’s nitrogen status analysis using georeferenced images. The scientific direction is extremely relevant because it allows to automate and replace the manual process of data labeling, significantly reducing the cost of preparing training samples. In the paper, four new synthetic data generation methods are proposed for training neural networks aimed at UAV image segmentation by the level of crop nitrogen supply on an agricultural field. In particular, the paper gives a description of synthetic data generation algorithms based on nitrogen covering with lines, parabolas, and areas. Experiments were carried out to test and evaluate the quality of these algorithms using eight modern image segmentation methods: two classical machine learning methods (Random Forest and XGBoost), four convolutional neural network methods based on U-Net architecture, and two transformers (TransUnet and UnetR). The results showed that two algorithms based on areas gave the best accuracy for convolutional neural networks and transformers — 98–100 %. Classical machine learning methods showed very low values for all quality metrics — 27–44 %.

KW - UAV images

KW - deep learning

KW - machine learning

KW - nitrogen level segmentation

KW - remote sensing data labeling

KW - smart agriculture

KW - synthetic data generation

UR - https://applmathjournal.spbu.ru/article/view/17515

UR - https://www.mendeley.com/catalogue/4211dc53-7fb1-3996-9b6a-aa2cc494a9ac/

U2 - 10.21638/11701/spbu10.2024.103

DO - 10.21638/11701/spbu10.2024.103

M3 - статья

VL - 20

SP - 20

EP - 33

JO - ВЕСТНИК САНКТ-ПЕТЕРБУРГСКОГО УНИВЕРСИТЕТА. ПРИКЛАДНАЯ МАТЕМАТИКА. ИНФОРМАТИКА. ПРОЦЕССЫ УПРАВЛЕНИЯ

JF - ВЕСТНИК САНКТ-ПЕТЕРБУРГСКОГО УНИВЕРСИТЕТА. ПРИКЛАДНАЯ МАТЕМАТИКА. ИНФОРМАТИКА. ПРОЦЕССЫ УПРАВЛЕНИЯ

SN - 1811-9905

IS - 1

ER -