Characterization of Hazelnut Trees in Open Field Through High-Resolution UAV-Based Imagery and Vegetation Indices

Sensors, Journal Year: 2025, Volume and Issue: 25(1), P. 288 - 288

Published: Jan. 6, 2025

The increasing demand for hazelnut kernels is favoring an upsurge in cultivation worldwide, but ongoing climate change threatens this crop, affecting yield decreases and subject to uncontrolled pathogen parasite attacks. Technical advances precision agriculture are expected support farmers more efficiently control the physio-pathological status of crops. Here, we report a straightforward approach monitoring trees open field, using aerial multispectral pictures taken by drones. A dataset 4112 images, each having 2Mpixel resolution per tree covering RGB, Red Edge, near-infrared frequencies, was obtained from 185 located two different orchards Piedmont region (northern Italy). To increase accuracy, especially reduce false negatives, image divided into nine quadrants. For quadrant, vegetation indices (VIs) were computed, parallel, quadrant tagged as “healthy/unhealthy” visual inspection. Three supervised binary classification algorithms used build models capable predicting VIs predictors. Out considered, only five (GNDVI, GCI, NDREI, NRI, GI) good predictors, while NDVI SAVI, RECI, TCARI not. Using them, model accuracy about 65%, with 13% negatives reached way that rather independent algorithms, demonstrating some allow inferring condition these trees. These achievements use drone-captured images performing rapid, non-destructive physiological characterization This offers sustainable strategy supporting their decision-making process during agricultural practices.

Language: Английский

Object-Based Detection of Hazelnut Orchards Using Very High Resolution Aerial Photographs

Published: July 15, 2024

Language: Английский

Hazelnut Classification and Weight Estimation

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: Dec. 16, 2024

Hazelnut is a nutritious and delicious nut that consumed fondly. Hazelnuts also stand out with their economic value. Traditional methods have long been used to classify hazelnuts. However, these are time consuming, costly require expertise. Today, using image-based deep learning classifiers automatically hazelnut varieties can provide significant advantages in agricultural activities. It increase the productivity of products. The main objective this study make weight estimation artificial neural networks (ANN). In study, data set was created by collecting total 700 photo samples from five different varieties, namely Çakıldak, Koca kulak, Palaz, Yağlı Yomra. Then, new novel network model developed. achieved 100% success classification 72% prediction. These results demonstrate potential intelligence applications agriculture. future studies, it recommended test algorithms for

Language: Английский