TY  - JOUR
AU  - Lazić, Olivera
AU  - Cvejić, Sandra
AU  - Dedić, Boško
AU  - Kupusinac, Aleksandar
AU  - Jocić, Siniša
AU  - Miladinović, Dragana
PY  - 2024
UR  - http://fiver.ifvcns.rs/handle/123456789/4807
AB  - Efficient water supply and timely detection of drought stress in crops to increase yields is an important task considering that agriculture is the primary consumer of water globally. This is particularly significant for plants such as sunflowers, which are an important source of quality edible oils, essential for human nutrition. Traditional detection methods are labor-intensive, time-consuming, and rely on advanced sensor technologies. We introduce an innovative approach based on neural networks and transfer learning for drought stress detection using a novel dataset including 209 non-invasive rhizotron images and 385 images of manually cleaned sections of sunflowers, subjected to normal watering or water stress. We used five neural network models: VGG16, VGG19, InceptionV3, DenseNet, and MobileNet, pre-trained on the ImageNet dataset, whose performance was compared to select the most efficient architecture. Accordingly, the most efficient model, MobileNet, was further refined using different data augmentation mechanisms. The introduction of targeted data augmentation and the use of grayscale images proved to be effective, demonstrating improved results, with an F1 score and an accuracy of 0.95. This approach encourages advances in water stress detection, highlighting the value of artificial intelligence in improving crop health monitoring and management for more resilient agricultural practices.
PB  - Basel : MDPI
T2  - Applied Sciences
T1  - Transfer Learning in Multimodal Sunflower Drought Stress Detection
IS  - 14
SP  - 6034
VL  - 14
DO  - 10.3390/app14146034
ER  - 

@article{
author = "Lazić, Olivera and Cvejić, Sandra and Dedić, Boško and Kupusinac, Aleksandar and Jocić, Siniša and Miladinović, Dragana",
year = "2024",
abstract = "Efficient water supply and timely detection of drought stress in crops to increase yields is an important task considering that agriculture is the primary consumer of water globally. This is particularly significant for plants such as sunflowers, which are an important source of quality edible oils, essential for human nutrition. Traditional detection methods are labor-intensive, time-consuming, and rely on advanced sensor technologies. We introduce an innovative approach based on neural networks and transfer learning for drought stress detection using a novel dataset including 209 non-invasive rhizotron images and 385 images of manually cleaned sections of sunflowers, subjected to normal watering or water stress. We used five neural network models: VGG16, VGG19, InceptionV3, DenseNet, and MobileNet, pre-trained on the ImageNet dataset, whose performance was compared to select the most efficient architecture. Accordingly, the most efficient model, MobileNet, was further refined using different data augmentation mechanisms. The introduction of targeted data augmentation and the use of grayscale images proved to be effective, demonstrating improved results, with an F1 score and an accuracy of 0.95. This approach encourages advances in water stress detection, highlighting the value of artificial intelligence in improving crop health monitoring and management for more resilient agricultural practices.",
publisher = "Basel : MDPI",
journal = "Applied Sciences",
title = "Transfer Learning in Multimodal Sunflower Drought Stress Detection",
number = "14",
pages = "6034",
volume = "14",
doi = "10.3390/app14146034"
}

Lazić, O., Cvejić, S., Dedić, B., Kupusinac, A., Jocić, S.,& Miladinović, D.. (2024). Transfer Learning in Multimodal Sunflower Drought Stress Detection. in Applied Sciences
Basel : MDPI., 14(14), 6034.
https://doi.org/10.3390/app14146034

Lazić O, Cvejić S, Dedić B, Kupusinac A, Jocić S, Miladinović D. Transfer Learning in Multimodal Sunflower Drought Stress Detection. in Applied Sciences. 2024;14(14):6034.
doi:10.3390/app14146034 .

Lazić, Olivera, Cvejić, Sandra, Dedić, Boško, Kupusinac, Aleksandar, Jocić, Siniša, Miladinović, Dragana, "Transfer Learning in Multimodal Sunflower Drought Stress Detection" in Applied Sciences, 14, no. 14 (2024):6034,
https://doi.org/10.3390/app14146034 . .

TY  - JOUR
AU  - Marčeta, Una
AU  - Vučinić-Vasić, Milica
AU  - Ninkov, Jordana
AU  - Ilić, Strahinja
AU  - Vujić, Bogdana
PY  - 2023
UR  - http://fiver.ifvcns.rs/handle/123456789/3859
AB  - Four monitoring campaigns of sedimentable dust were provided from
two large non-sanitary landfills (Zrenjanin and Novi Sad) in Serbia during
2021. Particle size analysis by laser diffraction and inductively coupled
plasma-optical emission spectrometry were carried out in order to obtain the
particle size distribution (PSD) and the toxic elements (TEs) concentrations.
The health risk assessment of the landfill employees was performed according
to the United States Environmental Protection Agency methods based on TEs
concentrations. The PSD results demonstrated that the majority of sedimentable
dust samples mass were not concentrated neither within PM2.5 neither
within PM10 fraction. Analysis revealed high concentration of TEs at both
landfills: an extremely high concentrations of Cr and Zn in samples from
Zrenjanin landfill was detected. Health risk potential of elements was as follows
for both landfills: Cr > Co > Pb > Ni > Zn > Cu. According to the results,
maximal hazard index for landfill employees in Zrenjanin (0.197) and Novi
Sad (0.113) showed that non-cancer risk was very low. For both landfill sites,
cancer risk was highest for Cr (2.75×10-5 for Zrenjanin and 2.02×10-7 for Novi
Sad), though still within the defined threshold for tolerable cancer risk.
AB  - Na dve velike nesanitarne deponije u Srbiji (Zrenjanin i Novi Sad) sprovedene su četiri kampanje monitoringa taložne prašine. Analiza raspodele veličine čestica urađena je laserskom difrakcijom, dok je koncentracija toksičnih elemenata (TE) određena pomoću optičke emisione spektrometrije sa induktivno spregnutom plazmom. Procena rizika po zdravlje zaposlenih na deponijama sprovedena je na osnovu koncentracija TE prema metodama Agencije za zaštitu životne sredine Sjedinjenih Američkih Država. Rezultati raspodele veličine čestica su pokazali da najveći deo uzoraka taložne prašine ne pripada ni PM2,5 ni PM10 frakciji. Analiza je pokazala visoku koncentraciju TE na obe deponije: ustanovljena je izuzetno visoka koncentracija Cr i Zn u uzorcima sa zrenjaninske deponije. Potencijal analiziranih toksičnih elemenata u pogledu rizika po zdravlje za obe deponije je sledeći: Cr > Co > Pb > Ni > Zn > Cu. Prema rezultatima, ukupni (maksimalni) indeks opasnosti za zaposlene na deponijama u Zrenjaninu (0,197) i Novom Sadu (0,113) pokazao je da je rizik od kancera veoma nizak. Za obe deponije, rizik od kancera je bio najveći za Cr (2,75×10-5 za Zrenjanin i 2,02×10-7 za Novi Sad), čija je vrednost takođe u okviru definisanog praga tolerancije za rizik od kancera.
PB  - Beograd : Srpsko hemijsko društvo
T2  - Journal of the Serbian Chemical Society
T1  - Health risk assessment of toxic elements in sedimentable dust from landfills
T1  - Procena zdravstvenog rizika od toksičnih elemenata u taložnoj prašini sa deponija
EP  - 791
IS  - 7-8
SP  - 777
VL  - 88
DO  - 10.2298/JSC230413032M
ER  - 

@article{
author = "Marčeta, Una and Vučinić-Vasić, Milica and Ninkov, Jordana and Ilić, Strahinja and Vujić, Bogdana",
year = "2023",
abstract = "Four monitoring campaigns of sedimentable dust were provided from
two large non-sanitary landfills (Zrenjanin and Novi Sad) in Serbia during
2021. Particle size analysis by laser diffraction and inductively coupled
plasma-optical emission spectrometry were carried out in order to obtain the
particle size distribution (PSD) and the toxic elements (TEs) concentrations.
The health risk assessment of the landfill employees was performed according
to the United States Environmental Protection Agency methods based on TEs
concentrations. The PSD results demonstrated that the majority of sedimentable
dust samples mass were not concentrated neither within PM2.5 neither
within PM10 fraction. Analysis revealed high concentration of TEs at both
landfills: an extremely high concentrations of Cr and Zn in samples from
Zrenjanin landfill was detected. Health risk potential of elements was as follows
for both landfills: Cr > Co > Pb > Ni > Zn > Cu. According to the results,
maximal hazard index for landfill employees in Zrenjanin (0.197) and Novi
Sad (0.113) showed that non-cancer risk was very low. For both landfill sites,
cancer risk was highest for Cr (2.75×10-5 for Zrenjanin and 2.02×10-7 for Novi
Sad), though still within the defined threshold for tolerable cancer risk., Na dve velike nesanitarne deponije u Srbiji (Zrenjanin i Novi Sad) sprovedene su četiri kampanje monitoringa taložne prašine. Analiza raspodele veličine čestica urađena je laserskom difrakcijom, dok je koncentracija toksičnih elemenata (TE) određena pomoću optičke emisione spektrometrije sa induktivno spregnutom plazmom. Procena rizika po zdravlje zaposlenih na deponijama sprovedena je na osnovu koncentracija TE prema metodama Agencije za zaštitu životne sredine Sjedinjenih Američkih Država. Rezultati raspodele veličine čestica su pokazali da najveći deo uzoraka taložne prašine ne pripada ni PM2,5 ni PM10 frakciji. Analiza je pokazala visoku koncentraciju TE na obe deponije: ustanovljena je izuzetno visoka koncentracija Cr i Zn u uzorcima sa zrenjaninske deponije. Potencijal analiziranih toksičnih elemenata u pogledu rizika po zdravlje za obe deponije je sledeći: Cr > Co > Pb > Ni > Zn > Cu. Prema rezultatima, ukupni (maksimalni) indeks opasnosti za zaposlene na deponijama u Zrenjaninu (0,197) i Novom Sadu (0,113) pokazao je da je rizik od kancera veoma nizak. Za obe deponije, rizik od kancera je bio najveći za Cr (2,75×10-5 za Zrenjanin i 2,02×10-7 za Novi Sad), čija je vrednost takođe u okviru definisanog praga tolerancije za rizik od kancera.",
publisher = "Beograd : Srpsko hemijsko društvo",
journal = "Journal of the Serbian Chemical Society",
title = "Health risk assessment of toxic elements in sedimentable dust from landfills, Procena zdravstvenog rizika od toksičnih elemenata u taložnoj prašini sa deponija",
pages = "791-777",
number = "7-8",
volume = "88",
doi = "10.2298/JSC230413032M"
}

Marčeta, U., Vučinić-Vasić, M., Ninkov, J., Ilić, S.,& Vujić, B.. (2023). Health risk assessment of toxic elements in sedimentable dust from landfills. in Journal of the Serbian Chemical Society
Beograd : Srpsko hemijsko društvo., 88(7-8), 777-791.
https://doi.org/10.2298/JSC230413032M

Marčeta U, Vučinić-Vasić M, Ninkov J, Ilić S, Vujić B. Health risk assessment of toxic elements in sedimentable dust from landfills. in Journal of the Serbian Chemical Society. 2023;88(7-8):777-791.
doi:10.2298/JSC230413032M .

Marčeta, Una, Vučinić-Vasić, Milica, Ninkov, Jordana, Ilić, Strahinja, Vujić, Bogdana, "Health risk assessment of toxic elements in sedimentable dust from landfills" in Journal of the Serbian Chemical Society, 88, no. 7-8 (2023):777-791,
https://doi.org/10.2298/JSC230413032M . .