Search

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.

Drought conditions can be significant constraint is sunflower cropping. Determination of variability in sunflower drought response is necessary step in breeding effort to select tolerant genotypes. Experiments were conducted to compare two approaches aiming to isolate drought stress from other sources of stress, control water availability to plants and to determine impact of drought on plant growth. Two inbred lines, HA-26-PR and IMI-AB-14, were selected based on their response to drought. In the first experiment plants were grown for period of 14 days in rhizotrons and in the second for period of 60 days in pots. Both experiments were placed in growth chamber. Drought conditions in rhizotron experiment were on constant level compared to pot experiment where plants were gradually exposed to drought at one stage of development. Plants were phenotyped during and at the end of rhizotron experiment and only at the end of pot experiment. Growing plant in rhizotrons gave insight in root morphology and growth, but for the limited duration of time and only for early stages of sunflower growth. Pot experiment enables one to monitor plant growth and development over longer time and to impose drought at selected moment making it closer to the drought scenarios in field conditions. Both methods are reliable for drought stress induction and determination of drought impact and can be considered as complementary. Further research will aim the comparison of methods in controlled conditions with field drought experiments.

Among the effects of wide-ranging climate change, drought presents a significant threat to global agricultural production. Drought reduces yield quantity and quality of crops, particularly in the semi-arid and arid regions. Sunflower is known for its moderate tolerance to water stress conditions thanks to its well-developed root system and ability to grow in different agroecological conditions. However, prolonged exposure to drought significantly reduces sunflower seed and oil yields, impacting global oil quality. Drought stress in sunflower mostly occurs during the seedling stage, vegetative phase or reproductive phase. Understanding the mechanisms and factors influencing drought tolerance is essential for developing resilient sunflower capable of sustaining agricultural productivity in the face of climate variability and water scarcity. Therefore, breeding for drought tolerance is an economical, feasible, and environmentally friendly control method However, it requires the use of the unexploited approach in sunflower research due to the complexity of the trait. Initially, determination of the genetic diversity existing within and between sunflower species remains the basis for elucidating of the genetic structure and for improvement of traits, including drought tolerance. Manipulation of this diversity to improve drought tolerance among sunflower genotypes may be achieved through selection of adaptive mechanisms that include drought escape, avoidance and tolerance. The breeding tools and technologies for drought tolerance are based on so-called “-omics” techniques, including phenomics, genomics, transcriptomics, proteomics, metabolomics, and epigenomics. Phenomics refers to a quantitative description of the plant's morphological, anatomical, ontogenetic, physiological, and biochemical properties. Recent advancements in phenomics, facilitated by highthroughput phenotyping tools, empower researchers to explore multivariate phenotypic information. Utilizing mainly no