Aerial Wildlife Image Repository for animal monitoring with drones in the age of artificial intelligence.

Drones (unoccupied aircraft systems) have become effective tools for wildlife monitoring and conservation. Automated animal detection and classification using artificial intelligence (AI) can substantially reduce logistical and financial costs and improve drone surveys. However, the lack of annotated animal imagery for training AI is a critical bottleneck in achieving accurate performance of AI algorithms compared to other fields. To bridge this gap for drone imagery and help advance and standardize automated animal classification, we have created the Aerial Wildlife Image Repository (AWIR), which is a dynamic, interactive database with annotated images captured from drone platforms using visible and thermal cameras. The AWIR provides the first open-access repository for users to upload, annotate, and curate images of animals acquired from drones. The AWIR also provides annotated imagery and benchmark datasets that users can download to train AI algorithms to automatically detect and classify animals, and compare algorithm performance. The AWIR contains 6587 animal objects in 1325 visible and thermal drone images of predominantly large birds and mammals of 13 species in open areas of North America. As contributors increase the taxonomic and geographic diversity of available images, the AWIR will open future avenues for AI research to improve animal surveys using drones for conservation applications. Database URL: https://projectportal.gri.msstate.edu/awir/.

Fine-tuning digital FIR filters with gray wolf optimization for peak performance.

The design of optimum filters constitutes a fundamental aspect within the realm of signal processing applications. The process entails the calculation of ideal coefficients for a filter in order to get a passband with a flat response and an unlimited level of attenuation in the stopband. The objective of this work is to solve the FIR filter design problem and to compare the optimal solutions obtained from evolutionary algorithms. The design of optimal FIR low pass (LP), high pass (HP), and band stop (BS) filters is achieved by the utilization of nature-inspired optimization approaches, namely gray wolf optimization ,cuckoo search, particle swarm optimization, and genetic algorithm. The filters are evaluated in terms of their stop band attenuation, pass band ripples, and departure from the anticipated response. In addition, this study compares the optimization strategies applied in the context of algorithm execution time which is achievement of global optimal outcomes for the design of digital finite impulse response (FIR) filters. The results indicate that when the Gray wolf algorithm is applied to the development of a finite impulse response (FIR) filter, it produces a higher level of performance than other approaches, as supported by enhanced design precision, decreased execution time, and achievement of an optimal solution.

Are unmanned aerial vehicle-based hyperspectral imaging and machine learning advancing crop science?

The past few years have seen increased interest in unmanned aerial vehicle (UAV)-based hyperspectral imaging (HSI) and machine learning (ML) in agricultural research, concomitant with an increase in published research on these topics. We provide an updated review, written for agriculturalists, highlighting the benefits in the retrieval of biophysical parameters of crops via UAVs relative to less sophisticated options. We reviewed >70 recent papers and found few consistent results between similar studies. Owing to their high complexity and cost, especially when applied to crops of low value, the benefits of most of the research reviewed are difficult to explain. Future effort will be necessary to distill research findings into lower-cost options for end-users.

Impact of soil moisture stress during the silk emergence and grain-filling in maize.

Suboptimal soil moisture during the growing season often limits maize growth and yield. However, the growth stage-specific responses of maize to soil moisture regimes have not been thoroughly investigated. This study investigated the response of maize to five different soil moisture regimes, that are, 0.25, 0.20, 0.15, 0.10, and 0.05 m3 m-3 volumetric water content (VWC), during flowering and grain-filling stages. Sub-optimal soil moisture at the flowering and grain-filling stages reduced ear leaf stomatal conductance by 73 and 64%, respectively. An increase in stress severity caused significant reductions in ear leaf chlorophyll content and greenness-associated vegetation indices across growth stages. Fourteen days of soil moisture stress during flowering delayed silk emergence, reduced silk length (19%), and silk fresh weight (34%). Furthermore, sub-optimal soil moisture caused a significant reduction in both kernel number (53%) and weight (54%). Soil moisture stress at the flowering had a direct impact on kernel number and an indirect effect on kernel weight. During grain-filling, disruption of ear leaf physiology resulted in a 34% decrease in kernel weight and a 43% decrease in kernel number. Unlike grain-filling, treatments at the flowering significantly reduced kernel starch (3%) and increased protein by 29%. These findings suggest that developing reproductive stage stress-tolerant hybrids with improved resilience to soil moisture stress could help reduce the yield gap between irrigated and rainfed maize.

Phenological stage and vegetation index for predicting corn yield under rainfed environments.

Uncrewed aerial systems (UASs) provide high temporal and spatial resolution information for crop health monitoring and informed management decisions to improve yields. However, traditional in-season yield prediction methodologies are often inconsistent and inaccurate due to variations in soil types and environmental factors. This study aimed to identify the best phenological stage and vegetation index (VI) for estimating corn yield under rainfed conditions. Multispectral images were collected over three years (2020-2022) during the corn growing season and over fifty VIs were analyzed. In the three-year period, thirty-one VIs exhibited significant correlations (r ≥ 0.7) with yield. Sixteen VIs were significantly correlated with the yield at least for two years, and five VIs had a significant correlation with the yield for all three years. A strong correlation with yield was achieved by combining red, red edge, and near infrared-based indices. Further, combined correlation and random forest an alyses between yield and VIs led to the identification of consistent and highest predictive power VIs for corn yield prediction. Among them, leaf chlorophyll index, Medium Resolution Imaging Spectrometer (MERIS) terrestrial chlorophyll index and modified normalized difference at 705 were the most consistent predictors of corn yield when recorded around the reproductive stage (R1). This study demonstrated the dynamic nature of canopy reflectance and the importance of considering growth stages, and environmental conditions for accurate corn yield prediction.

Fusion of visible and thermal images improves automated detection and classification of animals for drone surveys.

Visible and thermal images acquired from drones (unoccupied aircraft systems) have substantially improved animal monitoring. Combining complementary information from both image types provides a powerful approach for automating detection and classification of multiple animal species to augment drone surveys. We compared eight image fusion methods using thermal and visible drone images combined with two supervised deep learning models, to evaluate the detection and classification of white-tailed deer (Odocoileus virginianus), domestic cow (Bos taurus), and domestic horse (Equus caballus). We classified visible and thermal images separately and compared them with the results of image fusion. Fused images provided minimal improvement for cows and horses compared to visible images alone, likely because the size, shape, and color of these species made them conspicuous against the background. For white-tailed deer, which were typically cryptic against their backgrounds and often in shadows in visible images, the added information from thermal images improved detection and classification in fusion methods from 15 to 85%. Our results suggest that image fusion is ideal for surveying animals inconspicuous from their backgrounds, and our approach uses few image pairs to train compared to typical machine-learning methods. We discuss computational and field considerations to improve drone surveys using our fusion approach.

Effects of drought and heat stresses during reproductive stage on pollen germination, yield, and leaf reflectance properties in maize (Zea mays L.).

Drought and heat stresses are the major abiotic stress factors detrimental to maize (Zea mays L.) production. Much attention has been directed toward plant responses to heat or drought stress. However, maize reproductive stage responses to combined heat and drought remain less explored. Therefore, this study aimed to quantify the impact of optimum daytime (30°C, control) and warmer daytime temperatures (35°C, heat stress) on pollen germination, morpho-physiology, and yield potential using two maize genotypes ("Mo17" and "B73") under contrasting soil moisture content, that is, 100% and 40% irrigation during flowering. Pollen germination of both genotypes decreased under combined stresses (42%), followed by heat stress (30%) and drought stress (19%). Stomatal conductance and transpiration were comparable between control and heat stress but significantly decreased under combined stresses (83% and 72%) and drought stress (52% and 47%) compared with the control. Genotype "Mo17" reduced its green leaf area to minimize the water loss, which appears to be one of the adaptive strategies of "Mo17" under stress conditions. The leaf reflectance of both genotypes varied across treatments. Vegetation indices associated with pigments (chlorophyll index of green, chlorophyll index of red edge, and carotenoid index) and plant health (normalized difference red-edge index) were found to be highly sensitive to drought and combined stressors than heat stress. Combined drought and heat stresses caused a significant reduction in yield and yield components in both Mo17 (49%) and B73 (86%) genotypes. The harvest index of genotype "B73" was extremely low, indicating poor partitioning efficiency. At least when it comes to "B73," the cause of yield reduction appears to be the result of reduced sink number rather than the pollen and source size. To the best of our awareness, this is the first study that showed how the leaf-level spectra, yield, and quality parameters respond to the short duration of independent and combined stresses during flowering in inbred maize. Further studies are required to validate the responses of potential traits involving diverse maize genotypes under field conditions. This study suggests the need to develop maize with improved tolerance to combined stresses to sustain production under increasing temperatures and low rainfall conditions.

Improving Animal Monitoring Using Small Unmanned Aircraft Systems (sUAS) and Deep Learning Networks.

In recent years, small unmanned aircraft systems (sUAS) have been used widely to monitor animals because of their customizability, ease of operating, ability to access difficult to navigate places, and potential to minimize disturbance to animals. Automatic identification and classification of animals through images acquired using a sUAS may solve critical problems such as monitoring large areas with high vehicle traffic for animals to prevent collisions, such as animal-aircraft collisions on airports. In this research we demonstrate automated identification of four animal species using deep learning animal classification models trained on sUAS collected images. We used a sUAS mounted with visible spectrum cameras to capture 1288 images of four different animal species: cattle (Bos taurus), horses (Equus caballus), Canada Geese (Branta canadensis), and white-tailed deer (Odocoileus virginianus). We chose these animals because they were readily accessible and white-tailed deer and Canada Geese are considered aviation hazards, as well as being easily identifiable within aerial imagery. A four-class classification problem involving these species was developed from the acquired data using deep learning neural networks. We studied the performance of two deep neural network models, convolutional neural networks (CNN) and deep residual networks (ResNet). Results indicate that the ResNet model with 18 layers, ResNet 18, may be an effective algorithm at classifying between animals while using a relatively small number of training samples. The best ResNet architecture produced a 99.18% overall accuracy (OA) in animal identification and a Kappa statistic of 0.98. The highest OA and Kappa produced by CNN were 84.55% and 0.79 respectively. These findings suggest that ResNet is effective at distinguishing among the four species tested and shows promise for classifying larger datasets of more diverse animals.