Maize is a vital global crop, essential for food security but highly susceptible to diseases that threaten yield and quality. Traditional methods for detecting these diseases are computationally intensive and rely on ...
详细信息
Agile autonomous drones are becoming increasingly popular in research due to the challenges they represent in fields like control, state estimation, or perception at high speeds. When all algorithms are computed onboa...
详细信息
ISBN:
(数字)9798350357882
ISBN:
(纸本)9798350357899
Agile autonomous drones are becoming increasingly popular in research due to the challenges they represent in fields like control, state estimation, or perception at high speeds. When all algorithms are computed onboard the UAV, computational limitations make the task of agile flight even more difficult. One of the most computationally expensive tasks in agile flight is the generation of optimal trajectories. When these trajectories must be updated online due to changes in the environment or uncertainties, this high computational cost may result in insufficient time to reach the desired waypoints, which could cause a drone crash in cluttered environments. In this paper, we present Local Gaussian Modifiers (LGMs), a fast and lightweight way of modifying computationally heavy trajectories when recalculating them in time is not possible due to computational limitations. Moreover, we propose a strategy for deciding when is convenient to use these modifiers or recalculate the whole trajectory based on an estimation of the computational time of this trajectory generation. A trajectory blending procedure is also proposed to ensure smoothness in UAV control when a new trajectory is computed. Our approach was validated in simulation, being able to pass through a race circuit with moving gates, achieving speeds up to 16.0 m/s. Real flight validation was also performed achieving speeds up to 4.0 m/s in a fully autonomous pipeline using onboard computing.
Maize is a vital global crop, essential for food security but highly susceptible to diseases that threaten yield and quality. Traditional methods for detecting these diseases are computationally intensive and rely on ...
详细信息
ISBN:
(数字)9798350379037
ISBN:
(纸本)9798350379044
Maize is a vital global crop, essential for food security but highly susceptible to diseases that threaten yield and quality. Traditional methods for detecting these diseases are computationally intensive and rely on high-quality images, limiting their practical use in diverse field conditions. This research addresses these challenges by proposing MF-Net, a novel model that leverages two EfficientNet-b0 architectures with ReLU and Swish activation functions. Our approach involves partially training Model R with frozen early layers using ReLU, and fully training Model S with Swish activation for enhanced optimization and robustness. We employ model concatenation and feature fusion techniques to create a lightweight yet powerful model, complemented by additional layers for improved feature extraction and regularization. Experimental results on the Corn Disease and Severity (CD&S) dataset are compelling: MF-Net achieves a remarkable 95.7% accuracy and 96% precision in disease classification, and 88.3% accuracy and 88.46% precision in severity level detection. These results highlight the model's effectiveness under challenging conditions such as cluttered backgrounds and variable lighting, significantly reducing computation time without compromising accuracy. Hence, MF-Net presents an innovative and efficient solution for maize disease detection and severity assessment, offering significant advancements for agricultural practices and crop management.
暂无评论