Abstract:In crop production plant diseases cause significant yield losses. Therefore, the detection and scoring of disease occurrence is of high importance. The quantification of plant diseases requires the identification of leaves as individual scoring units. Diseased leaves are very dynamic and complex biological object which constantly change in form and color after interaction with plant pathogens. To address the task of identifying and segmenting individual leaves in agricultural fields, this work uses unmanned aerial vehicle (UAV), multispectral imagery of sugar beet fields and deep instance segmentation networks (Mask R-CNN). Based on standard and copy-paste image augmentation techniques, we tested and compare five strategies for achieving robustness of the network while keeping the number of labeled images within reasonable bounds. Additionally, we quantified the influence of environmental conditions on the network performance. Metrics of performance show that multispectral UAV images recorded under sunny conditions lead to a drop of up to 7% of average precision (AP) in comparison with images under cloudy, diffuse illumination conditions. The lowest performance in leaf detection was found on images with severe disease damage and sunny weather conditions. Subsequently, we used Mask R-CNN models in an image-processing pipeline for the calculation of leaf-based parameters such as leaf area, leaf slope, disease incidence, disease severity, number of clusters, and mean cluster area. To describe epidemiological development, we applied this pipeline in time-series in an experimental trial with five varieties and two fungicide strategies. Disease severity of the model with the highest AP results shows the highest correlation with the same parameter assessed by experts. Time-series development of disease severity and disease incidence demonstrates the advantages of multispectral UAV-imagery for contrasting varieties for resistance, and the limits for disease control measurements. With this work we highlight key components to consider for automatic leaf segmentation of diseased plants using UAV imagery, such as illumination and disease condition. Moreover, we offer a tool for delivering leaf-based parameters relevant to optimize crop production thought automated disease quantification imaging tools.

Advanced UAV - based leaf disease detection: Deep Radial Basis Function Networks with multidimensional mixed attention

Optimal Fuzzy Deep Neural Networks-Based Plant Disease Detection and Classification on UAV-Based Remote Sensed Data

Autonomous Detection of Plant Disease Symptoms Directly from Aerial Imagery

Using UAV-Imagery for Leaf Segmentation in Diseased Plants via Mask-Based Data Augmentation and Extension of Leaf-based Phenotyping Parameters

Deep Learning-Based Leaf Disease Detection in Crop Using Images for Agricultural Application

An Improved Convolutional Neural Network for Plant Disease Detection Using Unmanned Aerial Vehicle Images

Detection of healthy and diseased crops in drone captured images using Deep Learning

Recent Advances in Crop Disease Detection Using UAV and Deep Learning Techniques

LEVERAGING DEEP LEARNING FOR ACCURATE RICE LEAF DISEASE RECOGNITION

A Two-Step Machine Learning Approach for Crop Disease Detection Using GAN and UAV Technology

UAV hyperspectral remote sensor images for mango plant disease and pest identification using MD-FCM and XCS-RBFNN

Application of Drone Surveillance for Advance Agriculture Monitoring by Android Application Using Convolution Neural Network

Tuned weighted feature fusion with hybridized DNN-RNN framework for plant disease detection and classification

Deep Learning-Based Leaf Disease Detection in Crops Using Images for Agricultural Applications

Optimal Models for Plant Disease and Pest Detection Using UAV Image

Optimized deep learning network for plant leaf disease segmentation and multi-classification using leaf images

Deep Learning-Based Trees Disease Recognition and Classification Using Hyperspectral Data

Advancing the Future of Agricultural Sustainability: Utilizing Deep Learning for Plant Disease Detection

A Channel Attention-Driven Hybrid CNN Framework for Paddy Leaf Disease Detection

Next‐Gen Crop Monitoring: MTEG‐RTU Algorithm and UAV Synergy for Precise Disease Diagnosis

Artificial Intelligence-Based Drone System for Multiclass Plant Disease Detection Using an Improved Efficient Convolutional Neural Network