Optimization techniques in deep convolutional neuronal networks applied to olive diseases classification

Plants diseases have a detrimental effect on the quality but also on the quantity of agricultural production.However,the prediction of these diseases is proving the effect on crop quality and on reducing the risk of production losses.Indeed,the detection of plant diseases-either with a naked eye or using traditional methods-is largely a cumbersome process in terms of time,availability and results with a high-risk error.The present work introduces a depth study of various CNN architectures with different optimization algorithms carried out for olive disease detection using classification techniques that recommend the best model for constructing an effective disease detector.This study presents a dataset of 5571 olive leaf images collected manually on real conditions from different regions of Morocco,that also includes healthy class to detect olive diseases.Further,one of the goals of this research was to study the correlation effects between CNN architectures and optimization algorithms evaluated by the accuracy and other performance metrics.The highest rate in trained models was 100%,while the highest rate in experiments without data augmentation was 92,59%.Another subject of this study is the influence of the optimization algorithms on neuronal network performance.As a result of the experiments carried out,the MobileNet architecture using Rmsprop algorithms outperformed the others combinations in terms of performance and efficiency of disease detector.

DetEEktor:Mask R-CNN based neural network for energy plant identification on aerial photographs

Decentralized renewable energies(RE)represent one of the most important elements for future energy systems.For an optimal integration,a comprehensive knowledge of the installed RE systems and their characteristics is required.In Germany,the’Marktstammdatenregister’(MaStR)gives access to most of this information but has significant deficits in terms of up to date data covering all installed systems,later illustrated using a ground truth data set,and completely neglects some plant types e.g.solar thermal systems.To address this,we developed the novel model’DetEEktor’,with which six different RE plant types can be simultaneously detected and characterized on aerial photographs by means of a Mask R convolutional neural network.As first contribution of this paper,we give a detailed overview of the structure,design and data base of the model.Afterwards,we contribute by demonstrating its capability to attack the lack of knowledge of existing RE systems:Dependent on the plant type,we succeed in detecting 63 to 75%of the existing systems.Applied on the ground truth use case,we reduce the number of missing photovoltaic systems in the MaStR by 44%and identify 72%of the existing solarthermal systems.Applied to a high RE share use case,we identified an estimated number of about 1,468 photovoltaic(including 18 free-field),1,063 solarthermal,6 biomass and 4 wind power plants missing in the MaStR.In conclusion,with our presented’DetEEktor’,we aim to contribute to a more detailed data basis of existing RE systems.