Plant AI in Agriculture: Innovative Approaches to Sunflower Leaf Disease Detection with Federated Learning CNNs

Abstract

This research presents a new use of Convolutional Neural Networks (CNN) integrated with Federated Learning (FL) for detecting and classifying illnesses on sunflower leaves. The centre of this investigation is a systematic analysis of findings obtained from local data., which are transformed into global knowledge with the help of sophisticated federated averaging approaches. The findings show the efficiency of our model., which is outstanding. The macro averages for accuracy., recall and F1-scores also reveal an upward trend of continuous improvement starting kr _1 (82.95) to up between 83.74 as at kr_3 met kp_ 4atkr(translation truncated). Weighted and micro averages also demonstrate a rising tendency., reflecting that the model becomes more accurate in terms of its ability to manage class imbalances. Specifically., the micro averages also improve similarly., starting at 83.05 and ending at 94.88 for kr_5. Still., the weighted ones change from just two decimal points below (from an average of close to 17 times ten minus six). That difference changes drastically because now we are going from having about. The research finds that federated averaging helps transform local data into global data. It demonstrates how the individual client can contribute individually to the learning of a model. Still., it also shows that in federated learning., collaborative and distributed nature has enhanced performance over time. In conclusion., the study demonstrates a convincing case for utilizing FL and CNN in rural situations., particularly for recognizing illnesses afflicting sunflower leaves. The experimental and statistical data reveal the efficiency of this method., which may lead to its broad utilization in plant pathology for better precision agriculture.

Code and data availability

公開本文の所在を確認できませんでした。非公開または購読が必要な可能性があります。

No evidence-backed public reproduction asset is currently recorded.