Apple Diseases Detection: Based On Deep Learning

Abstract

Apple’s quality and output often go down because of diseases, which affects the growth of the agricultural economy. Proper diagnosis and decision-making of apple diseases are crucial for reducing agricultural losses and boosting economic development. Using a unique Multi-scale Dense classification network, this article diagnoses 11 photos, including healthy and damaged apple fruits and leaves. Several illnesses were diagnosed. First, using typical image augmentation approaches, the Cycle-GAN algorithm was used to expand anthracnose and ring rot datasets. Cycle-GAN can simulate blemishes like anthracnose and ring rot on otherwise perfect apples using what it’s learned about healthy and sick apple images. When Cycle-GAN-generated photos of infected apples were introduced to the training set, the diagnostic accuracy increased compared to baseline approaches the user’s text was too short to be rewritten academically. Two models were created using Dense Net and Multi-scale connection techniques to facilitate the reprocessing of image characteristics from lower layers in organization neural networks. The research paper discusses the use of Multi-scale Dense Inception-V3 and Multi-scale Dense Inception-Resnet-V2 models. Both algorithms were able to identify 11 distinct picture kinds correctly. The overall accuracy and classification accuracy were higher than the DenseNet-121 network, representing a state-of-the-art performance.

Abstract from the paper’s OpenAlex record, shown for indexing and research. Copyright remains with the authors or the publisher. Read at the source ↗

The text the model read, with the source record’s inline markup (italics, sub- and superscripts) rendered and its character entities decoded; quote offsets refer to the raw text.