Apple Leaf Disease Recognition and Sub-Class Categorization Based on Improved Multi-Scale Feature Fusion Network

Abstract

Apple diseases cause a lot of economic losses to fruit growers in China. Early diagnosis and accurate recognition of apple diseases can control the spread of disease and reduce production costs. However, the significance of disease characteristic of apple leaves in complex environment is relatively weak, and the fine-grain among different diseases of apple leaves is high, and the conventional feature extraction methods will lose the discrimination information. To solve these problems, an apple disease classification model based on multi-scale feature fusion is proposed in this paper. Firstly, the information flow of conventional residual network (ResNet) was improved to achieve efficient information circulation through changing the position of batch normalization and rectified linear unit (ReLU). Secondly, in order to solve the problem of serious loss of information in ResNet downsample, the channel projection and spatial projection of downsample were separated. Lastly, the 3×3 conv in ResBlocks was replaced by pyramid convolution, and the dilated convolution with different dilation rate was introduced into pyramid convolution to enhance the output scale of feature maps and improve the robustness of the model. The optimized model was verified on the dataset of this paper, and the optimized model had stronger anti-noise ability and better robustness, excellent learning effect and fast convergence speed. The classification accuracy on the original dataset is 94.24%, and that on the preprocessed dataset is 94.99%. The results demonstrate that the optimal model has a high accuracy, which can provide a reference for the prevention and control of apple leaf diseases.INDEX TERMS Apple disease, multi-scale feature fusion, deep learning, ResNet, classification.

An improved segmentation technique for multilevel thresholding of crop image using cuckoo search algorithm based on recursive minimum cross entropy

“…The size of the image is $$\,2048\times 1365$$ with a bit depth of $$24$$ and is taken from Refs. [10, 11]. The nature of these images is highly dense.…”

“…Mostly, the symptoms of crop disease are visible on the leaf, stem, and fruits. Hence, the disease can be identified by the colour changes on leaves, stems, and fruits [7,[9][10][11]. The healthy one is identified by changing colour while a drastic change in colour shows an unhealthy one.…”

“…In the literature, several approaches for image segmentation have been used [11][12][13][14][15]. Among all, thresholding plays a key role.…”

Dilated convolution capsule network for apple leaf disease identification

“…The results proved that the proposed method outperforms conventional state-of-theart deep learning models. Luo et al (2021) proposed an apple disease classification model based on a multi-scale conventional ResNet. To solve the problem of serious loss of information in the ResNet downsample, the channel projection and spatial projection of downsample were separated, the 3 × 3 convention in ResBlocks was replaced by pyramid convolution, and the dilated convolution with different dilation rates was introduced into pyramid convolution to enhance the output scale of feature maps and improve the robustness of the model.…”

“…The results proved that the proposed method outperforms conventional state-of-the-art deep learning models. Luo et al. (2021) proposed an apple disease classification model based on a multi-scale conventional ResNet.…”