RESUMO
With an emphasis on classifying diseases of sugarcane leaves, this research suggests an attention-based multilevel deep learning architecture for reliably classifying plant diseases. The suggested architecture comprises spatial and channel attention for saliency detection and blends features from lower to higher levels. On a self-created database, the model outperformed cutting-edge models like VGG19, ResNet50, XceptionNet, and EfficientNet_B7 with an accuracy of 86.53%. The findings show how essential all-level characteristics are for categorizing images and how they can improve efficiency even with tiny databases. The suggested architecture has the potential to support the early detection and diagnosis of plant diseases, enabling fast crop damage mitigation. Additionally, the implementation of the proposed AMRCNN model in the Android phone-based application gives an opportunity for the widespread use of mobile phones in the classification of sugarcane diseases.
RESUMO
Deep learning practices in the agriculture sector can address many challenges faced by the farmers such as disease detection, yield estimation, soil profile estimation, etc. In this paper, disease classification for the sugarcane plant and the experimentation involved thereby is thoroughly discussed. Experimental results include the performances of the well-known existing transfer learning techniques and proposed ensemble deep learning based architecture that incorporates stack ensemble of two networks with one having level-wise spatial attention helping to provide better generalization. A Self-created database of sugarcane leaf diseases is introduced to the research community through this paper. It involves 5 categories with a total of 2569 images. Here, it is observed that best performing transfer learning method, MobileNet-V2 shows an accuracy of around 84% with the lowest number of parameters whereas ensemble model reaching to 86.53% with less epochs and with acceptable number of parameters.
