Enhancing automated detection of plant leaf diseases through quantum-based triple-level depth separable residual networks and morphological feature extraction techniques

Timely detection of plant disease in agriculture is significant to prevent the from spreading and reduce crop loss. Recent developments in deep learning have resulted in raised interest in automated systems in order to detect plant disease, because of their capability to quickly and precisely categorize diseases on the basis of images. This work proposes a novel deep learning architecture, namely Quantum-based Triple-Level Depth Separable Residual Network, which integrates the strengths of quantum-inspired encoding, depth separable convolutions, and a triple-level attention mechanism. Initially, the input images are obtained from widely used plant leaf datasets, which provide multiple disease samples for disease detection. The collected images go through various pre-processing phases. Several morphological operations are utilized to evaluate the leaf’s structural properties. The pre-processed images are sent into the proposed model for extracting crucial attributes and classifying plant diseases. Especially, performance evaluation demonstrated that the proposed model’s performance is the finest compared to other methods concerning diverse performance metrics and computational efficiency. By the utilization of depth separable convolutions and the triple-level attention mechanism, the proposed model attains higher precision and accuracy in plant disease classification while needing minimal computational resources. The incorporation of quantum-inspired encoding assisted in capturing complex patterns within the image. Nevertheless, the proposed model demonstrated that it has superior performance in plant disease detection. The greater accuracy for the proposed model is 98.67% and the precision for the proposed model is 98.12%. This demonstrates the proposed mechanism’s remarkable performance rates in plant disease detection using advanced techniques.