Automated Plant Disease Recognition using Tasmanian Devil Optimization with Deep Learning Model

Plant diseases have devastating effects on crop production, contributing to major economic loss and food scarcity. Timely and accurate recognition of plant ailments is vital to effectual disease management and keeping further spread. Plant disease classification utilizing Deep Learning (DL) has gained important attention recently because of its potential to correct and affect the detection of plant diseases. DL approaches, particularly Convolutional Neural Networks (CNNs) demonstrate that extremely effective in capturing intricate patterns and features in plant leaf images, allowing correct disease classification. In this article, a Tasmanian Devil Optimization with Deep Learning Enabled Plant Disease Recognition (TDODL-PDR) technique is proposed for effective crop management. The TDODL-PDR technique derives feature vectors utilizing the Multi-Direction and Location Distribution of Pixels in Trend Structure (MDLDPTS) descriptor. Besides, the deep Bidirectional Long Short-Term Memory (BiLSTM) approach gets exploited for the plant disease recognition. Finally, the TDO method can be executed to optimize the hyperparameters of the BiLSTM approach. The TDO method inspired by the foraging behaviour of Tasmanian Devils (TDs) effectively explores the parameter space and improves the model's performance. The experimental values stated that the TDODL-PDR model successfully distinguishes healthy plants from diseased ones and accurately classifies different disease types. The automated TDODL-PDR model offers a practical and reliable solution for early disease detection in crops, enabling farmers to take prompt actions to mitigate the spread and minimize crop losses