Amidst changing climatic circumstances and the proliferation of plant diseases, the agricultural industry confronts formidable obstacles in guaranteeing food security. For crop management and production optimisation, early and precise diagnosis of plant diseases is essential. There is a need for more effective solutions since traditional illness detection techniques are frequently labour and time intensive. Deep learning methods have become effective instruments in this field, providing automatic and precise illness identification from pictures. The benefits and drawbacks of deep learning techniques used in plant disease detection are discussed in this paper's review of the field. Along with methods for gathering datasets, augmenting them, and optimising models, it covers a variety of topologies, including convolutional neural networks (CNNs) and its variations. It also looks at how to improve the scalability and efficiency of disease detection systems by integrating cutting edge technology like transfer learning with respective to smartphones along with like Internet of Things (IoT) gadgets and drones. The study highlights the significance of interdisciplinary collaboration across the domains of agriculture, computer science, and data analytics for sustainable agricultural practices. It also discusses future possibilities for research and possible applications.