Agriculture is the base for the growth of the economy of a country. Maintaining the health of crops is a significant part of it. Early identification of plant leaf diseases can help in dealing with and eradicating the various challenges. Using deep learning in detecting leaf diseases is a crucial and essential research area in the agriculture field. Such kind of research focuses on seventeen percent of the total gross domestic product (GDP) is contributed by agriculture in India. Enhancing the growth of crops can lead to growth in farmer’s profit, which further boosts the economy of the country. This paper presents a comprehensive review of various research works being carried out in the field of plant disease detection using deep-learning-based techniques. Using handcrafted feature-based approaches can be challenging. Models based on deep learning give an edge over such handcrafted feature-based techniques. The research shows that various deep learning approaches show significant accuracy on particular data sets, however, the accuracy declines when tested on field image conditions or on a separate data set. While reviewing it was found that models like GoogleNet and InceptionV3 have shown better results. Using transformer-based techniques could be a potential approach to detect plant leaf diseases.