Preserving plants is preserving life itself, as they are the green lungs of our planet, sustaining all the living beings”. It is well-known that agriculture is the primary source of food production. Efficient and early detection of plant disease is crucial for preserving agricultural yield and quality. It requires a tremendous amount of work, time and expertise to detect the disease. Automated disease detection systems can be scaled to accommodate different crop types, regions, and environmental conditions. They can be adapted to various agricultural settings, including smallholder farms, large-scale commercial operations, and greenhouse environments. Hence to address these challenges, image processing techniques are employed by capturing the images of the leaf and comparing it with the dataset. It reviews methods like image acquisition, pre-processing, segmentation, feature extraction and classification steps. Along with this, it explores the usage of deep learning techniques, specifically Convolutional Neural Networks and the Residual Network(DenseNet) architecture, to automate and enhance the accuracy of plant disease detection. The utilization of CNN enables the system to learn discriminative features from the input images, allowing for accurate classification of diseased and healthy leaves. The trained models exhibit robust performance in distinguishing between different disease types and severity levels, enabling early detection and precise diagnosis of plant diseases.