Rose Plant Leaves: Disease Detection and Pesticide Management using CNN

Agriculture is one of the main livelihoods and a major source of income for farmers. Rose cultivation is one of the highest contributing income sources to India. But the main problem in the cultivation of rose plant is the diseases on the plant leaves which will affect its growth. The most commonly occurring diseases on rose plant leaves are Blackspot, Rose mosaic, Rose aphids, and Powdery mildew. Manual methods for disease detection take more time and expensive. Hence, there is a need for an intelligent system based on advanced Deep Learning (DL) architectures which are the sub-field of artificial intelligence. Convolution Neural Network (CNN) is one of the deep neural networks used for many image based object detection tasks. The proposed work employs the CNN model for the detection of the above-said four types of diseases on rose plant leaves. The images were acquired from the real field and some are from an existing data set. The pre-processing such as image resizing and image augmentation of the data set were performed before training with CNN architecture using the optimized parameters. The trained architecture was tested using the test data set and the results were analyzed. Based on the results, an appropriate pesticides was prescribed.