Identification of the crop is an important aspect for preventing the losses of the agricultural product. The study of crop diseases means studying the visually observable patterns observed on the crop. Disease detection on crops is very unfavourable for sustainable agriculture. It is very difficult to monitor the crop diseases manually. It requires an enormous amount of work, professionals in the crop diseases, and requires extreme processing time. Hence, image processing is used for the detection of crop diseases. The proposed system involves five important steps: a) image capturing, b) image pre-processing, c) image segmentation, d) feature extraction and e) classification. Finally, features are trained with the machine learning algorithm and tested on unknown images. Results are being calculated based on qualitative and quantitative analysis. In qualitative analysis, pre-processing and segmentation methods are operated on the whole data set. In quantitative analysis, the accuracy of all machine learning algorithms is computed. This paper discussed the methods used for the detection of crop diseases using their leaf image. This paper also discussed some segmentation and feature extraction algorithms used in crop disease detection. The classification is done using Gradient Boosting Classifiers with the highest accuracy.