Classification of Crop diseases through Remote-sensed data using Multi-class SVM

One of the main obstacles to quantitative yield prediction and choosing the control parameters that increase the yield of any crop is the detection of illness in crops. Therefore, a flawless system is required for illness identification to enhance it productivity. An innovative, cutting-edge method in other words, machine learning is playing an essential part in most real-time applications, including smart farming. The objective of this Documentation describes a system that uses image processing and machine learning to disease prediction for crops. Various dimensionality reduction, atmospheric correction, and other pre-processing methods before classification, end member extraction is carried out. Classification is done using Support Vector Machines, suitable aspects of images, such as colour and texture classifier. The SVM approach is regarded as the best and most accurate in machine learning. The deceased area is extracted by segmentation; colour-based segmentation is used in the process. The results obtained are classified using SVM as infected or not. If sick, the classifier identifies the illness and offers a potential cure. The diagnosis of leaf diseases is extremely labor intensive and requires knowledge of plant diseases. The purpose of this work is to create a computer program that automatically locates and categorized illnesses. The actions like loading a picture, segmenting, and pre-processing Disease detection involves extraction and classification The images of leaves are utilized for the diagnosis of plant diseases, using an image processing approach to locate and it is helpful to categorize illnesses in agricultural applications.