A multi-spectral image-based high-level classification based on a modified SVM with enhanced PCA and hybrid metaheuristic algorithm

Abstract

In the context of Precision Agriculture (PA), an extremely fine crop classification has recently grown in importance. Satellite photos are often used to identify land use. Furthermore, this classification poses new difficulties since in addition to utilising the multi-temporal features of the multi-spectral images; it also calls for pixel-based examination and a greater number of trainings. Multiple machine learning methods are applied to multi-spectral and multi-temporal satellite data in this paper to create crop categorization models. The UAV platform is used for monitoring the crop field. The multispectral photo is caught as the intake figure which is taken out of the dataset. In precision agriculture, the pre-processing is performed for removing the noise content present in the input image. Feature section is done through PCA and the kernel modified SVM is used for classification. The implementation is performed using PYTHON platform. Paddy and Wheat are selected for the similarity. The accomplishment of the considered methodology is paralleled with and without the optimization, and also compared with the existing methods like Naïve bayes (NB), K- Nearest Neighbour (KNN), K- means and Random Forest (RF).