RETRACTED ARTICLE: A structured support vector machine for hyperspectral satellite image segmentation and classification based on modified swarm optimization approach

The hyperspectral image analysis improved by the most powerful and fastest growing technologies in the field of remote sensing in recent years. The hyperspectral image classification involves the identification and recognition by capturing spectral information over the region and consequently analysis by the pixel value. The conventional method uses the wiener filter for pre-processing and GLCM approach to extract the second order statistical features with dragonfly optimization technique for image extraction. The machine learning techniques used in the conventional technique is extreme learning machine and relevance vector machine. Here the high-resolution hyperspectral remote sensing datasets are taken from hyperspectral remote sensing scenes. This scene acquired by the AVIRIS sensor during a flight campaign over the Indian pines test site in Northwestern Indian. The hyperspectral images are filtered by a modified swarm optimization approach and these images are extracted by threshold-based segmentation process with the use of OTSU’s binary threshold method. The structured support vector machine is proposed for the classification of the satellite image. By the use of the optimization process, the structured support vector machine is improved its performance. Since overall sensitivity, specificity, and accuracy is improved. The simulation part carried out the data set for Indian pines and Salinas’s scene and the overall design is done with MATLAB.