Landscape identification and detection of various geographical features from satellite imagery using intellectual soft computing methods

Abstract

Modern remote sensing satellites generate multispectral data from a variety of sensors. Timely analysis of this excess data already presents a formidable challenge. Development of advanced technique for improving remote sensing image classification accuracy is essential for deriving reliable land cover information of both cultural and natural resources applications. Soft Computing methods are among the optimal tools for this type of applications. In the present research, the training time, classification time and the accuracy of the neural network and other conventional algorithms in classifying the satellite images of different study area are compared. The performance of different classifiers such as Deep Neural Network (DNN), Versatile Supervised Multi Resolution (VSMR) and Feed Forward Multilayer Perceptron (FF-MLP) Deep Learning classifiers are tested on two datasets. One is Landsat and another one is Space net. In DNN, the input image is preprocessed and removes the features like gray esteem and unkind gray distribution. It yields higher efficient segmentation quality and consumes less space and time complexity. The segmented image is utilized to achieve area established grayscale estimate, and then the strategy computes the grayscale distributional feature for different zones of the segmented image. Based on the grayscale distributional feature and the approached area value the technique classifies the satellite image under a diverse class of regions. This technique produces useful results in flood image classification and improved the accuracy and reduces the false classification ratio also. The work yields classification accuracy of 94.12% which is far better than earlier results in this engrossed arena of research. The presentation of the execution is examined, a comparison is also maderegarding the clustering accuracy, time complexity, and false classification ratio is presented. Versatile Supervised Multi-Resolution based method that automatically classifies the different regions from spatiotemporal re