Enhance satellite image classification based on fuzzy clustering and Marr-Hildreth algorithm

The Satellite image classification is a significant method recycled in remote detecting for the automated study and structure recognition of the satellite data, which facilitate the automated understanding of the huge amount of data or information. These days, there happen the various kinds of classification methods, such as parallelepiped and less distance classifiers, but it is static essential to get better their presentation in relations of correctness rate. After existing work, we study the cross breed fuzzy, c-mean clustering algorithm and FFNN classifier for rocks cover planning of lands, shadow, construction. It initializes with the separate step pre-processing process to create the picture appropriate for segmentation. The processing of the image is refined using the cross-genetic-Artificial Bee Colony (ABC) algorithm that is implemented by cross breading the ABC and Fuzzy c-means to find the effective division in satellite image and categorized using NN. The presentation of the offered hybrid-algorithm is compared performance with the algorithms like, Artificial Bee Colony (ABC) algorithm, ABC-GA algorithm, Affecting KFCM. We implement the edge detection using Mar Hildreth Edge Detection Algorithm; Fuzzy c means Clustering using for segmentation of the satellite image, to extract the features used Principle component analysis, to optimize the extracted feature using Bacteria Foraging Algorithm and classification SVM (Support Vector Machine) to execute the satellite picture classification. In minimum time and evaluate the better accuracy based on the support vector machine algorithm. We design the framework in satellite images classifies uses MATLAB 2013a simulation tool and evaluate the performance in the Xdb-Index and DB-Index.