Accuracy assessment for SAR image classification using laws mask features performance analysis of SAR image classification

Synthetic aperture radar (sar) provides very high resolution imaging by taking the advantages of long-range propagation characteristics of radar signals and the complex information processing capability of modern digital electronics. in this project a proposed method for the supervised classification of sar images has been done by using laws textural features. here we also introduced traditional classification methods namely k means clustering, fuzzy c means clustering, and tamura features. since the input sar image is a form of supervised classification training sets data are taken and then laws mask features are calculated and then classification is carried out based on minimum distance classifier. this is done because specific laws texture mask extracts feature information efficiently, compared to other texture classification methods. to prove the efficiency of the proposed work, sar image is classified using previous traditional methods and the results are compared using error matrix. in this error matrix two main parameters (user accuracy and producer accuracy) are calculated for comparison purpose. The output classification results show that the proposed method gives better classification accuracy (overall 80.83%) when compared to the other traditional methods.