Change Detection of Images Based on Multivariate Alteration Detection Method

Advancement of Satellite technology gives rapid response to the complex problems and decision making. Due to the impact of climate change, loss of rainfall, drought, risen of sea level are some of the consequences. Recently, the cyclone Fani hits oddisa severely and the livelihood of the people is in standstill. The satellite images of before and after Fani shows the severity of the cyclone. This paper, we introduced three methods which includes Large Scale Mean Shift (LSMS), Multivariate Alteration Detection (MAD) and Local statistic Feature extraction methods are used to detect the changes in the heterogeneous images. The satellite images of before and after the cyclone Fani is considered for processing and the severity is estimated. The satellite images are calibrated and preprocessed for removing the speckle noises. With the application of regression prediction algorithm used to tune the accuracy of segmentation and change detection process. The model is tested with the satellite image and the results are very promising. This approach can also be applied in medical images to detect the changes in features, which helps the physician to diagnose the decease.