Abstract

Effect Of Compression On Classification Of Remote Sensing Image

Balasubramanian Ramadurai
SG Systems Operator
College of Engineering Guindy, Anna University, India
vsrsburvr@annauniv.edu

Dr. S. S. Ramakrishnan
Professor & Director-SEASTAR
Institute of Remote Sensing, Dept. of Civil Engg.,
College of Engineering Guindy, Anna University, India
ssr@annauniv.edu

M. V. Hemadhri
SG Systems Operator
Institute of Remote Sensing, Dept. of Civil Engg.,
College of Engineering Guindy, Anna University, India
hemadhri@annauniv.edu

Abstract :
Data Compression is an important aspect of frame work which is expressed as a means and tools for the data originating from different sources. It aims at obtaining information of greater quality; the real definition of greater quality depends on the user requirements and the real world applications without loss or change of pixel values.The compression and decompression is achieved using Wavelet Transformation. The present scenario provides greater thrust for data compression. It finds greater use in medical and remote sensing images where, each pixel information is of greater importance to the decision maker. The other advantage being that large voluminous image data can be transmitted over the networks. This permits the decision maker to process and send back after modifications, in the base image data. Further, the importance of data compression must be lossless after decompression. Wavelets find a greater application in obtaining a lossless compression technique. The advantage is that it decomposes the data into various frequency components which helps to study and analyze each component with a resolution matched to its scale. They have advantages over traditional Fourier methods in analyzing physical situations where the signal is discontinuous random noise/spikes. Wavelet models are developed independently for various applications in the field of Mathematics, Remote Sensing, Turbulence, Seismic activities. Development of coding for compression using Wavelets involves more skill and time.

The results obtained are : Achieve excellent compression to the required size.

While classification there are mixing of classes in the compressed image-this is because there are changes in pixel value during decompression and

A feedback mechanism to be developed to overcome the changes in the pixel values on the compressed image so that the original values in the decompressed image can be reconstructed.

|