Abstract
Abstract. This article presents techniques for noise filtering of remotely sensed images based on Multi-resolution Analysis (MRA). Multiresolution techniques provide a coarse-to-fine and scale-invariant decomposition of images for image interpretation. The multiresolution image analysis methods have the ability to analyze the image in an adaptive manner, capturing local information as well as global information. Further, noise being one of the biggest problems in image analysis and interpretation for further processing, is effectively handled by multi-resolution methods. The paper aims at the analysis of noise filtering of image using wavelets and curvelets on high resolution multispectral images acquired by the Quickbird and medium resolution Landsat Thematic Mapper satellite systems. To improve the performance of noise filtering an iterative thresholding scheme for wavelets and curvelets is proposed for restoring the image from its noisy version. Two comparative measures are used for evaluation of the performance of the methods for denoising. One of them is the signal to noise ratio and the second is the ability of the noise filtering scheme to preserve the sharpness of the edges. By both of these comparative measures, the curvelet with iterative threshold has proved to be better than the others. Results are illustrated using Quickbird and Landsat images for fixed and iterative thresholding.
Highlights
In recent past, advancements in observing the Earth from space have led to a new class of images with very high spatial resolution
The application of multi-resolution analysis (MRA) for image analysis and interpretation has become very popular in recent past
The first version (Candes and Donoho, 2000) used a complex series of steps involving the ridgelet analysis of the radon transform of an image. Their performance was very slow; an improved version was developed known as Fast Discrete Curvelet Transform (FDCT) (Candes et al, 2006)
Summary
Advancements in observing the Earth from space have led to a new class of images with very high spatial resolution. Remote sensing images are used to extract some features, detect the presence of various phenomena, and for interpretation. These applications require high signal-tonoise ratio (SNR) to get correct results and better performance. By decomposing the image into a series of high-pass and low-pass filter bands, the wavelet transform extracts directional details that capture horizontal, vertical, and diagonal details. These three linear directions are limiting and might not capture enough directional information in remotely sensed images. The results with various performance measures are discussed with interpretations and direction of future work is presented
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
