Landsat TM Satellite Image Restoration Using Kalman Filters

The quality of satellite images propagating through the atmosphere is affected by phenomena such as scattering and absorption of light, and turbulence, which degrade the image by blurring it and reducing its contrast. The atmospheric Wiener filter, which corrects for turbulence blur, aerosol blur, and path radiance simultaneously, is implemented in the digital restoration of Landsat Thematic Mapper (TM) imagery. Digital restoration results for Landsat TM imagery using the atmospheric Wiener filter were presented in the past. Here, a new approach for digital restoration of Landsat TM imagery is presented by implementing a Kalman filter as an atmospheric filter, which corrects for turbulence blur, aerosol blur, and path radiance simultaneously. Turbulence MTF is calculated from meteorological data. Aerosol MTF is consistent with optical depth. The product of the two yields atmospheric MTF, which is implemented in both the atmospheric Wiener and Kalman filters. Restoration improves both resolvable detail and contrast. Restorations are quite apparent even under clear weather conditions. Although aerosol MTF is dominant, slightly better results are obtained when the shape of atmospheric MTF includes turbulence, in addition to that of aerosol MTF, as shown by the use of criteria for restoration success. In general, the Kalman restoration is superior. Introduction Satellite images obtained by imaging vertically through the atmosphere are blurred, distorted, and exhibit poorer contrast relative to ideal images (i.e., images without degradation). During recent years, there has been an effort to develop methods of restoration using the atmospheric optical transfer function in order to compensate for image degradation. Images propagating through the atmosphere are attenuated by absorption and large angle scattering by aerosols. They are also blurred by small angle scattering caused by aerosols, and by optical turbulence (Kopeika, 1998a). In the remote sensing community essentially all atmospheric blur in satellite imagery is attributed to small-angle light scatter by aerosols and is called the adjacency effect because it causes photons to be imaged in pixels adjacent to those in which they ought to have been imaged. A detailed summary of much of the adjacency effect literature, including numerical calculations, experimental results, Monte Carlo simulations, resolution measurements, image correction methods based on aerosol scatter blur, etc., is found in Kopeika et al. (1998a). In part of the propagation community, however, image blur is often attributed solely to turbulence. This assumption often leads to questionable results which do not correlate well with Landsat TM Satellite Image Restoration Using Kalman Filters D. Arbel, E. Cohen, M. Citroen, D.G. Blumberg, and N.S. Kopeika turbulence theory, as described in Kopeika et al. (1998b). It is our view that image blur through the atmosphere involves both turbulence and small-angle forward scatter of light by aerosols (Kopeika, 1998a; Kopeika et al., 1998b). Interestingly enough, much of the coherence studies of turbulence and aerosol light scatter were carried out by the same researchers, whose pioneering works are also well known in each community. The aerosol modulation transfer function (MTF) contributions of Fante, de Wolf, Ishimaru, Lutomirski, etc., who are well known for their works on turbulence, are summarized in Kopeika et al. (1998b) and Kopeika (1998b). Those references also contain numerous examples of the problems arising in works by more recent members of the turbulence community when they ignore blur contributions caused by small-angle forward scatter of light by aerosols. Here, we consider the opposite phenomenon, i.e., the inadvisability of ignoring turbulence blur in the remote sensing community, which generally assumes all atmospheric blur is caused by aerosols (adjacency effect). A question to be considered is whether image restoration should be based on aerosol MTF alone or, instead, on the product of aerosol and turbulence MTF. Results are compared here. The influence of aerosol and optical turbulence strength on laser beam widening in the atmosphere is described in Zilberman et al. (2001) and Kopeika et al. (2001). This is equivalent to point spread functions at various elevations. It was mentioned there that the beam widening caused by atmospheric aerosols is significant at higher levels of the atmosphere (up to 20 km). On the other hand, those measurements indicate that turbulence is dominant at the lower level of the atmosphere, and then decreases sharply as the altitude increases. Little information about higher levels of the atmosphere is available yet. During recent years, there has been an effort to develop methods of restoration and filtering of images while using atmospheric optical transfer functions (MTF and phase transfer function) in order to compensate for image blur and distortions. Use of the standard Wiener filter for correction of atmospheric blur is often not effective because, although aerosol MTF is rather deterministic, turbulence MTF is random. The atmospheric Wiener filter (Sadot et al., 1995; Kopeika, 1998a) is one method for overcoming turbulence jitter (fluctuations). The atmospheric Wiener filter, which corrects for turbulence blur, aerosol blur, and path radiance simultaneously, is implemented here in digital restoration of Landsat Thematic Mapper (TM) imagery. Digital restoration results for Landsat TM imagery using the atmospheric Wiener filter were presented in the past (Sadot et al., 1995; Arbel et al., 1998; Arbel et al., 1999; Arbel and Kopeika, 2000). Here, a new approach for P H OTO G R A M M E T R I C E N G I N E E R I N G & R E M OT E S E N S I N G J anuary 2004 9 1 Department of Electrical and Computer Engineering and Department of Geography and Environmental Development, Ben-Gurion University of the Negev, P.O.B. 653, Beer Sheva, 84105, Israel ([email protected]; blumberg@bgumail. bgu.ac.il; [email protected]). Photogrammetric Engineering & Remote Sensing Vol. 70, No. 1, January 2004, pp. 91–100. 0099-1112/04/7001–0091/$3.00/0 © 2004 American Society for Photogrammetry and Remote Sensing 02-060.qxd 12/11/03 4:01 PM Page 91