Sovremennye problemy distantsionnogo zondirovaniya Zemli iz kosmosa, 2025, Vol. 22, No. 2, pp. 28-41
Selection of reference spectrum in the constrained renormalization method to reduce speckle noise in an ALOS-1 radar image
1 Kotelnikov Institute of Radioengineering and Electronics RAS, Fryazino Branch, Fryazino, Moscow Region, Russia
Accepted: 10.02.2025
DOI: 10.21046/2070-7401-2025-22-2-28-41
The application of the constrained renormalization method (CRM) to images obtained using one of synthetic-aperture radar (SAR) systems is considered. This is done in order to combat speckle noise. The possibility of a significant reduction in the speckle noise level appears due to the fact that the CRM renormalizes the spectrum of the source image to the reference spectrum model, which is a model of the spectrum of an optical image of good quality. Due to peculiarities of the ALOS-1 (Advanced Land Observing Satellite) radar data collection technique, the image under study is strongly stretched along one of the coordinates and compressed along the other. It is shown that these transformations lead to changes in the spatial spectrum that should be used as a reference in the CRM procedure, therefore, the application of the classical model of universal reference spectrum in CRM leads to unsatisfactory results. The method of selecting the reference spectrum is demonstrated step by step. It has been established that correct choice of the initial data significantly impacts the final result. An optical aerospace image without distortion (without stretching and compression along the axes) is taken as the reference image. In order for the renormalization procedure to lead to correct results, it is necessary to produce the above distortions over the reference image. That is, an image of good quality, for this particular task, is this optical reference image stretched and compressed along the axes in exactly the same way as the SAR image constructed from ALOS-1 radar data. The use of CRM in combination with median filtering has been tested. It is shown that the combined use of CRM with a method that is different in ideology gives a gain in the final result, since possible loss of useful information by one of the methods will be compensated by its backup.
Keywords: digital images, remote sensing, image processing, speckle noise, reference spectrum, constrained renormalization method
Full textReferences:
- Gulyaev Yu. V., Zrazhevsky A. Yu., Kokoshkin A. V., Korotkov V. A., Cherepenin V. A., Correction of the spacial spectrum distorted by the optical system using the reference image method. Part 3. General-purpose reference spectrum, J. Radio Electronics, 2013, No. 12, 16 p. (in Russian), http://jre.cplire.ru/jre/dec13/3/text.html.
- Ilyushin S. V., Speckle suppression in medical ultrasound images using fractal coding, T-Comm, 2011, No. 3, pp. 22–26 (in Russian).
- Kokoshkin A. V., Using the method of renormalization with limitation for medical ultrasound image processing, J. Radio Electronics, 2020, No. 10, 12 p. (in Russian), https://doi.org/10.30898/1684-1719.2020.10.1.
- Kokoshkin A. V., Application of the method of renormalization with limitation to remote sensing images obtained with synthetic aperture radars, J. Radio Electronics, 2021, No. 3, 12 p. (in Russian), https://doi.org/10.30898/1684-1719.2021.3.4.
- Kokoshkin A. V., Features of an application of the constrained renormalization method to images with speckle noise, J. Communications Technology and Electronics, 2022, V. 67, No. 12, pp. 1419–1429, DOI: 10.1134/S1064226922120075.
- Kokoshkin A. V., Korotkov V. A., Korotkov K. V., Novichihin E. P., Using the method of renormalization limited to restore the distorted image in the presence of interference and noise with unknown parameters, J. Radio Electronics, 2015, No. 7, 11 p. (in Russian), http://jre.cplire.ru/jre/jul15/4/text.html.
- Kokoshkin A. V., Novichihin E. P., Smolyaninov I. V., Application of spectral and spatial processing methods to sonar images, Radioelektronika, Nanosistemy, Informacionnye Tehnologii, 2021, V. 13, No. 3, pp. 377–382 (in Russian), DOI: 10/17725/rensit.2021.13.377.
- Razmanov V. M., Krivtsov A. P., Dolotov S. A., Specific features of measuring seafloor topography with an interferometric side-looking sonar, J. Communications Technology and Electronics, 2006, V. 51, No. 1, pp. 52–58, https://doi.org/10.1134/S1064226906010074.
- Achim A., Bezerianos A., Tsakalides P., Novel Bayesian multiscale method for speckle removal in medical ultrasound images, IEEE Trans. Medical Imaging, 2001, V. 20, Iss. 8, pp. 772–783, DOI: 10.1109/42.938245.
- Isar A., Firoiu I., Nafornita C., Moga S., SONAR images denoising, In: Sonar Systems, Kolev N. Z. (ed.), 2011, DOI: 10.5772/19190.
- Ghazel M., Freeman G. H., Vrscay E. R., Fractal image denoising, IEEE Trans. Image Processing, 2003, No. 12, pp. 1560–1578.
- Gonzalez R. C., Woods R. E., Digital image processing, 3rd ed., Pearson Prentice Hall, New Jersey, 2008, 976 p.