Home         Authors   Papers   Year of conference   Themes   Organizations        To MES conference

Radar Image Autofocus in Conditions of High Vehicle Motion Instability
Authors
	Kuzmin I.A.
	Lyalin K.S.
	Meleshin Yu.
	Khasanov M.
Date of publication
	2018
DOI
	10.31114/2078-7707-2018-4-149-154
Abstract
	In the case of obtaining radar image in the conditions of high motion instability it’s impossible to define the synthesis parameters as they are changing during the aperture span. In real systems (mostly airborne SAR) the position of the vehicle is described by some curve. If the coherence span (for required resolution) is more than the aperture length, then instability of the motion can be considered as insignificant and to synthesize the image it’s enough to find just the velocity for each coherence span. If the span is less than aperture length, then the instability is considered as notable. In this case the autofocus process becomes more complicated by the necessity of selection the right velocity for each small coherence span and also each synthesized image is not a part of the global grid: each fragment may have random scale and relative offset. In this paper an approach to autofocus in the high motion instability conditions without using navigation data is proposed. As it’s been said the fragment to autofocus size is much less than aperture length, nevertheless each fragment contains the information almost about the whole aperture (the area being in the radiation pattern beam). This allows using the conventional autofocus algorithms, based on numerical optimization of synthesis parameters (velocity, radiation pattern angle offset) by some image quality criterion: entropy, contrast, or sharpness. Initially it was assumed that for obtaining the whole image it would be enough to know the velocities and the fragments sizes (to sum them up with corresponding scales and offsets). It turned out that the secondary motions (velocity fluctuation over a duration of one fragment, lateral motions) produce discrepancy between the real average velocity and that found as a result of the autofocus. This means that two neighbor images may have random scale and offset. In the work to find these parameters it was used the maximum correlation criterion. The scale is found by the sequential search with the search range decreasing at each iteration. The proper scale is considered that providing the best correlation maximum, the proper offset is the argument index of the maximum. The grid mismatch problem is also connected with the necessity of inter-fragment phase compensation. Skipping this stage produces artifacts and distortions in the final image (resolution degradation, false peaks appearance). The compensation is solved by the sequential search by minimum entropy criterion. The method was tested and approved on the real radar data. It has shown the possibility of performing the autofocus in the case of absence of navigation data even in the presence of high vehicle fluctuations and trajectory deviation. The obtained image azimuth resolution is at least 7 times better than the same obtained by the conventional autofocus algorithms and at least 4 times better than in the case of using high-precision navigation data.
Keywords
	synthetic aperture, autofocus, sub aperture synthesis, movement instability, correlation processing.
Library reference
	Kuzmin I.A., Lyalin K.S., Meleshin Yu., Khasanov M. Radar Image Autofocus in Conditions of High Vehicle Motion Instability // Problems of Perspective Micro- and Nanoelectronic Systems Development - 2018. Issue 4. P. 149-154. doi:10.31114/2078-7707-2018-4-149-154
URL of paper
	http://www.mes-conference.ru/data/year2018/pdf/D132.pdf