350 rub
Journal Achievements of Modern Radioelectronics №11 for 2010 г.
Article in number:
Automatic recognition method of low detectable ground targets by airborne laser location assets
Authors:
I. M. Khmarov, V. Yu. Kanivets, N. G. Kondrashov
Abstract:
Now modern and perspective complexes of aviation arms are equipped with the integrated optical-electronic systems which structure includes laser range finders and laser systems of reception of three-dimensional (3D) images of the ground targets. Thus the problem of increase of probability of recognition airborne laser location assets the low detectable ground targets is actual. The work purpose is using the expressions, allowing to define factor of conformity of the accepted image to its standard, to develop the methodical approach of an estimation of degree of recognition of the low detectable ground targets two-dimensional and three-dimensional airborne laser location assets. Thus offered algorithms should provide identification of the land purposes in real time, and in difference from known (for example [1]) to have universal character. For a basis at calculation of probability of recognition of the low detectable ground targets the method of comparison with the standard is taken. At standard formation that fact is considered that the computing system has the data on inclined range D and angle of a place to ε location of object, and unknown persons still have only a purpose azimuth β. Thus, the quantity of generated standards depends on the chosen step of digitization on an azimuth and Δβ numbers of types z the analyzed low detectable ground targets. For formation of standards the specialized information calculating system "Rapier" is used developed with participation of authors. Information calculating system is developed in the programming system - Borland Delphi-Delphi-2006 - in the operating system "Windows-XP" and it is adaptive to a wide range of applied problems. By means of information calculating system the set of standards consisting from 2D and 3D of images for such land objects, as the helicopter, lorry ZIL, car and the tank depending on parameters D, ε, and Δβ was recired. Assuming a corner ε is known it is possible to draw a conclusion that at definition of type of the purpose which the accepted image concerns it is necessary to find the maximum value of factor of conformity for each of standards at all possible values of a bearing angle (which quantity depends from Δβ). And at a following stage it is necessary to define the greatest value among already known maxima for each type ground targets. Proceeding from it the problem of recognition of the land purposes with the help 2D and 3D laser location assets dares in four stages: 1. Preliminary classification ground targets taking into account their geometrical form and the sizes. 2. Tracking the image centre. 3. Comparison of the alarm and reference image by definition of the maximum factor of conformity for each type z. 4. Definition of the maximum value of factor of conformity among available maximum values for each type z ground targets and what of types the given factor concerns. It is the closing stage of recognition ground targets. As a whole the given stages form the algorithm of automatic identification ground targets which is based on the analysis of three-dimensional images and comparison of reference and alarm images. We will notice that mosern information technologies quite allow us to process the given volumes of the information in real time. Thus, on the basis of the expressions, allowing to define factor (measure) of conformity of the accepted image to its standard, the methodical approach of an estimation of degree of recognition of the low detectable ground targets on their binary two-dimensional and three-dimensional portraits received by means of airborne laser location assets is developed.
Pages: 86-90
References
  1. Nyberg, S., Borman, L., Assessing camouflage using textural features // Proc. SPIE. 2001. V. 4370. P. 60-71.
  2. Loyd, J., Sanders, J., Physically realistic camouflage net for visualizations and signature generations // Proc. SPIE. 2001. V. 4370. P. 72-83.
  3. Sobel, E., Douglas, J., Etinger, G., 3D LADAR ATR Based on Recognition By Parts // Proc. SPIE. 2003. V. 5094.