V.A. Franz, O.M. Levina, V.V. Voronin, R.A. Kozhin

Currently, computer vision systems are widely used in many fields of technology. The solution to many problems is simplified by using a depth map in addition to the two-dimensional image. The depth map is a two-dimensional single-channel image containing information on the distance from the plane of the sensor to the objects in the scene. Unfortunately, the data obtained from the sensor, usually contain some defects that make the big error in the estimation of the distance to the object. An algorithm of automatic data preprocessing depth map that allows to detect and recover most of the defects present in the depths of the original map. To find the damaged sections of depth map is proposed to use the morphological «Top-hat transformation» and to restore the depth map reconstruction method «Fast Marching». Provides an assessment of the effectiveness of image processing algorithm to test image processing example test sequences from the sensor Kinect.

 

1. Szeliski R. Computer Vision. Algorithms and Applications. Washington: Springer. 2011.
2. Zach Ch. Fast and High Quality Fusion of Depth Maps. Department of Computer Science. University of North Carolina at Chapel Hill. USA. 2008.
3. Wei Huo, Gene Cheung, Xin Li, Oscar Au. Depth Map Super-Resolution Using Synthesized View Matching for Depth-Image-Based Rendering.Hong Kong University of Science and Technology.
4. JAne B. Cifrovaja obrabotka izobrazhenijj. M.: Tekhnosfera. 2007.
5. Telea A.An image inpainting technique based on the fast marching method // Journal of Graphics Tools. 2004. V. 9. № 1. P. 25-36.
6. Marchuk V.I., Voronin V.V. Rekonstrukcija znachenijj utra­chennykh pikselîâ izobrazhenijj v uslovijakh ogranichennojj apriornojj informacii // Nauchno-tekhnicheskie vedomosti SPbGPU. 2009. № 72. S. 52-56.
7. Voronin V.V., Marchuk V.I., Egiazarian K.O. Images reconstruction using modified exemplar based method // Image Processing: Algorithms and Systems IX. Ed. by J.T. Astola, K.O. Egiazarian. Proceedings of SPIE. V. 7870. 2011.
8. Frantc V.A., Voronin V.V., Marchuk V.I., Egiazarian K.O. Fast texture and structure image reconstruction using the perceptual hash // Proc. SPIE. V. 8655. Image Processing: Algorithms and Systems XI. 86550V. 2013.
9. Marchuk V.I., Voronin V.V., Franc V.A. Modificirovannyjj metod vosstanovlenija dvumernykh signalov // Nauchno-tekhnicheskie vedomosti SPbGPU. 2011. № 115. S. 31-36.
10. Voronin V.V. Metod rekonstrukcii izobrazhenijj na osnove interpoljacii granic obektov kubicheskimi splajjnami // Uspekhi sovremennojj radioehlektroniki. 2012. № 6. S. 26-30.