CONSIDERATION OF THE DESTABILIZING FACTORS INFLUENCE FOR INCREASING OF THE MEASUREMENTS ACCURACY OF A RANGEFINDER BASED ON STEREO IMAGES

The wide using of digital photography has led to significant progress in the development of the theory and methods of restoring the three-dimensional space picture on base of two-dimensional digital images. To solve the problem of increasing the measurements accuracy of such systems, it is necessary to take into account the influence of a number of destabilizing factors. The aim of this work was development of technique for accounting and compensating of destabilizing factors influence, such as the deviation from the horizontal position line of the stereo pair lens, the non-parallelism of the lenses optical axes, the mutual inclination of the photo detector matrices, and the distortion of the stereo camera optical system for increasing of the measurements accuracy of rangefinder based on the correlation analysis of the stereo image.

A software application has been developed for analyzing the optical distortions of serially produced lenses, which allows to visually demonstrate the distortions nature and to determine the polynomial coefficients for compensating of the optical distortion.It is obtained that for the Fujifilm FinePix Real 3D stereo camera the distortion of the digital image reaches ± 20–35 pixels at the edges of the photo detective matrix and is not the same for the first and second lenses. The difference in the optical distortion values is due to the unequal slope of the photo detector matrix to the optical axis of the objective. Compensating polynomials for the optical system distortions of the first and second lenses of the stereo camera are experimentally determined.

The range object expression from the stereo images taking into account the optical distortion compensation is obtained. It is shown for increasing of the measurements accuracy, the determining factor is not the absolute value of the lenses distortion, but the difference in the optical distortions of the stereo camera lenses, depending on the difference of the measured object coordinates of the on the photo detective matrices. Experimental studies of the developed technique for distortions compensation showed a decrease of the absolute measurements error more than by an order of magnitude at distances up to 100 m. 

1. Shapiro, L. Computer vision / L. Shapiro, D. Stockman – Moscow: BINOM, 2006. – 752 p.

2. Zolotukhin, D.A. Reconstruction and filtration of a three-dimensional shape of a microobject by a stereopair of images with a parallel projection. Zolotukhin, I.V. Safonov, K.A. Kryzhanovsky // Mechanics, management and informatics. – 2012. – V. 9, No. 9. – P. 174-179. (in Russian)

3. Forsythe, D. Computer vision. Modern approach / D. Forsythe, J. Pons. Moscow: Williams, 2004. – 928 p.

4. Hartley, R.I. Theory and practice of projective rectification / R.I. Hartley // International Journal of Computer Vision. – 1999. – Vol. 35, № 2. – P. 115–127, doi: 10.1023/A:1007957826135.

5. Zhang, Z. Camera calibration with lens distortion from low-rank textures / Z. Zhang, Y. Matsushita, Y. Ma // Computer Vision and Pattern Recognition (CVPR), 2011 IEEE Conference. 2011. P. 2321-2328 DOI: 10.1109/ CVPR.2011.5995548.

6. Alemán-Flores, M. Automatic Lens Distortion Correction Using One-Parameter Division Models / M.AlemánFlores, L. Alvarez, L. Gomez, D. Santana-Cedres // Image Processing On Line. 2014. Vol. 4. P. 327-343., doi: 10.1007/s10851-012-0342-2.

7. Bugaenko, EI A method for automatic detection and correction of radial distortion on digital images / EI Bugaenko, MI Trufanov // Izv. Universities. Instrument making. – 2008. – V. 51, No. 2. – P. 16-22. (in Russian)

8. Method for measuring distances on a digital camera: patent 2485443 Russian Federation, / VL Kozlov. – 2013. – Bul. No. 17.

9. Method for determining the distance to the object through a digital camera: pat. 18666 Rep. Belarus, / V.L. Kozlov; – 2014. – No. 5. – P. 95.

10. Titov D. V., Shirabakina T. A. Correction Device of Distortion // „Information and Telecommunication Technologies in Intelligent Systems". Proc. 5th Int. Conf. Spain, 2007. – P. 122–124.

11. Kunina, I.A. Blind compensation of radial distortion on a single image using a fast Hafa transformation. Kunina, S.A. Gladilin, D.P. Nikolaev // Computer Optics. – 2016. – V. 40, No. 3. – P. 395-403. (in Russian)

12. Gonzalez, R. Digital image processing / R. Gonzalez, R. Woods. – Moscow: Technosphere, 2005. – 1072 p.