Abstract :
Today, according to product automatisation systems, which are used for building of three dimension models of technological environment, enginery and machinery, became widespread. Many modern fast - acting scanning laser systems are used only in unmanned situations because of danger for human. It means development of passive optical systems of three - dimensional vision. In this work algorithm of inverse viewing transformation is suggested, which based on determination unknown parameters of system of equations for point in three - dimensional space: v(x, y, z) = A(ρ1, ρ2, φ1, φ2, θ1, θ2, d)· w(x1, x2, y1, y2), where w - vector of screen coordinates, v - vector of coordinates of point in three - dimensional space, A - generalized matrix of rotation and transfer. Common area of images is detected using correlation analysis with optimization of computational complexity and growing of data accuracy on each step of calculation of correlation function of images of object.
Keywords :
computational complexity; computer vision; correlation methods; matrix algebra; optimisation; stereo image processing; 3D model building; 3D vision; computational complexity; inverse viewing transformation algorithm; multistage correlation analysis; optimization; product automatisation systems; rotation matrix; stereo image analysis; transfer matrix; Algorithm design and analysis; Application software; Cameras; Clouds; Computational modeling; Computer vision; Data visualization; Image generation; Layout; Robot vision systems; 3D recognition; inverse transformation;