| 摘要 |
<p>Digital image processing process for constructing a calculated image, termed the target image (I<*>), which represents a source image (I) reconstructed together with modifications to the angles of elevation and azimuth, and to the scale factor in relation to the source image, comprising the determination of a common vanishing point (P) for the source image (I) and the target image (I<*>), the origin of an orthonormal reference frame; the calculation, for each pixel (M<*>) with address (X<*>, Y<*>) in the target image (I<*>), of spherical coordinates ( alpha , ss) by means of a linear function G<-1> such that ( alpha , ss) = G<-1> (X<*>, Y<*>), the coefficients of which are functions of the modified angles of elevation and azimuth and modified scale factor; and the reading, from a table, by the matching up, to the resolution of the table, of the spherical coordinates ( alpha , ss) and one of the addresses ( alpha 0, ss0) of the table, of a pair (X, Y) which is stored therein and which constitutes the address of a point (M) in the source image (I) corresponding to the pixel (M<*>) in the target image (I<*>). Application: surveillance cameras. <IMAGE></p> |
