Combines two rotation-and-shift transformations

S = cv.composeRT(rvec1, tvec1, rvec2, tvec2)

Input

• rvec1 First rotation vector, 3x1 float vector.
• tvec1 First translation vector, 3x1 float vector.
• rvec2 Second rotation vector, 3x1 float vector.
• tvec2 Second translation vector, 3x1 float vector.

Output

• S A scalar struct with the following fields:
  • rvec3 Rotation vector of the superposition, 3x1 float vector.
  • tvec3 Translation vector of the superposition, 3x1 float vector.
  • dr3dr1, dr3dt1, dr3dr2, dr3dt2, dt3dr1, dt3dt1, dt3dr2, dt3dt2 Derivatives of rvec3 or tvec3 with regard to rvec1, tvec1, rvec2, and tvec2, respectively. Each derivative is a 3x3 float matrix.

The function computes:

rvec3 = rodrigues^-1( rodorigues(rvec2) * rodrigues(rvec1) )
tvec3 = rodrigues(rvec2) * tvec1 + tvec2

where rodrigues denotes a rotation vector to a rotation matrix transformation, and rodrigues^-1 denotes the inverse transformation. See cv.Rodrigues for details.

Also, the function can compute the derivatives of the output vectors with regards to the input vectors (see cv.matMulDeriv). The function us used inside cv.stereoCalibrate but can also be used in your own code where Levenberg-Marquardt or another gradient-based solver is used to optimize a function that contains a matrix multiplication.