How to compute the homography

46 visualizaciones (últimos 30 días)
Sushil  Sharma
Sushil Sharma el 18 de Abr. de 2019
Comentada: Matt J el 28 de Dic. de 2019
Dear Community,
I have some issue to getting the trasnform image for homography
I want to apply the homography to an image, I do have homography martix from the general formulation
H = R+1/d*n*t'
H = rotation_matrix + t.* Normal
H =
0 -1.0000 0
-0.3420 0 -0.9397
5.9397 5.0000 4.6580
The basic concept of idea is to trying get image from vitrual camera, so the question I that how I can get the image from virtual camera ?
As you can see figure below , I have perspective image but I want to use inverse perspective mapping
Thanking you so much
3D2D.jpg

Respuesta aceptada

darova
darova el 19 de Abr. de 2019
Hi, that green trapezoid is what you want. Can be found as line intersections of plane
mmm.png
But having all those point on a plane (in 3D) dont know how to convert them into 2D (to get that trapezoid in XY):
mm1.png
And how to convert image (pixels position) having that trapezoid?
mm2.png
See attached files
  3 comentarios
darova
darova el 21 de Abr. de 2019
Create vectors Nx, Ny first:
NX = cross([0 0 1],-NZ); % camera X-axis
NX = NX/norm(NX);
NY = cross(NX, NZ); % camera Y-axis
NY = NY/norm(NY);
Then using dot product convert each point in 2D:
Where, P = (x,y,z) - point on plane
Untitled.png
Sushil  Sharma
Sushil Sharma el 21 de Abr. de 2019
Thank you so much darova for your answer,
I get some idea, now I understand how to deal with situation,
Really apppericate for your help,
I wish in future you will clear my doubt as you did now
Thanks in advance

Iniciar sesión para comentar.

Más respuestas (1)

Matt J
Matt J el 19 de Abr. de 2019
tform = projective2d(H);
newImage = imwarp(oldImage,tform);
  2 comentarios
Omar Ali
Omar Ali el 28 de Dic. de 2019
how i can apply it on input image?
Matt J
Matt J el 28 de Dic. de 2019
That's what this line does
newImage = imwarp(oldImage,tform);

Iniciar sesión para comentar.

Categorías

Más información sobre Geometric Transformation and Image Registration en Help Center y File Exchange.

Etiquetas

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by