Hi,
I have various perspective photos of a vehicle and need to add structural details from these to my model drawn so far by referring to the photos. Can I somehow place them into a view window then pick points on the photo and same points on the model and have Rhino adjust model orientation and camera lens value etc so they match picture #1, so I can then work out on the model where the structural features I need to create, exist ?
If pic in front of model then make pic translucent via properties.
Save this view, add another picture, this time of other side of vehicle, do same thing, etc. Ability to restore that view with that picture.
I would do it all manually. You can position the images around the model and save them into layers. You can save views as NamedView. Camera length can be altered when you’re looking through its perspective, click into the background and on your right side you can see panel Properties with Lens Length option. However, I’m afraid using photos as a direct reference for tracing is usually used an option only in ortho views.
Hi,
I ran the perspective match video, I thought it was mapping a pic to an object, but I see it maybe of use, however it fails.
I import pic, then pic point#1 on pic (wish I could zoom in to pic to do so, having to be ultra careful, but need to zoom in else I cant pick the points accurately.
pick that point on model
then it asks point#2 pick that on picture and it says start over.
I cant get beyond this stage.
No way manually can I get rhino to orient and adjust perspective view and camera, take ages, to map to a photo. I did try it but it doesnt work,
This is what I need, but it fails every time. WHY DOES IT FAIL ?
I now remember this from a previous project, it did the same failure there, Never was able to get this to work at all. never gets beyond the second pick on model.
Once you Run the Perspectivematch command you cannot zoom, rotate or pan from what I remember. It’s a little tricky in the beginning. But I was able to get a good result after a few attempts, granted it has been quite some time since I’ve used it.
Are you saying once you start picking points dont attempt to zoom or move model to get at the second point etc ?
That makes life v tricky if not impossible as I get all sorts of picks happening unless I can zoom in on just where it should be., bad enough not being able to magnify the picture to pick an accurate point on it. One needs to be able to zoom both.
Furthermore one needs to move the model about so as to see the picture, as its shaded mode, ghosted and xray also deny visibility of picture, and wireframe is a mess, one cannot see the bit to click on without the sense of shaded.
If its failing due to that, then thats a real pain, its impractical ! Just cannot see the picture without moving the model about, totally obscured.
Could someone do a video with a detailed shaded model requiring accurate picks to show how its done ?
By turning wireframe on and off I managed to get to 7 points then hit enter and it failed to match the model to the picture, near to the view but also wrong angle at vehicle, front of lorry cabin wasnt visible yet was in picture. Try again with bottom of two tyres and cabin front, again fail.
This needs a MASSIVE makeover as a command. The vehicle is accurate and its a quality pic, but fails to alter camera and angle etc.
I try another pic, rear view, place 6 points, nothing happens. Try again, very carefully, nothing happens. Its impossible to place points on the model without zooming in to ensure the right place is picked. Maybe one point was out, but unless we can zoom to place, this will happen.
It needs something totally different, pre-place the points then run the command. Maybe use points on a layer and filter just points. Nope, they move when you zoom !
That video is all well and good but life is more than a simple box !
Hmm, but each model has bounding box and from there points of intersection PerspectiveMatch is derived. (two point perspective)
Example from Photoshopcafe…com
Hi,
The picture of a lorry, I am drawing up all the structures aft of the cab, I have drawn the cab, but without it already drawn I cant form a box to the structures. How do I create a box in the picture when the back end is not like a modern box juggernaut but has mudguards, a compartment between them, various fittings, it cant be boxed easily. One would have to establish perspective lines in pshop, there isnt enough on it for working those out. Even the blue car, if someone was presented with that and asked to draw a perspective box, there is only wheels at side that can give a vanishing line. One needs staright lines on the subject which are parallel to do so.
As I said life isnt that simple.
I didnt think the user had to draw a box on his photo. I cant box whats in this photo, I also cant box the model as only thin chassis is drawn and a cab, but no bonnet.
I need this Matchperspective so as to create those parts. is this a catch22 then ?
I cant see the point of this to assist creation if the creation has to be done beforehand.
I am taking cab far, corner, cab near front corner,cab rear nearside corner (cant see far side), rear chassis members, tyres in fact not good as if picture is under load, and mine is not, then the variables will fail.
I think Perspective match command doesn’t really work. At least it has never worked for me. I’ve had much more success doing this by hand matching the camera less to what I think the photo used, and then tweak the perspective view until it looks right.
Hi, my post seems to have vanished, I tried ortho view, established the sight lines of the photo on my model in top view, created a cone 168cm tall for my eye level, another for target. placed where sight lines met.
created a cplane, zero at the camera cone, set camera at the cone, set target at the other cone set up middle of view, going by the photo.
pic looking now sort of …ok,but model is lacking the perspective to match it.
viewport properties and alter parallel to perspective.
ah now we have it,
HOWEVER I cant draw a single thing, unless I start from an existing object !
unable to plot things from picture onto model !
any zoom etc loses the view.
Maybe I should draw on the photo, group it, then project that onto my model ?
Creating our own far better method of Perspective match, but its a poor alternative, but to have rhino match the models view and perspective and camera length is critical in this being accurate.
Cplane and camera target etc doesnt do that, picking points on the model and having them mapped to a picture is the important thing, but ther is no ability to freeze the scene and then draw on the model, even place points would be useful, but I need to zoom to be accurate with placement. and that loses the function.
NEED ZOOM WITHOUT ALTERING CAMERA TO TARGET SIGHTING LINE
yeah you cannot get a ‘closer look’ to what you are doing without altering the view, and without changing perspective. Rhino should have a magnifier glass that is GPU accelerated and refined. A sort of 2D zoom over a 3D viewport.
oh wait, there is a way! (total hack)
Step 1:
Save the view you want as your matched perspective, in this case:
Step 2. Create a new layout page using the command _Layout:
Make the default view of the layout active (in this case it was a top view) by double-clicking on it so it becomes a live viewport, and load your saved camera_match01 view:
step 3: double-click again to make viewport inactive (now you are back out in paper space):
Step 4: Zoom to where you need:
Step 5: Double-click again to make viewport active (no zooming/panning now!) and draw…
so, if I undestand the procedure to orient correctly a model in existing view, I could build the perspective grid using the existing view in Carapace.
Then I can insert in Rhino that grid as _BackgroundBitmap and I could start to play with Camera parameters until I can find the the most acceptable result.
Correct?
