Transcripts
1. Introduction: Okay, welcome to this class. Very happy to have you here. In this class, we're
going to see how to reconstruct and
create Trev objects from images with a technique using artificial intelligence,
salt photogrammetry. The core idea of
photogrammetry is reconstructing Trev
objects like puzzles, you get the images by circling
around these objects. All these images have been created using adobe
substance, the di sampler. And you're going to
learn this class how to create Trev objects
for your video games, content creation, or projects
using photogrammetry and art to leverage your work and empower the treaty
production pipeline. You need to create
more treaty models and more beautiful treaty models to empower the narrative
of your unique story. Thank you for seeing
this video and I can wait you to have
you in the class.
2. Converting a 3D Ancient Tree Piece into 3D Object: Okay, this is a tree
we're reconstructing using photogrammetry and
adobe substance sampler. Enjoy this amazing and
beautiful nature tree. This is there match
reconstruction using adobe substance
tread sampler, artificial intelligence. Let's start dictatorial. We're reconstructing this tree. I took a lot of photos from all the circular round
at different heights. And these are the photos
we're going to use to reconstruct the object. Okay, now we go to a
substance red sampler. You click on File, New Project. We're going to click
on Trey Object Beta, the plus nutri object. Here we're going to browse
for all the images. We open them, we wait until
all images are processed. 46 images is important. The sensor size and focal
length will auto generate mask. If you have the information, it will be great that
you provide from your custom camera
you use in Mask. We're going to click
Generate new badge. This process, mask
like in Photoshop, took this object and
removes all the noise. If you don't do this process, it can generate all the
scenery and the scene. But we only want this mask. Okay for you to understand
what is masking is you have this image and we're
going to click here on Object Selection tool. This is going to
mask the object with artificial intelligence
as you see is mask, this is the process of you can for example selecting verse, you have the object extracted. This is the process carle
masking weekly badge. We wait until the
program finishes processing quickly submit precision
and photo ordering. Default or you can
select sequence. We're going to choose default. We wait until the cloud
point is generated. This is the point
cloud generated. These are all the images
that reconstructed. As you see, we have five images that are not correct
geometry details. We're going to select full for
the geometry details here. You can change the
bounding boxes. You can stretch it,
make it bigger. But I think it has inferred. Okay, We click Submit to the final processing
of object Cat. It's important to say that
you will need a good GPU. Designers need a good
graphic processor unit. Now, programmers do because artificial intelligence
is based also on GPU. For training, you must get some savings and get a
good GPU ever recommend. But any other equipment that can do the work is
okay, that has the capacity. You also need a lot of Ram
and a good disc space. Let's wait until it
reaches processing. Here we can see reconstruction. That is okay. Finally we
have a good reconstruction. Amazing. You will click here on, this is the solid
wire frame checker. We're going to add two project. We want the map baking. The advanced decimation
here is important. If you want this
with more details, you can create a new
process version. You can change the target phase count if you want
a higher quality. If you like the result, you click on a project. Here we have it. The last step is to check the math process. Here we have the height map, the bench normal,
and the base color. Weekly file export. We are going to export it. You select your folder, output select folder, and then
Nuclear Material settings, we're exporting base color, normal height map of the
merge as a object wavefront. We export, We're going to have here very soon the export
object and materials. Okay, finally here
we have the version one dot object object, the tola hemp and normal maps. In the next video, we're going
to import this on blender.
3. Blender Ancient Tree Render: Okay, we're going to learn
how to create this on Blender that you can render this image for your portfolio
cloner and render image. Now let's see in how to render now let's
click File Project. And don't say I'm going
to delete this cube. We're going to
click here on File, Import Wavefront object. We're going to select
Object we have. And as you can see, the material is there because
it has a MTL file. But if for some case you
have only the PNG's, you don't have this, you don't have the material
you are going to click. You have the object and the
PNGs with the channels, with the base color,
the roughness, the normal, the metal. You can click here Plus. But before you will need to
click on it Preferences, you will search for
the node Angular and a look that is enabled. Then you're going to
click on New Material under the shaving tool, and you're going to click here. And you're going to
click Control Shift to select Principal Textures. And you'll go to your folder, and you select Interclnormalasclorilkrincipalexturreconstructionn. You're going to go back to
layout for rendering purposes. You can simply add, for example, me plane, You click on to scale. You can you click on this
subject and you click to move. You click on the light, You
click to move the light. You can play these values
and control C control. You click R to rotate, X to rotate under the X to move. This is not a trial of blender, but I only want to show how
you can make a simple render. You can also add Light Son
will be great. Move this here. You can even add
texts to the planes. But for the purpose
of this tutorial, we also the camera. You can click here on
Etat Camera to move it. Right click, Set Active Camera, and you can and move it away. If we hold the camera and we click on Grab
and we put it away. Move away the set active camera you like here to
take a photograph. You also can play
with the lights. The song, for example, the
song has a warm light. We go here to the light and you can change
for a warm light. This, if you like to render the image image
as you can see here, you can have for your portfolio, you can play and make some images of the
object in high quality.
4. Photogrammetry Explained: Okay, now let's see some photogrammetry
concepts and why we're using a double
substance red sampler to do this three
object reconstruction. Photogrammetry is a
technique where you have an object and
you will take photos of that object by going around in different
angles positions. And by rotating the camera. By taking pictures
from different angles, it can create threat models or maps of the
photographed subject. It's like piecing
together a puzzle, using images to understand
the shape and size of things. But why is it important? Photogrammetry,
It can accurately captures real world details. Enhancing realism in digital environments
is cost effective, is an alternative to
traditional surveying. Reducing time and resources and empowering the treaty
production pipeline. It enables treaty documentation, preserving historical and
environmental landmarks. Why Adobe substance red sampler has built in photogrammetry
and is one of the best. It generates high
quality treaty models from an image
series using treat. There are other AI
photogrametry for creating treaty models using
the capture like meta shape and photo
catch of apple. Okay, but why is important photogrammetry to
create a more robust, efficient treaty
production pipeline? If you're creating a video game, a film, or you are
a content creator, you need to be more effective
in creating models, more treaty models and
better treating models. This can enhance
your production in seabro substance
painter in cinema. For the blender, you can create a lot of more
resources in less time. Also, you can do photogrammetry
by using drones. And you can scan, for example, mountains, cliffs, or trees.
5. Class Project: For class project. It
will be very simple. You will go to any
renderer you like. I'm using Blender and you will
import your treaty object. The only thing is you
need to make a render for an image which
selective camera image. I will share this image as my class project
image. Thank you. Your amazing image of the treaty object
you have digitized. We are going to be very
happy to see and comment and give our impressions to
support your amazing project.
6. Outro: Thanks for taking this class. Hope you make your
own tree model. And you take a photography and chair it as
your class project. You need to start
collecting and scanning tread objects to create
your own library of models. And then you can use
it for your project, for your company, and
for your portfolio. Thank you very much
for seeing this.
Serpent Academy, Passion for teaching and learning
