3D Scanning of Real Objects
Luma AI is a user-friendly software designed for beginners, eliminating the need for a background in 3D modeling or graphic programming. It allows users to generate 3D digital models by directly scanning real objects. Below is a guide on how to use Luma AI to create 3D digital model assets that can be uploaded to the Unity SDK.
Download and Installation
Download and install the Luma AI app on your mobile device. Currently, it only supports the iOS system. Here are the requirements for mobile devices:
Register and Log in on the Mobile App and Official Website
Create
There are two ways to create a 3D model:
AI Generation
After logging in, click on the "Create" button on the right side of the home screen to enter the AI generation page.
In the text box below, enter a description of the model you want to generate, such as "A pink cup".
Then click on the generate button on the right side of the input box and wait for the corresponding model to be generated.
Click on the model you want.
Create Your Own
After logging in, click on the "+" in the middle of the bottom of the screen to enter the creation page. Choose the method of creating your project. There are three ways in total:
3.1 Method One: UPLOAD (upload a local video) - Automatically recognize the model object by uploading a local video.
3.2 Method Two: OBJECT
Capture the model by moving your phone. The steps are as follows:
Complete the tutorial as prompted;
Click on the "Confirm" button to start capturing the model.；
Take note of the following during the capture process:
Capture Orientation: Rotate the capture from three different heights:
a、Align the phone with your chest, looking straight ahead.
b、Position the phone slightly above the head, pointing towards the center of the scene.
c、 Finally, capture from just above knee level.
Capture Speed: Motion blur significantly reduces the quality of reconstruction. To achieve the best effect, move the phone slowly and avoid rapid movements, especially rotations.
Scene Coverage: For optimal results, capture the model or scene from as many unique viewpoints as possible. Also, move the phone in 3D space during capture, rather than rotating the model from a stationary position. Guided capture mode is a good choice to ensure sufficient coverage.
Model Size: For guided captures, the more comprehensive the coverage of the model from all angles (including top and bottom), the better the results. Therefore, capturing larger models may not yield perfect results.
Model Distance: For best results, try to keep the entire model within the frame during scanning. This provides the app with more information about reflections and model shape for more accurate reconstruction.
Model Material: Currently, the app struggles with complex reflections (like curved mirrors), transparent models (like car windows or plastic bottles), and very large textureless surfaces (like white walls). Most other materials work well.
Capture Environmental Lighting: The app can capture models in most lighting conditions as long as the texture remains recognizable (i.e., not colorless or completely dark).
Moving Models: Any movement in the scene during capture can reduce the quality of the final result. For example, leaves swaying in the wind may cause loss of detail, while moving people in the background may introduce artifacts. Be careful not to extend fingers/arms/legs into the frame while shooting.
After completing the model capture, click "Finish" on the top right corner.
3.3 Method Three: SCENECapture the model using your phone's camera to record a video. The steps are as follows:
3.3 Method Three: SCENE
Capture the model using your phone's camera to record a video. The steps are as follows:
1、Click the red start button in the middle at the bottom of the screen to shoot a video circling around the model.
2、After completing the shooting, click on the "Finish" button on the top right corner.
Take note of the following for video capture:
Video Setting: When using video capture, it is best to turn off the "HDR Video" option.
Exposure: t is recommended to use fixed exposure as much as possible.
Upload
After completing the above steps, you will be redirected to the upload page.
1、Click on the "Upload" button at the bottom of the screen, which will pop up a window to edit the model name.
2、Edit the model name and click "Upload" to start the upload process.
Rendering
After the upload is complete, wait for the rendering process (this step requires no action, just wait for the rendering to be completed).
Download
The steps are as follows:
Mobile APP:
1、After rendering is complete, pull up the details page from the bottom.
2、Select "Assets ---> OBJ ---> Low-Poly Diffuse".
3、A confirmation pop-up will appear, click on "Proceed" to download.
4、After the download is complete, select the location to store the compressed file.
PC：
1、Login and go to the homepage.
2、Click on the "Captures" button in the top right corner to navigate to the Fields Dashboard page.
3、On the Fields Dashboard page, select the model you want to download under Your Captures.
4、Hover over the model you want to download, and a download button will appear.
5、Click the download button, a page for selecting download parameters will appear.
6、Select "Object ---> OBJ ---> low poly" for the download.
Uploading Models Downloaded from Luma AI
Before uploading models downloaded from Luma AI, it's necessary to convert the type of the model:
1、Drag the entire model folder into your project folder.
2、Directly drag and drop the model into the scene.
3、Drag it down and click on “Original Prefab” to generate a prefab type model, which can then be uploaded as a prefab type model.
Last updated