Camera Array Tool for Blender

This is a Blender add-on that creates Camera Arrays on object surfaces.

You can build virtual Camera Array studios from spheres, cylinders, grid surfaces around your 3D object and render multiple cameras out as image sequences. Play and design your own Camera array setup inside Blender.

Image series can be used to create Gaussian Splatting models, for example.

NEW UPDATE INCLUDES OPTION TO GENERATE 4DGS DATA FOR SYNTHETIC ANIMATED GAUSSIAN SPLATTING TRAINING!!

Check the latest tutorials in here:

Basic Features:

• Create primitives for camera rigs or choose from Pre made ones
• Five pre made Array objects
• Choose camera direction (point inward or outward or both)
• Move, Scale and Edit your object and the cameras will follow the surface
• Change Focal Length of all cameras in Array at once.
• Render images from each cameras to the folder you want

Export COLMAP Features:

• Generate cameras.txt, images.txt and point3D.txt files
• These files helps you to skip Structure from Motion phase in Postshot and you can start directly the Gaussian Splatting training
• Optimize point cloud file size with a density setting
• Export point clouds with Color values. Colored points will speed up the Gaussian Splatting process.

Generate Animated 4DGS data:

• Choose animated object
• Render animation frames from each Array Cameras
• Include colmap data for your animated object

Basic idea is that this tool generates the camera at the center of each polygon of the selected 3D mesh. You can decide which direction of the surface normal camera will point. Add-on is very simple to use and you are able to modify your Array structure simply by just editing the original object from where the Camera Array was made.

Modify Array object and change Focal Lengths for all cameras.

Rendering. You are able to export images out from each cameras of your array.

Render 4DGS animation frames

Point Cloud file size optimization with Density slider

Colored point cloud affects to the speed how fast Gaussian Splatting model will be trained in Postshot

Applications:

One idea where the image sequences produced by Camera Array can be used is 3D Gaussian Splatting training. So this tool allows you to convert your 3D model to 3DGS format through image sequences.

Image sequences can be trained in Gaussian Splatting format in separate applications such as:

• Postshot (https://www.jawset.com/)
• NeRF Studio (https://docs.nerf.studio/)
• Brush for Mac and Linux (https://github.com/ArthurBrussee/brush)
• Free Postshot Batch Trainer v2.0 (https://bit.ly/m/ToppiNappiCodes)

Here you can Download a 4DGS Batman run loop PLY sequence and Gaussian Splatting PLY file of a Furby conversion :

Older tutorials