Month: May 2026

First I need to create Gpu texture for lidar file.

Then, go to Niagara panel, change the parameter

What’s more, I linked each lidar and the gpu texture in Niagara, and put it into sequencer to get the effect

In the UE5 scene, we designed a vertical space in which the rich area and the poor area are positioned in opposition—one above and one below. When the viewer looks from the perspective of the rich, the poor area lies beneath their feet; when viewing from the perspective of the poor, the rich area hovers above their head. These two perspectives face each other, and the viewer needs to freely rotate the camera to observe both areas up and down. This physical “head-turning” motion is intended to let the viewer intuitively feel the sense of “distance” between wealth and poverty in physical space—the brightness and openness of the rich area become more apparent in contrast to the cramped and dim atmosphere of the poor area when switching viewpoints. To further reinforce this visual contrast, we adjusted the size proportions and lighting intensity of the two areas: the rich area occupies a larger proportion and is lit brighter (simulating large windows and ample sunlight), while the poor area is slightly smaller with dimmer lighting (simulating narrow building spacing and limited daylight). Our hope is that this design allows viewers not just to “see” the wealth gap, but to “feel” its presence in space.

The building models we collected needed to be converted into point cloud data before they could be presented as particles in UE5. To save modeling time, we first downloaded ready-made architectural models from Sketchfab that matched the style of London neighborhoods, and then used point cloud processing software such as CloudCompare to export them as .ply or .xyz files. During the conversion process, we encountered some difficulties: the high texture repetition on some models confused the point cloud generation algorithm, leading to a large number of overlapping noise points. We had to manually inspect the point clouds frame by frame, selecting and deleting outliers that deviated from the main structures—a time-consuming step. After importing the processed point clouds into UE5, we assigned different point density and color settings to the rich and poor areas—the rich area had denser points (symbolizing concentrated resources) with warmer tones (gold, beige), while the poor area had sparser points with cooler tones (grey, blue). This visual distinction helps viewers perceive the economic differences between the two areas at a glance.

I tried a plugin that is able to convert lidar into Niagara particle, which allows me to have movement of the point cloud.

Here is a documentation of the plugin:

I discovered the workflow of how to convert a fbx model into a LiDAR point cloud in UE5

1. Embed texture on model: prepare fbx/obj model, drag it into maya/blender, link the texture image, export the fbx with “embed media” selected.
2. Convert the format into LAS in Cloud Compare: drag the fbx into Cloud Compare, edit>mesh>sample cloud, then select the point cloud, go to file> save, to export the LAS format.( you may need to go edit>apply transformation>Euler Angle to change the x-axis rotation 90 degrees before export, because the axis mode is different in UE5)

3. Import the lidar point cloud in UE5: Go to plugin> tick lidar, restart, then drag the LAS format file in content drawer.