Skip to content

README.md

Latest commit

 

History

History
159 lines (123 loc) · 7.23 KB

README.md

File metadata and controls

159 lines (123 loc) · 7.23 KB

Potree Core 2.0

npm version GitHub version

  • This project was originally based on Potree Viewer 1.6 and is now since version 2.0 based on the shiukaheng fork of the Potree-Loader.
  • Potree is a web based pouint cloud visualizer project created by Markus Schütz.
  • This project contains only the main parts of the potree project adapted to be more easily used as a independent library, the code was adapted from the original repositorys.
  • Support for pointclouds from LAS, LAZ, Binary files.
  • Some features require support for the following GL extensions
    • EXT_frag_depth, WEBGL_depth_texture, OES_vertex_array_object

Demo

Example

  • The project can be build running the commands npm install and npm run build.
  • Download the potree build from the build folder or add it to your project using NPM.
  • Include it alonside the worker folder in your project (can be found on the source folder).
  • The build is a ES module, that can be imported to other projects, threejs should be available as a peer dependency.
  • Bellow its a fully functional example of how to use this wrapper to load potree point clouds to a three.js project
const scene = new Scene();
const camera = new PerspectiveCamera(60, 1, 0.1, 10000);

const canvas = document.getElementById("canvas");

const renderer = new WebGLRenderer({canvas:canvas});

const geometry = new BoxGeometry(1, 1, 1);
const material = new MeshBasicMaterial({color: 0x00ff00});
const cube = new Mesh(geometry, material);
scene.add(cube);

const controls = new OrbitControls(camera, canvas);
camera.position.z = 10;

const pointClouds = [];

const baseUrl = "data/test/";
const potree = new Potree();
potree.loadPointCloud("cloud.js", url => `${baseUrl}${url}`,).then(function(pco) {
   scene.add(pco);
	pointClouds.push(pco);
});

function loop()
{
   potree.updatePointClouds(pointClouds, camera, renderer);

	controls.update();
	renderer.render(scene, camera);

	requestAnimationFrame(loop);
};
loop();

Clip Boxes

  • Clip boxes restrict the rendered region of a point cloud to a box-shaped volume.
  • Use the createClipBox(size, position) helper to build an IClipBox from a size and world-space position.
  • Set the desired ClipMode on the material and pass the clip boxes to setClipBoxes().
import { ClipMode, createClipBox } from 'potree-core';
import { Vector3 } from 'three';

// Create a 5×5×5 clip box centered at world position (2, 0, 0)
const clipBox = createClipBox(new Vector3(5, 5, 5), new Vector3(2, 0, 0));

// Highlight points inside the box (other modes: CLIP_OUTSIDE, CLIP_INSIDE, DISABLED)
pco.material.clipMode = ClipMode.HIGHLIGHT_INSIDE;
pco.material.setClipBoxes([clipBox]);
  • ClipMode.DISABLED – no clipping.
  • ClipMode.CLIP_OUTSIDE – only points inside the box are rendered.
  • ClipMode.CLIP_INSIDE – only points outside the box are rendered.
  • ClipMode.HIGHLIGHT_INSIDE – all points rendered; points inside the box are highlighted.

Clip Spheres

  • Clip spheres restrict the rendered region of a point cloud to a sphere-shaped volume.
  • Use the createClipSphere(center, radius) helper to build an IClipSphere from a center position and radius.
  • Set the desired ClipMode on the material and pass the clip spheres to setClipSpheres().
  • Clip boxes and clip spheres can be used together; a point is considered "inside" if it falls inside any box or any sphere.
import { ClipMode, createClipSphere } from 'potree-core';
import { Vector3 } from 'three';

// Create a sphere of radius 3 centered at world position (0, 1, 0)
const clipSphere = createClipSphere(new Vector3(0, 1, 0), 3);

// Highlight points inside the sphere (other modes: CLIP_OUTSIDE, CLIP_INSIDE, DISABLED)
pco.material.clipMode = ClipMode.HIGHLIGHT_INSIDE;
pco.material.setClipSpheres([clipSphere]);
  • ClipMode.DISABLED – no clipping.
  • ClipMode.CLIP_OUTSIDE – only points inside the sphere are rendered.
  • ClipMode.CLIP_INSIDE – only points outside the sphere are rendered.
  • ClipMode.HIGHLIGHT_INSIDE – all points rendered; points inside the sphere are highlighted.

Custom Request Manager

  • The potree core library uses a custom request manager to handle the loading of point cloud data.
  • The request manager can be replaced by a custom implementation, for example to use a custom caching system or to handle requests in a different way.
class CustomRequestManager implements RequestManager 
{
   fetch(input: RequestInfo | URL, init?: RequestInit): Promise<Response> {
      throw new Error("Method not implemented.")
   }

   async getUrl(url: string): Promise<string> {
      return url;
   }
}

Notes

  • Since potree-core is meant to be used as library and not as a full software as potree some features are not available.
  • EDL shading is not supported by potree core.
  • Removed Arena 4D point cloud support.
  • Removed Entwine Point Tile file support.
  • GUI elements were removed from the library
    • PotreeViewer
    • Controls, Input, GUI, Tools
    • Anotations, Actions, ProfileRequest
    • Potree.startQuery, Potree.endQuery and Potree.resolveQueries
    • Potree.timerQueries
    • Potree.MOUSE, Potree.CameraMode
    • PotreeRenderer, RepRenderer, Potree.Renderer
      • JQuery, TWEEN and Proj4 dependencies

Potree Converter

  • Use the (Potree Converter)[https://github.com/potree/PotreeConverter/releases] tool to create point cloud data from LAS, ZLAS or BIN point cloud files
  • Potree Converter 1.8 creates a multi file structure with each node as an individual file.
  • Potree Converter 2.1 creates a single file for all points and separates files for hierarchy index, its faster to create files. Requires a HTTP server configured for file streaming.
  • Tool to create hierarquical structure used for point-cloud rendering using potree-core.
  • There are two main versions 2.1 witch generates 4 contained files with point data, hierarchy,
  • To generate a folder output from a input file run the command .\PotreeConverter '..\input.laz' -o ../output

TXT2LAS

  • The potree converter tool only supports las and laz files, so textural file formats such as .pts, .xyz, have to be first converted into a supported format.
  • The TXT2LAS tool from the (LASTools)[https://github.com/LAStools/LAStools] repository can be used for this effect.
  • To run the tool use the command .\txt2las64 -i input.pts -ipts -parse xyziRGB -set_scale 0.001 0.001 0.001 -set_version 1.4 -o output.laz

To Do

  • Supports logarithmic depth buffer (just by enabling it on the threejs renderer), useful for large scale visualization.
  • Point clouds are automatically updated, frustum culling is used to avoid unnecessary updates (better update performance for multiple point clouds).