Table of Contents

構造化 E57 点群で Mega Block を作成する

この節では、レーザースキャンデバイスで一般的な E57 形式データを Mega で使用する方法を説明します。

Overall workflow

EasyAR Mega supports building spatial maps from structured E57 data generated by third-party scanning devices.

An E57 file must provide both the point cloud for spatial reconstruction and images with corresponding poses. After the data is uploaded, the Mega cloud mapping service automatically processes it and generates a Mega Block that can be used for localization and AR content deployment.

This method is suitable for medium and large scenes such as shopping malls, museums, exhibition halls, campuses, factories, warehouses, and continuous street blocks. It is especially suitable for projects that have already completed professional laser scanning or require relatively complete 3D structure.

The overall workflow for building a spatial map with E57 data is:

  • Upload E57 data to the Mega cloud to generate a localization map.
  • Place AR content and test remotely. View the reconstructed scene in Unity or other 3D engines, design AR content, debug remotely, package the application, and so on.
  • Test and experience onsite.

E57 types supported by Mega

An E57 file must provide both the point cloud for spatial reconstruction and images with corresponding poses. E57 data that meets these structural requirements can come from many professional scanning devices or processing platforms. Common sources include Matterport, Leica, NavVis, XGRIDS, CHCNAV, and other devices and software that can export structured E57.

注記

E57 files that contain only point clouds but no panoramic images or camera poses cannot be used with Mega.

E57 input requirements

  • Structured data: scan station information should be retained in the E57, rather than only a merged unstructured point cloud.
  • Registered point cloud: all scan stations must be correctly stitched in the same coordinate system, without obvious ghosting, layer offsets, breaks, or drift.
  • Panoramic images: each valid scan station should contain extractable color panoramic images; RGB point colors alone cannot replace panoramas.
  • Camera poses: panoramas must have extrinsics or scan station poses consistent with the point cloud, and they must be correctly associated with the corresponding scan positions.
  • Metric scale: data must use real-world scale. Meters are recommended; if other units are used, make sure the scale metadata is accurate during export.
  • Coordinate consistency: point clouds, scan stations, and panoramas should use the same coordinate system. After export, do not rotate, scale, or translate only part of the data.

Acquisition quality recommendations

Mega localization depends on stable visual features in the scan station panoramas. Acquisition quality affects not only the 3D model, but also the user's onsite localization success rate. To improve localization, follow these recommendations:

  • Scan station spacing: generally keep it at 1-3 m. Use denser stations indoors, at corners, doorways, and places with many occlusions; open areas can be slightly looser.
  • Route coverage: cover all planned localization corridors, entrances, corners, stairways, elevator halls, and content deployment areas.
  • Image quality: keep lenses clean and exposure stable; avoid motion blur, strong backlight, very dark areas, and large reflective surfaces.
  • Scene state: collect data when there are fewer people and displays are stable, reducing crowds, vehicles, and temporary occlusions.
  • Low-texture areas: add scan stations around solid-color walls, glass curtain walls, repeated shelves, and similar areas, while also covering nearby reference objects with clear textures.

Why high-density scan stations are needed

  • They provide more viewpoints and visual overlap, improving the usable range and stability of onsite localization.
  • They reduce localization blind spots caused by door openings, corners, columns, and large exhibits.
  • They improve the spatial association between point clouds and images, making later content placement and real-world occlusion easier.
  • They provide more sufficient data redundancy for scene changes, different lighting, and different device camera viewpoints.

Next steps

上記の E57 ファイルを EasyAR サーバーにアップロードする必要があります。詳細は E57 データを Mega サーバーへアップロードする を参照してください。