Mega basic concepts

Mega involves multiple key concepts and technical terms, which are core elements for building and using city-scale AR localization services. This document explains the primary concepts in non-technical language, including EMP, Mega Block, EIF, and EMA. Understanding these concepts helps you better plan projects, collect data, verify results, and optimize AR experiences.

EMP (mapping data)

EMP is the file format recorded when using a mobile phone to collect mapping data, containing video frames, IMU (inertial measurement unit) data, GNSS information, SLAM poses, etc. This data can be used to generate the corresponding mapping result—the Mega Block.

Function

• Rapid mapping: Suitable for small spaces like rooms or exhibition halls, or specific objects, requiring no professional equipment.
• Complementary to panoramic camera data: EMP offers greater flexibility.

How to obtain and use

• Obtain: Use the Mega Toolbox App to record by slowly moving the phone in the target area (see Using mobile phones to collect small spaces - collection methods or Using mobile phones to collect specific objects - collection methods), then export the EMP file.
• Use: Upload to the console to create small-space mobile collection tasks or object mobile collection tasks, generating a Block.

Notes

• EMP files are small in size, suitable for rapid iteration but not ideal for very large spaces.
• Collection essentials: Move slowly, scan from multiple angles, avoid blur or dynamic interference; refer to Collection checklist.
• If EMP results in low Block quality, check phone compatibility (requires ARCore/ARKit support).

Mega Block

Mega Block is the core localization map unit of Mega. Generated through mapping services, it includes a 3D model and localization map, containing geometric structures, visual features, and coordinate information of the target area, similar to a "digital twin map."

Features

• Provides high-precision localization: Devices achieve centimeter-level world coordinate anchoring by matching current camera images with features in the Block.
• Supports AR fusion: Including virtual-real occlusion and persistent content.
• Coverage: A single Block can cover millions of square meters, supporting multiple floors or large spaces.

How to obtain and use

• Obtain: Upload via data collection to the console, then create a mapping task to generate automatically.
• Use: Create localization services, localization libraries, add Blocks to localization libraries, then configure in applications.

Notes

• Block quality depends on collected data: Environments with rich features and uniform lighting typically yield the best results.
• For mapping issue troubleshooting, refer to Mapping failures or Model defects.

EIF (simulation data)

EIF is a data format used for simulated localization testing, containing video frames, IMU (inertial measurement unit) data, GNSS information, SLAM poses, etc., serving as input for Mega services.

Function

• Simulation verification: Preview localization and tracking effects on a computer by replaying EIF data without field testing.
• Issue reproduction: Record problematic segments to help debug tracking issues like drift or loss.
• Effect preview: Combine with annotation files to simulate AR content overlay and evaluate end-user experience.

How to obtain and use

• Obtain: Record with Mega Toolbox in the target area (see Simulating operational effect preview - collecting EIF data or Visual anomaly data collection - using Toolbox to record mobile EIF files), then export the EIF file.
• Use: Import into Mega Studio to replay and view localization performance; or upload for issue reporting.

Notes

• Recorded EIF duration should not exceed 10 minutes; longer collection times may impact device SLAM performance.
• For testing only, not for formal mapping.

EMA (annotation file)

EMA is the format for annotation data files, used to define the position, orientation, and behavior of AR content within a Mega Block. It is a JSON-like structured file containing annotation information such as coordinate points, 3D model references, and trigger conditions.

Function

• Content anchoring: Specifies where virtual objects (e.g., models, text, animations) should appear in the real world with precise positioning.
• Interaction definition: Sets trigger rules (e.g., display when user approaches), animation logic, sound, etc.
• Effect preview: Simulates complete AR experiences when combined with EIF or field testing.

Notes

• EMA is optional: Basic localization doesn't require EMA, but rich interactions do.
• If EMA causes content misalignment, check coordinate system consistency (must match the Block).

Understanding these concepts enables more efficient use of Mega services. For deeper details, refer to the Overview or specific sections.