Mega space solution
For tourist attractions, urban blocks, large commercial complexes, and other scenarios covering hundreds of thousands of square meters or even millions of square meters, the traditional "single capture, single large map" approach is insufficient. This document introduces how to handle the capture and usage of mega space scenarios, addressing issues such as excessive cumulative errors and memory overflow caused by a single oversized map.
Core challenges
When dealing with millions of square meters of ultra-large spaces, attempting to build a single giant map faces many technical obstacles:
- Data collection difficulty: A single collection session that takes too long can cause device overheating and excessively large data volumes (hundreds of GB), making upload and processing extremely difficult.
- Accumulated error: Mapping errors accumulate with collection distance. Over paths spanning several kilometers, even a 1% error can result in deviations of tens of meters, which are hard to correct.
- Performance bottleneck: Loading and optimizing map data covering millions of square meters can instantly exhaust device memory.
Solution: Chunking and fusion
The standard solution to the ultra-large space problem is divide and conquer, logically dividing the entire large scene into multiple sub-regions that physically overlap with each other, collecting each sub-region separately, and finally fusing and optimizing all sub-regions. This enables seamless roaming through dynamic loading at runtime.
Data acquisition
You need to plan according to the large-space data acquisition method.
- Segmentation principle: Divide the large scene by area, scene, or functional zone, with each zone covering approximately
100,000 square meters. - Overlap requirement: Adjacent sub-zones must have
a minimum of 200 meters of fully overlapping paths. The larger the overlap area, the better the fused map. Ensure sufficient overlap to optimize the fusion of adjacent sub-maps.
Before starting data acquisition for each sub-zone:
- Perform acquisition route planning
- Check GoPro Max device configuration
- Read and familiarize yourself with the acquisition method
- Understand the steps for exporting GoPro Max data
Create a mapping task
- If you are using Mega mapping for the first time, you need to create a new mapping project first.
- In the Mega development center, you can create a large-scale multi-map fusion task. Upload multiple segmented videos, and the cloud mapping algorithm will automatically stitch and optimize them.
View the mapping results
After the mapping is completed, you can view the mapping results:
Tip
- If the mapping fails, refer to the Mapping failure troubleshooting guide for resolution.
- If issues like layering occur during mapping, refer to the Model defect troubleshooting guide for resolution.
Test positioning effect
- If you are using Mega positioning for the first time, you need to configure the positioning database
- Quickly verify if the positioning database is available
- View simulation run effect
- View real-time on-site run effect
Run strategy
The fused map can be used directly as a single map without any additional processing.
- GNSS-based query: In outdoor scenarios, GNSS is used to assist in querying the map, enabling precise positioning even in large maps.
- On-demand loading: When placing content, the mesh of each sub-area can be loaded individually.
- Seamless switching:
- The positioning algorithm runs simultaneously in sub-map A and sub-map B.
- In overlapping areas, content will also switch seamlessly.