Pocket 3D: A Turnkey Solution and an Open Development Kit

Built for Both 3D Modeling and Development

One Device, Multiple Application Paths

As a new-generation lightweight 3D modeling system, Pocket 3D enables rapid point-cloud acquisition and large-scale 3DGS mapping. It produces four types of modeling outputs while significantly improving reconstruction efficiency in complex environments.

Its capabilities extend far beyond standard modeling.

Unlike conventional 3D modeling systems, Pocket 3D was designed with an extensible architecture that supports the export of raw imagery, pose data, and other source data. This design also makes Pocket 3D a professional development kit. Universities, developers, software vendors, and enterprise R&D teams can use its open data interfaces for advanced secondary development and integrated 3D vision applications. Its flexible and scalable architecture provides greater development potential than conventional systems and supports broader industrial integration and deployment.

As a Professional 3D Modeling System

Efficient 3D Modeling Through a Simplified Workflow

This workflow is designed for users who need simple operation and fast results:

Capture: Move around the target scene with the handheld device to perform a complete scan. Before operation, place the device on a level surface for approximately 20 seconds to complete automatic calibration, then begin data acquisition.

Processing: Import the captured data into the accompanying software and run one-click post-processing to generate colored point clouds and 3DGS reconstructions. The workflow is optimized for real-world use, removing unnecessary parameter settings and reducing operational complexity.

Outputs: Point-cloud models support accurate measurement and analysis. 3DGS models support free-view navigation, preset-path tours, and combined visualization with point clouds and panoramic imagery.

Pocket 3D also reduces long-term ownership costs. All supporting software for mapping, model training, and result viewing includes a perpetual free license and free lifetime updates, with no annual subscription fees or additional charges for future features. A one-time device purchase provides continued access to the complete software suite, significantly reducing long-term operating costs.

As a Development Kit

Open Access to Raw Pose and Image Data

A core step in the Pocket 3D workflow is to build a scene map from LiDAR point clouds and calculate the panoramic camera pose at each capture position.

For turnkey users, this is simply an intermediate background process. For developers, however, its outputs — a set of panoramic images with absolute poses and a dense point cloud with real-world scale — are highly valuable.

By comparison, vision-only open-source solutions typically rely on COLMAP and offline Structure-from-Motion calculation to estimate camera poses. This process is time-consuming and prone to failure in low-texture, low-light, or large-scale environments. Pocket 3D uses LiDAR sensors to estimate poses directly, delivering reliable results with inherent physical scale while eliminating the time-consuming SfM stage that often interrupts conventional workflows.

Once the imagery and pose data are available, developers can bypass this critical bottleneck and proceed directly to downstream applications or algorithm development.

How Can the Data Be Used?

Third-Party 3DGS Workflows: Import pose-tagged panoramic images directly into your preferred Gaussian Splatting training framework and process them using a third-party 3DGS pipeline.

Mesh Reconstruction Software: Import point-cloud and image data into third-party modeling tools to generate triangular mesh models for visualization, editing, and rendering.

Proprietary Algorithm Development: For developers working on SLAM, 3D reconstruction, or pose estimation, the dataset provides multimodal sensor data—including LiDAR, imagery, and pose data—with real-scale ground truth for algorithm comparison, error analysis, and validation.

Teaching and Laboratory Training: In universities and vocational education, a system that captures real-world data and exports intermediate results provides a clearer demonstration of SLAM and 3D reconstruction principles than simulation-only datasets.

The data supports different user requirements. General users can follow the standardized 3D modeling workflow, developers can extract pose data for independent training and optimization, and evaluation organizations can conduct performance testing and secondary development.

What We Offer Software Partners?

For commercial software developers working in 3D reconstruction, digital twins, or industry application platforms, Pocket 3D provides open and flexible integration capabilities.

Our supports technical integration through open data formats, allowing imagery, pose data, and point clouds captured by Pocket 3D to be incorporated into your products. Cooperation is also available for SDK access, embedded product integration, and other development models. Interested teams are welcome to contact us.