HJCheng0602/SLAM3R_onlinecam

Enhanced SLAM3R: Real-Time Reconstruction via Online Camera Stream

38

stars

63

commits

Python

primary language

Dec 4, 2025

updated

README

Enhanced SLAM3R: Real-Time Reconstruction via Online Camera Stream

🌟 Project Overview

This is an enhanced and improved version of the well-known SLAM3R framework, designed to significantly expand its input versatility. The core advancement lies in enabling the use of online camera feeds, empowering users to perform remote and real-time 3D reconstruction using readily available mobile device cameras, thus overcoming the limitations of static local video files.

The overall functionality remains consistent with the original SLAM3R GitHub repository.


✨ Key Added Feature

The single, crucial feature introduced in this improved version is:

  • 🌐 Online Camera Support of gradio demo: The system now supports using an IP Camera (Internet Protocol Camera) stream as the input source for 3D reconstruction.

How to Use the New Feature


Demo

  • first run python app.py --online to open the gradio interface.

To leverage your mobile device camera for reconstruction from any location, follow these steps within the Gradio interface:

  1. Locate the Input Type Menu: Navigate to the Type dropdown menu within the Gradio interface.

  2. Select "web camera": You will find an additional option: "web camera". Select this option.

  3. Provide Connection Details: Once selected, you must input the necessary credentials to connect to your mobile device's camera stream:

    Configuration ItemDescription
    IP Camera Sharing URLEnter the streaming URL (e.g., rtsp://...) provided by your IP camera application.
    PasswordEnter the required password for accessing the shared IP camera stream.
  4. Initiate Reconstruction: After entering the details, you can begin the 3D reconstruction process using the live camera feed.


Script

look up the hint in the scripts/demo_wild.sh:

######################################################################################
# Your input can be a directory of images, a video file, and a webcamera url.
# For directory of images, set the TEST_DATASET to path like: data/test/myimages
# For video, set the TEST_DATASET to path like: data/test/myvideo.mp4
# For webcamera, set the TEST_DATASET to url like: http://rab:12345678@192.168.137.83:8081
######################################################################################

then you can try

bash scripts/demo_wild.sh

Now you can start reconstruction now !


⚠️ Performance Note on IP Camera Usage

It is crucial to understand that the practical usability of the IP camera stream may be significantly affected by:

  • Latency (Delay): High network latency can introduce inconsistencies and instability, potentially impacting the quality and real-time performance of the reconstruction.

πŸ“š IP Camera Information

For detailed instructions and information on how to set up, share, and obtain the necessary URL and password for your mobile device as an IP camera, please refer to this resource:

  • πŸ”— IP Camera Bridge Reference: ipcamera

Contributors

HJCheng0602

39 commits

Ly-kc

18 commits

siyandong

6 commits

HJCheng0602/SLAM3R_onlinecam

Enhanced SLAM3R: Real-Time Reconstruction via Online Camera Stream

38

stars

63

commits

Python

primary language

Dec 4, 2025

updated

README

Enhanced SLAM3R: Real-Time Reconstruction via Online Camera Stream

🌟 Project Overview

This is an enhanced and improved version of the well-known SLAM3R framework, designed to significantly expand its input versatility. The core advancement lies in enabling the use of online camera feeds, empowering users to perform remote and real-time 3D reconstruction using readily available mobile device cameras, thus overcoming the limitations of static local video files.

The overall functionality remains consistent with the original SLAM3R GitHub repository.


✨ Key Added Feature

The single, crucial feature introduced in this improved version is:

  • 🌐 Online Camera Support of gradio demo: The system now supports using an IP Camera (Internet Protocol Camera) stream as the input source for 3D reconstruction.

How to Use the New Feature


Demo

  • first run python app.py --online to open the gradio interface.

To leverage your mobile device camera for reconstruction from any location, follow these steps within the Gradio interface:

  1. Locate the Input Type Menu: Navigate to the Type dropdown menu within the Gradio interface.

  2. Select "web camera": You will find an additional option: "web camera". Select this option.

  3. Provide Connection Details: Once selected, you must input the necessary credentials to connect to your mobile device's camera stream:

    Configuration ItemDescription
    IP Camera Sharing URLEnter the streaming URL (e.g., rtsp://...) provided by your IP camera application.
    PasswordEnter the required password for accessing the shared IP camera stream.
  4. Initiate Reconstruction: After entering the details, you can begin the 3D reconstruction process using the live camera feed.


Script

look up the hint in the scripts/demo_wild.sh:

######################################################################################
# Your input can be a directory of images, a video file, and a webcamera url.
# For directory of images, set the TEST_DATASET to path like: data/test/myimages
# For video, set the TEST_DATASET to path like: data/test/myvideo.mp4
# For webcamera, set the TEST_DATASET to url like: http://rab:12345678@192.168.137.83:8081
######################################################################################

then you can try

bash scripts/demo_wild.sh

Now you can start reconstruction now !


⚠️ Performance Note on IP Camera Usage

It is crucial to understand that the practical usability of the IP camera stream may be significantly affected by:

  • Latency (Delay): High network latency can introduce inconsistencies and instability, potentially impacting the quality and real-time performance of the reconstruction.

πŸ“š IP Camera Information

For detailed instructions and information on how to set up, share, and obtain the necessary URL and password for your mobile device as an IP camera, please refer to this resource:

  • πŸ”— IP Camera Bridge Reference: ipcamera

Contributors

HJCheng0602

39 commits

Ly-kc

18 commits

siyandong

6 commits

Languages

Python

96.8%

Shell

1.9%