Robots for 360 photographer to automate or semi automate capturing process15111
 WGAN ForumFounder & WGAN-TV Podcast Host Atlanta, Georgia |
DanSmigrod private msg quote post Address this user | |
Quote:Originally Posted by @Sdoughtie |
||
| Post 76 • IP flag post | ||
|
WGAN Forum Founder & WGAN-TV Podcast Host Atlanta, Georgia |
DanSmigrod private msg quote post Address this user | |
 |
||
| Post 77 • IP flag post | ||
|
WGAN Forum Founder & WGAN-TV Podcast Host Atlanta, Georgia |
DanSmigrod private msg quote post Address this user | |
 (Camera is on a robot) |
||
| Post 78 • IP flag post | ||
|
WGAN Forum Founder & WGAN-TV Podcast Host Atlanta, Georgia |
DanSmigrod private msg quote post Address this user | |
| Video: DJI Osmo Pocket 3 + Portable Electric Dolly | video courtesy of Roget Seng YouTube Channel | 1 January 2024 Portable Electric Dolly I wonder if a 360 photographer will make this work for their needs. Your thoughts? Dan |
||
| Post 79 • IP flag post | ||
|
WGAN Forum Founder & WGAN-TV Podcast Host Atlanta, Georgia |
DanSmigrod private msg quote post Address this user | |
 |
||
| Post 80 • IP flag post | ||
|
WGAN Forum Founder & WGAN-TV Podcast Host Atlanta, Georgia |
DanSmigrod private msg quote post Address this user | |
 |
||
| Post 81 • IP flag post | ||
|
WGAN Forum Founder & WGAN-TV Podcast Host Atlanta, Georgia |
DanSmigrod private msg quote post Address this user | |
|
||
| Post 82 • IP flag post | ||
|
WGAN Forum Founder & WGAN-TV Podcast Host Atlanta, Georgia |
DanSmigrod private msg quote post Address this user | |
| Video: Unitree Go2 New Upgrade | Video courtesy of Unitree Robotics YouTube Channel | 30 July 2024 | ||
| Post 83 • IP flag post | ||
|
WGAN Forum Founder & WGAN-TV Podcast Host Atlanta, Georgia |
DanSmigrod private msg quote post Address this user | |
| Video: HP SitePrint Autonomous Layout Robot in use in Melbourne. | Video courtesy of C.R. Kennedy Geospatial YouTube Channel | 23 August 2024 | ||
| Post 84 • IP flag post | ||
|
WGAN Forum Founder & WGAN-TV Podcast Host Atlanta, Georgia |
DanSmigrod private msg quote post Address this user | |
 |
||
| Post 85 • IP flag post | ||
|
WGAN Forum Founder & WGAN-TV Podcast Host Atlanta, Georgia |
DanSmigrod private msg quote post Address this user | |
| Video: Googles New Humanoid Robots Are Incredible - Gemini 2 Robotics | Video courtesy of TheAIGRID YouTube Channel | 13 March 2025 | ||
| Post 86 • IP flag post | ||
|
WGAN Forum Founder & WGAN-TV Podcast Host Atlanta, Georgia |
DanSmigrod private msg quote post Address this user | |
| Video: Tesla Bot Gen 3 Shocks With Its New Role 2025! 500 Tasks But Elon Musk Will Be Scared For NEW Rivals | Video courtesy of Tesla Car World YouTube Channel | 13 March 2025 | ||
| Post 87 • IP flag post | ||
|
Sdoughtie private msg quote post Address this user | |
| https://www.unitree.com/go2 From Grok... who know's if it's accurate: Can the Unitree Go2 Explore a Building? The Unitree Go2, particularly in its higher-end configurations like the Go2 Pro or Go2 Edu, is well-suited for autonomous building exploration. It’s a quadruped robot designed with agility and advanced sensing capabilities, making it capable of navigating indoor environments. Key features supporting this include: 4D LIDAR L1: Standard on all Go2 models (Air, Pro, Edu), this provides a 360°x90° hemispherical field of view with a detection range starting at 0.05 meters and extending up to 20 meters (at 90% reflectivity). This ultra-wide-angle LIDAR allows the Go2 to map its surroundings in real-time, detect obstacles, and navigate complex indoor spaces like rooms, hallways, or staircases. Mobility: With a top speed of 3.7 m/s (Pro/Edu models) and the ability to handle uneven terrain, the Go2 can traverse typical building features like steps or cluttered floors. Its adaptive joints and obstacle avoidance algorithms enhance its ability to move autonomously. AI and Autonomy: The Go2 integrates embodied AI, including large language model support (e.g., GPT empowerment) and optional NVIDIA Jetson Orin NX modules in the Edu version. This enables intelligent path planning and real-time decision-making, crucial for exploration without constant human input. App Integration: The Unitree Go app allows you to construct a point cloud map of an area and specify autonomous paths for the Go2 to follow, making it practical for exploring a building without pre-programmed routes. For example, the Go2 could roam a house or office, mapping walls, furniture, and openings while avoiding collisions, much like Boston Dynamics’ Spot but at a lower cost and with a consumer-friendly design. The Edu version, with its enhanced computing power and ROS (Robot Operating System) compatibility, is particularly suited for research-grade exploration tasks. Can the LIDAR Data Be Downloaded? Yes, the LIDAR data from the Unitree Go2 can be downloaded, though the process depends on your setup and technical expertise: Standard Models (Air/Pro): The 4D LIDAR L1 generates point cloud data that can be accessed via the Unitree Go app or through the robot’s onboard systems. The app supports real-time visualization of point clouds, and Unitree provides an open-source SLAM (Simultaneous Localization and Mapping) solution using the L1’s data (point cloud and IMU). While the app doesn’t natively export files in formats like .pcd (Point Cloud Data), you can connect to the robot wirelessly (WiFi 6 or 4G/5G in some configs) and extract raw data via SDK tools or custom scripts. Unitree’s website offers resources like the Unilidar Point Cloud Software and SDK for this purpose (see unitree.com). Edu Model: The Go2 Edu, with its ROS2 Humble support and optional 2D/3D LIDAR upgrades (beyond the standard L1), makes data extraction more straightforward. Users have reported success in visualizing LIDAR topics in ROS and converting them into downloadable files. For instance, a forum post on MYBOTSHOP from September 2024 details a user connecting a Go2 Edu to ROS2, accessing the LIDAR topic, and seeking to output point cloud data for use in tools like Blender. The process typically involves: Establishing a wireless connection to the robot. Subscribing to the LIDAR topic in ROS (e.g., /pointcloud). Using ROS tools like rosbag to record the data or a script to save it as a .pcd or .ply file. L2 Upgrade Option: As of November 2024, Unitree released the 4D LIDAR L2 ($419), an upgrade boasting 64,000 points/s and a 30-meter range. It includes downloadable sample data (e.g., indoor and park point clouds) and an open-source SLAM solution at unitree.com. If equipped with the L2, the Go2 could generate even denser maps, with similar data export methods. Practical Details How to Download: For non-Edu models, you’d need to use Unitree’s SDK (available on their site) or reverse-engineer the data stream via USB-C or network ports, as some enthusiasts have done (e.g., Reddit posts from 2023 document Go2 teardowns revealing RockChip RK3588 hardware). For Edu, ROS simplifies this—tutorials from DroneBlocks or Unitree’s GitHub (github.com/unitreerobotics/unitree_ros) guide you through saving LIDAR outputs. Formats: Common formats like .pcd or .ply are achievable with ROS or post-processing tools like CloudCompare or Blender, aligning with the user’s goal from the MYBOTSHOP forum. Storage: The Go2 doesn’t store large datasets onboard long-term; data is processed in real-time, so you’d need an external computer or cloud connection to save it. Cost Context MSRP: Go2 Air ($1,600), Go2 Pro ($2,800), Go2 Edu ($8,500+), per Unitree’s site and distributors like Airpuria. The L2 LIDAR adds $419. Exact Edu pricing varies with add-ons (e.g., Jetson module, extra battery). In short, the Go2 can explore a building with its LIDAR-driven autonomy, and you can download the data—most easily with the Edu model via ROS, or with some tinkering on Air/Pro models using Unitree’s tools. Want a step-by-step guide for a specific model? Let me know! |
||
| Post 88 • IP flag post | ||
|
WGAN Forum Founder & WGAN-TV Podcast Host Atlanta, Georgia |
DanSmigrod private msg quote post Address this user | |
| Video: New UPDATE! Elon Musk LEAKED Tesla Bot Gen 3 10K Mass Production & All Real-Life Tasks Testing! | Video courtesy of Tesla Car World YouTube Channel | 14 March 2025 | ||
| Post 89 • IP flag post | ||
 WGAN FanClub Member Queensland, Australia |
Wingman private msg quote post Address this user | |
Quote:Originally Posted by Sdoughtie There is a guy in official Matterport group who is testing it with a Matterport camera. Here is his post. I think this dog is too small and lightweight to work as he is expecting. Even if it can balance with a Pro3 sitting on a monopod it till just worn all gear much faster. It will be more interesting to put a slam lidar on it. But I worry it may be to low point to collect good data from a this height. I actually contacted them when they first started to advertise on Linked in 3-5 years ago with the sae idea. It was their first model and their engineer confirmed my concerns that it may not be able to carry a big rotating camera on a monopod. I think the lidar they use is very cheap and does not have enough density.. I actually have seen them selling the lidar for $2xx on their website about 2 months ago. May be the one you mention is more expensive even though it is the same 65000 ptps. Anyway it should be ok for navigation but not really good for collecting point cloud data. I still cannot let go that the one below(Snoopa rover) has never happened. It could carry anything as it is just tripod on wheels, could be mapped to move from point to point. It would be ideal if they actually could make it  |
||
| Post 90 • IP flag post | ||
|
WGAN Fan Club Member Queensland, Australia |
Wingman private msg quote post Address this user | |
Quote:Originally Posted by DanSmigrod I actually have not thought about using this dolly with my Pocket 3. However I have another interesting question. What if we put it under a tripod leg making one leg motorised and add just normal wheels to two other legs? |
||
| Post 91 • IP flag post | ||
|
|
Sdoughtie private msg quote post Address this user | |
Quote:Originally Posted by Wingman I was looking at the interior of the Pro3 and it uses a Livox sensor, that appears to have the same accuracy of the lidar sensor on the Go2 Pro Interior view of Pro3: https://www.wegetaroundnetwork.com/topic/20631/page/1/pro3-camera---3dgs-model---interior-view/#5 Livox Sensor: https://www.livoxtech.com/mid-40-and-mid-100 Livox on a dog robot: https://www.livoxtech.com/showcase/13 |
||
| Post 92 • IP flag post | ||
|
|
Sdoughtie private msg quote post Address this user | |
| Granted the MP Pro3’s accuracy is likely enhanced through software… | ||
| Post 93 • IP flag post | ||
|
WGAN Fan Club Member Queensland, Australia |
Wingman private msg quote post Address this user | |
Quote:Originally Posted by SdoughtieYes I can confirm that Pro3 models are using Livox Lidars, more likely mid-40 model. However Unitree 4D Lidars(L1 and L2) are not the same specs as mid-40. They are 25000 points/s for L1 model and 64000 points/s for L2. Livox mid-40 is 100000 points/s I do not believe L1 or L2 with such low density can produce the same 2cm accuracy as claimed. Quote: Originally Posted by Sdoughtie That looks like Livox mid-360 lidar model and it shoots 200000 points/s. The dog however does not look like unitree dog. Unitree offers B2 model and on its picture on their website it looks like it comes with quite expensive Oyster Lidar or similar. It says it is 32 wire(I think they meant lines) but does not mention its speed/density. However B2 is $100 000 |
||
| Post 94 • IP flag post | ||
|
|
Sdoughtie private msg quote post Address this user | |
Quote:Originally Posted by Wingman LOL, oops, I missed some 0's there. |
||
| Post 95 • IP flag post | ||