PLANIX and R1 SLAM Explained
SLAM Explanation
Last updated
Was this helpful?
SLAM Explanation
Last updated
Was this helpful?
Introduction to SLAM (Simultaneous Localization and Mapping)
The new SLAM feature helps users understand the camera's position in relation to previous scans, giving a sense of how well future scans will match up. This is achieved by tracking the camera’s position in real-time using 2D depth measurements (captured by a time-of-flight lidar sensor), which builds a simplified map of the area. This data helps the system calculate how well a new scan will align with what has already been scanned.
The SLAM system uses live depth data from the camera’s 2D lidar sensor along with previously captured scans to locate the camera on a floor plan. Tracking begins once at least one scan has been completed, and the camera’s position will only display on the screen if the “show live lidar data” option is enabled in the app (See Image 1 below).
If the camera moves too far from the locations of previous scans, tracking may struggle and stop. When this happens, the system tries several times to re-establish the position. If these attempts don’t work, the tracking feature will turn off, and the camera’s position will stop updating on the app. Moving the camera closer to its last known position can help it recover. Starting a new scan or moving to a different floor can also restart tracking. Keeping the camera level and scanning frequently will improve tracking quality.
When tracking is active, and live lidar data is displayed, the app offers indicators to show the tracking status and the alignment quality of a new scan at the current camera position. The color of the 2D lidar point cloud and the camera position marker changes from green to red, indicating how well a new scan will align. Green means alignment is good, yellow or brown suggests alignment may be off, and red means tracking is struggling to maintain the position. If the center marker appears red without the point cloud, tracking has turned off automatically. When the point cloud appears brown, moving farther from previous scan positions may cause the camera tracker to lose its location (See Image 2 below).
These visual cues help users understand the alignment quality and make adjustments if needed.
Coverage refers to the number of scans required to capture an entire property and the details captured with each scan. Generally speaking, better coverage makes for a better 3D tour. Additionally, coverage influences floor plan accuracy, as rooms that have not been scanned will not be included in the floor plan.
The number of scans affects the time required to scan a property. Knowledge of ideal camera placement is required to reduce time spent scanning but still capture the data needed for a professional, accurate output. Ideal camera placement is based on line-of-sight and distance from the previous scan. SLAM helps the Operator determine placement by showing them where the camera rests in real-time. Below are examples of good coverage (Image 3), and poor coverage (Image 4).
Data alignment greatly benefits from SLAM because it helps the Operator create scans in optimal locations for automatic data alignment. Below are examples of Manual Alignment (Image 5) vs. Automatic Alignment (Image 6).
