AI Offset Calibration

Use AI to refine sensor offsets by comparing sensor readings with a known point cloud.

Calibration Inputs

Provide data for the AI to analyze.

Known Point Cloud

Ground truth data (PLY, XYZ).

Sensor Readings

Data from sensors (PLY, XYZ).

Create a "Ground Truth" Scan: Using a high-precision 3D scanner or a known 3D model, create a clean point cloud of a simple reference scene (e.g., the corner of a room). Save this as a .ply or .xyz file. This is your "perfect" reference.
Capture Real Sensor Data: With your ESP32 rig, scan the *exact same scene* from the same position. Go to the "Export" page and download the point cloud data, which will contain the raw, uncalibrated readings.
Upload Both Files: Upload the "ground truth" file to the "Known Point Cloud" field and your exported sensor data to the "Sensor Readings" field.
Provide Initial Guesses: Input your best estimate for the physical offsets of your sensors relative to a central point on your rig.
Calibrate: Click the button to send all this data to the AI. It will compare your messy data to the perfect data and calculate the precise offsets needed to align them.

Calibration Results

The AI's suggested adjustments will appear here.

Submit the form to begin calibration.