Long Radius Calibration (Auto)

## Long Radius Calibration (Auto)

Same as the Camera Calibration module, it calculates the correspondence between the image coordinate system and the mechanical coordinate system, generating a calibration for picking or alignment projects.
## Module Principle
By selecting the movement coordinates of the robot arm and inputting the image coordinates corresponding to the robot arm movement, the calibration relationship between the image pixel coordinates and the real-world coordinates is established and saved in the global variables.
![](/media/202601/2026-01-15_105502_1508570.3570884028101299.png)
The Camera Calibration module generates mechanical point data through the selected mechanical model, and the image point data is acquired externally (usually by acquiring feature points through Optimized Feature Matching).
The mechanical center pose is the center coordinate of the machine during calibration (the fifth point or the second point). The calibrated mechanical point is calculated and output through this center coordinate and the mechanical type. Therefore, after modifying the mechanical type, you need to click [Mechanical Parameters] to regenerate the mechanical point data once.
Calibration only has XYR and XR, YR types with angles. The actual rotation count is twice the [Rotation Count]. First rotate around the mechanical rotation center, then rotate around the image matching point as the rotation center.
The photographing method refers to the relationship between the robot arm and the camera during calibration.
<Eye-in-Hand> The camera moves with the robot arm, and the movement direction of the acquired image points will be opposite to the actual robot arm running direction.
<Eye-to-Hand> The camera is fixed and the robot arm moves, and the movement direction of the acquired image points will be the same as the actual robot arm running direction.
## Usage Scenarios
Performed before picking or alignment calculation. Visual calibration is the process of determining parameters such as camera intrinsics, extrinsics, and lens distortion through image data collected by the camera. The purpose is to establish a precise mapping relationship between image pixel coordinates and actual physical coordinates.

## Usage Method

### Pre/Post-Sequence Modules
Camera image acquisition, Optimized Feature Matching to acquire image point data, run multiple times to generate calibration. The output of Camera Calibration is the mechanical point that needs to be reached next time.

![](/media/202603/2026-03-12_152539_5688130.3782010936453578.png)
## Parameter Configuration Description
【Mechanical Parameters】 Set the coordinates of mechanical movement, and obtain image coordinates through input images. 
Note: In calibration mode, the first picture taken is not included in the calculation; in simulation mode, the first photo will be used as the first point of the image points.
【Calibration Group】 Select the corresponding coordinate system group, and the calibration data will be recorded in the calibration file under this coordinate system group. The coordinate system group must be set in the Coordinate System Calibration in Global Data Management.
【Calibration File】 Select the calibration file under the corresponding coordinate system group, and the calibration data will be recorded in this file. (The calibration file must be set in the Coordinate System Calibration in Global Data Management. First add a coordinate system group, then drag the camera and robot to the window under the corresponding coordinate system group, select the connection line to connect the two icons, and the calibration file can be generated) 
Running Mode:
【Simulation Model】 Generally used when calibrating with a group of pictures. In this mode, the first photo will be used as the first point of the image points. Also, images will not be saved in simulation mode. 
【Calibration Mode】 Calibrate according to actual conditions. (Taking 12-point calibration as an example, when the reference image selects a simulation image, 13 images need to be provided, and the first image is not actually calculated. Simulation mode does not need to provide an extra image) 
【Save Image】 Save the reference image.
Mechanical Model:
【XYR】 Mechanical structure XY axes can translate and rotate. 
【XR (Y Separated)】 Mechanical structure only has X-axis translation and can rotate, Y-axis structure is separated. 
【YR (X Separated)】 Mechanical structure only has Y-axis translation and can rotate, X-axis structure is separated. 
Photographing Method: 
【Eye-to-Hand】 The camera is installed in a fixed position and does not shift or rotate with the machine. 
【Eye-in-Hand】 The camera is fixedly installed on the mechanical rotation axis and shifts and rotates with the machine. 
#### Mechanical Parameters 
Mechanical Movement Settings:
【Translation Row (Column) Count】 Set the number of rows/columns for calibration mechanical movement. 
【Rotation Count】 Set the number of rotations for calibration mechanical movement. 
【Rotation Radius】 The physical distance from the rotation center to the feature matching point. Just input an approximate value to improve calculation speed. 
【Coordinate System Type】 Determine the coordinate system type of the robot arm. 
【Pose Step】 The distance of single movement or rotation during mechanical calibration. 
【Rotation Angle Around Center】 The rotation angle when rotating around the image point during mechanical calibration. (This angle can be set to a larger value without jumping out of the field of view) 
【Center Pose】 The mechanical pose of the center point during translation and rotation in calibration. When subscribe is selected, the value can be referenced in the port as the center pose.
Calibration Judgment: 
【Max Displacement Error】 The maximum error threshold between calibration data and calculation data on the mechanical axis. 
【Max Angle Error】 The maximum error threshold between calibration data and calculation data angle. 
【Movement Data】 The mechanical movement data corresponding to the mechanical calibration. 
【Confirm】 Save the mechanical parameters. Be sure to click Confirm to save after setting the mechanical parameters, otherwise the set values will be restored to default values! 
## Output Description

![](/media/202603/2026-03-12_152607_2447040.8853923579898205.png)
The total run result outputs true after the calibration is successful, otherwise false.
The single run result outputs the run result of each photographing movement.
When calibration generation fails, both the total run result and the single output are false.
## Precautions
1. When using a bottom camera, X-axis mirroring is required.
2. The height of the plane for calibration needs to be the same as the height during use, otherwise there will be errors.
3. In normal mode (calibration mode), the first image of camera calibration is not included in the calculation, and the camera needs to take one more photo to complete the calibration.
## Problem Solving
Build the following flow to test the accuracy of the calibration in xy directions. Move the matching object 1-10mm, observe whether the calibration conversion result changes correspondingly to test the accuracy.
![](/media/202603/2026-03-12_152203_6181840.08111837465297045.png)
During alignment or picking, when the offset angle increases, causing the xy error to increase correspondingly, it is a calculation error of the calibration rotation center. You can solve the problem by increasing the calibration angle movement (more than 30 degrees). Or by increasing the number of rotations.

[【Attachment】Long Radius Calibration (Auto) Example Project.rar](/media/attachment/2026/03/%E9%95%BF%E5%8D%8A%E5%BE%84%E6%A0%87%E5%AE%9A%E8%87%AA%E5%8A%A8%E7%A4%BA%E4%BE%8B%E5%B7%A5%E7%A8%8B.rar)

---

Click here to jump to the corresponding Chinese manual page: [Long Radius Calibration (Auto)](/doc/126)