# ZED 2i Calibration and Skeleton Tracking

**URL:** <https://community.stereolabs.com/t/zed-2i-calibration-and-skeleton-tracking/10648>\
**Category:** ZED2\
**Created:** [January 16, 2026, 1:02pm UTC](https://community.stereolabs.com/t/zed-2i-calibration-and-skeleton-tracking/10648 "2026-01-16T13:02:39Z")\
**Posts on this page:** 15\
**Page:** 1

<div class="post-metadata">

**Author:** ![sachithwelarathna](https://community.stereolabs.com/letter_avatar_proxy/v4/letter/s/b5ac83/32.png) [@sachithwelarathna](https://community.stereolabs.com/u/sachithwelarathna)\
**Post date:** [January 16, 2026, 1:02pm UTC](https://community.stereolabs.com/t/zed-2i-calibration-and-skeleton-tracking/10648/1 "2026-01-16T13:02:39Z")

</div>

Hi [StereoLabs](https://www.stereolabs.com/) team and community,

I am working with multiple ZED 2i cameras for skeleton tracking (tracking distances between 2–10 m, testing one camera at a time under normal lighting conditions but for calibration I used a dark environment as advised). After evaluating five different units using the same procedure, I have observed some unexpected behaviour in both calibration and distance measurements. I would appreciate clarification before moving to the next stage of my experiments.

What I’ve observed

· As [StereoLabs](https://www.stereolabs.com/) recommends, some cameras performed better with factory calibration (FC) compared to user calibration.

· However, for other units, user calibration sometimes resulted in better performance than FC.

These inconsistencies raised several questions:

**My questions**

**1. Under what conditions does [StereoLabs](https://www.stereolabs.com/) recommend performing user calibration instead of relying on factory calibration?**

Are there specific indicators, symptoms, or accuracy thresholds that suggest a recalibration is necessary?

**2. Can user calibration degrade performance if environmental conditions are suboptimal** (e.g., poor lighting, monitor glare, low screen resolution, reflections)?

In other words, is the calibration process sensitive enough that suboptimal conditions may produce intrinsics worse than the factory defaults?

**3. Does the ZED Calibration Tool provide a quality score, confidence measure, or any metric** to help determine whether a calibration was successful or should be repeated?

**4. Do calibration parameters drift over time** , or should the factory calibration remain stable unless the camera experiences mechanical shock, temperature extremes, or physical damage?

**5. What is the recommended robust procedure for user calibration?**

The process appears susceptible to user-induced errors—particularly while moving the camera in front of the screen.

Is it preferable to minimize camera movement by mounting the camera during calibration?

Or is it acceptable to hand‑hold the camera and follow the on‑screen markers?

**6. Do all calibration parameters (intrinsics, distortion coefficients, stereo baseline, etc.) contribute equally to the final calibration quality?**

Or are certain parameters more critical for depth accuracy and downstream tasks such as skeleton tracking?

7. Across all five cameras tested, I observed systematic depth‑measurement biases: **significant overestimation (6–14%) at approximately 2 m, improved accuracy between 4–8 m (typically \<3% error), and increasing underestimation beyond 8 m**. Could you explain the factors that may cause these distance‑dependent variations in measurement accuracy?

Any guidance or best practices would be greatly appreciated.

Thank you!

---

<div class="post-metadata">

**Author:** ![Myzhar](https://community.stereolabs.com/user_avatar/community.stereolabs.com/myzhar/32/11_2.png) [@Myzhar](https://community.stereolabs.com/u/Myzhar)\
**Post date:** [January 20, 2026, 3:57pm UTC](https://community.stereolabs.com/t/zed-2i-calibration-and-skeleton-tracking/10648/2 "2026-01-20T15:57:46Z")

</div>

Hi @sachithwelarathna  
Welcome to the [Stereolabs](https://www.stereolabs.com/) community.

> [@sachithwelarathna](#):
>
> Under what conditions does [StereoLabs](https://www.stereolabs.com/) recommend performing user calibration instead of relying on factory calibration?

Normally, we recommend performing a manual calibration only if required:

- Using the camera in non-standard conditions (underwater, with modified optics, behind glasses, etc)
- The camera received a hard shock that could have invalidated the factory calibration

> [@sachithwelarathna](#):
>
> Are there specific indicators, symptoms, or accuracy thresholds that suggest a recalibration is necessary?

If you notice that the depth estimation is not as good as expected, then you can try to improve it by manual calibration.

> [@sachithwelarathna](#):
>
> **2. Can user calibration degrade performance if environmental conditions are suboptimal** (e.g., poor lighting, monitor glare, low screen resolution, reflections)?
> 
> In other words, is the calibration process sensitive enough that suboptimal conditions may produce intrinsics worse than the factory defaults?

Yes, you must perform the calibration always in **optimal conditions**.

> [@sachithwelarathna](#):
>
> **3. Does the ZED Calibration Tool provide a quality score, confidence measure, or any metric** to help determine whether a calibration was successful or should be repeated?

No, it does not. You can use the ZED OpenCV Calibration tool that provides score metrics at the end of the calibration process:

> **[GitHub - stereolabs/zed-opencv-calibration: Calibration tools to calibrate ZED cameras using...](https://github.com/stereolabs/zed-opencv-calibration)**
>
> Calibration tools to calibrate ZED cameras using OpenCV

> [@sachithwelarathna](#):
>
> **4. Do calibration parameters drift over time** , or should the factory calibration remain stable unless the camera experiences mechanical shock, temperature extremes, or physical damage?

Adverse environmental conditions can invalidate the factory calibration. Very low or very high temperatures and shocks can damage or slightly move the optics, causing a modification of the intrinsic values.

> [@sachithwelarathna](#):
>
> **5. What is the recommended robust procedure for user calibration?**
> 
> The process appears susceptible to user-induced errors—particularly while moving the camera in front of the screen.

We recommend using the validated process described here:

> **[GitHub - stereolabs/zed-opencv-calibration: Calibration tools to calibrate ZED cameras using...](https://github.com/stereolabs/zed-opencv-calibration)**
>
> Calibration tools to calibrate ZED cameras using OpenCV

> [@sachithwelarathna](#):
>
> 1. Do all calibration parameters (intrinsics, distortion coefficients, stereo baseline, etc.) contribute equally to the final calibration quality?

Yes, they all affect the quality of the calibration. There are no parameters that are more important than others.

> [@sachithwelarathna](#):
>
> 1. Across all five cameras tested, I observed systematic depth‑measurement biases: **significant overestimation (6–14%) at approximately 2 m, improved accuracy between 4–8 m (typically \<3% error), and increasing underestimation beyond 8 m**. Could you explain the factors that may cause these distance‑dependent variations in measurement accuracy?

Please provide more information regarding:

- the test setup
- the camera configuration
- the depth mode
- etc

Please find detailed information regarding the expected accuracy in this section of the online documentation:

> **[Depth Modes - Stereolabs](https://www.stereolabs.com/docs/depth-sensing/depth-modes)**
>
> Stereolabs AI Depth leverages advanced neural networks to generate high-quality depth maps from stereo images, delivering reliable results even in challenging scenarios. Compared to traditional …

---

<div class="post-metadata">

**Author:** ![sachithwelarathna](https://community.stereolabs.com/letter_avatar_proxy/v4/letter/s/b5ac83/32.png) [@sachithwelarathna](https://community.stereolabs.com/u/sachithwelarathna)\
**Post date:** [May 13, 2026, 10:04am UTC](https://community.stereolabs.com/t/zed-2i-calibration-and-skeleton-tracking/10648/3 "2026-05-13T10:04:10Z")

</div>

Thank you for the detailed and useful response. I would also like to know the following: I am particularly interested in using the ZED 2i camera in two different resolutions (HD1080 and 2K). My understanding is that the camera needs to be calibrated at the specific resolution in order to obtain optimal results.

I would like to ask whether it is acceptable to first calibrate the camera at the 2K resolution and then scale the calibration for other resolutions?

---

<div class="post-metadata">

**Author:** ![Myzhar](https://community.stereolabs.com/user_avatar/community.stereolabs.com/myzhar/32/11_2.png) [@Myzhar](https://community.stereolabs.com/u/Myzhar)\
**Post date:** [May 14, 2026, 2:23pm UTC](https://community.stereolabs.com/t/zed-2i-calibration-and-skeleton-tracking/10648/4 "2026-05-14T14:23:44Z")

</div>

> [@sachithwelarathna](#):
>
> My understanding is that the camera needs to be calibrated at the specific resolution in order to obtain optimal results.

This is not true. Al the ZED stereo cameras are factory calibrated using a precise automated process. You do not need to manually calibrate them, unless you use them in a modified optical condition (behind non-homogeneous glasses, underwater, etc).

---

<div class="post-metadata">

**Author:** ![sachithwelarathna](https://community.stereolabs.com/letter_avatar_proxy/v4/letter/s/b5ac83/32.png) [@sachithwelarathna](https://community.stereolabs.com/u/sachithwelarathna)\
**Post date:** [May 14, 2026, 2:45pm UTC](https://community.stereolabs.com/t/zed-2i-calibration-and-skeleton-tracking/10648/5 "2026-05-14T14:45:33Z")

</div>

Thank you for the response. I am using the OpenCV method for camera calibration. I use HD1080 and 2K modes for my experiments most of the time. But I want to clarify: if I use the OpenCV method as you mentioned in your previous response, do I need to perform calibration separately for the HD1080 mode and the 2K mode (not the factory calibration)?

---

<div class="post-metadata">

**Author:** ![Myzhar](https://community.stereolabs.com/user_avatar/community.stereolabs.com/myzhar/32/11_2.png) [@Myzhar](https://community.stereolabs.com/u/Myzhar)\
**Post date:** [May 14, 2026, 4:10pm UTC](https://community.stereolabs.com/t/zed-2i-calibration-and-skeleton-tracking/10648/6 "2026-05-14T16:10:46Z")

</div>

> [@sachithwelarathna](#):
>
> do I need to perform calibration separately for the HD1080 mode and the 2K mode (not the factory calibration)?

No, you can use the maximum resolution. The ZED SDK will automatically calculate the correct values for the lower resolutions.

---

<div class="post-metadata">

**Author:** ![sachithwelarathna](https://community.stereolabs.com/letter_avatar_proxy/v4/letter/s/b5ac83/32.png) [@sachithwelarathna](https://community.stereolabs.com/u/sachithwelarathna)\
**Post date:** [May 18, 2026, 11:00am UTC](https://community.stereolabs.com/t/zed-2i-calibration-and-skeleton-tracking/10648/7 "2026-05-18T11:00:01Z")

</div>

Small clarification: I used the OpenCV method to calibrate the ZED 2i at 2K resolution and generated a calibration file in `.yml` format, which contains the intrinsic, distortion, and stereo parameters (except Baseline) for that specific resolution. Then, I replaced the corresponding values in the factory calibration to use it as a `.conf` file for the camera. Do you mean that the ZED SDK will automatically compute the correct calibration values for the lower resolutions during use?

---

<div class="post-metadata">

**Author:** ![Myzhar](https://community.stereolabs.com/user_avatar/community.stereolabs.com/myzhar/32/11_2.png) [@Myzhar](https://community.stereolabs.com/u/Myzhar)\
**Post date:** [May 18, 2026, 12:48pm UTC](https://community.stereolabs.com/t/zed-2i-calibration-and-skeleton-tracking/10648/8 "2026-05-18T12:48:25Z")

</div>

> [@sachithwelarathna](#):
>
> Do you mean that the ZED SDK will automatically compute the correct calibration values for the lower resolutions during use?

Yes, that’s the behavior of the ZED SDK.

> [@sachithwelarathna](#):
>
> I used the OpenCV method to calibrate the ZED 2i at 2K resolution and generated a calibration file in `.yml` format, which contains the intrinsic, distortion, and stereo parameters (except Baseline) for that specific resolution. Then, I replaced the corresponding values in the factory calibration to use it as a `.conf` file for the camera.

You are not required to perform this. You can pass the YAML file directly to the ZED SDK using the parameter [optional\_opencv\_calibration\_file](https://www.stereolabs.com/docs/api/structsl_1_1InitParameters.html#a9eab2753374ef3baec1d31960859ba19) of `InitParameters`.

---

<div class="post-metadata">

**Author:** ![sachithwelarathna](https://community.stereolabs.com/letter_avatar_proxy/v4/letter/s/b5ac83/32.png) [@sachithwelarathna](https://community.stereolabs.com/u/sachithwelarathna)\
**Post date:** [May 18, 2026, 2:26pm UTC](https://community.stereolabs.com/t/zed-2i-calibration-and-skeleton-tracking/10648/9 "2026-05-18T14:26:03Z")

</div>

Thank you for the clarification.

1. Do you have any suggestions regarding the checkerboard and setup for the OpenCV calibration method?

2. I tried using checkerboards with different square sizes in my experiments (25.4 mm, 48 mm, and 72 mm). I placed the camera (I am using multiple cameras for my experiments) at a fixed location with appropriate lighting and then moved the checkerboard to different positions for sample collection. However, most of the time the reprojection error exceeded 0.5 (stereo error was more prominent), and the calibration was not successful, so the .yml file was not generated (it only worked successfully a few times), even though I achieved 100% on all four metrics (X coverage, Y coverage, skew, and size).

3. Do you also have any data or an opinion on the stability of the ZED 2i camera once it has been calibrated (i.e., stability of depth measurements over time)?

---

<div class="post-metadata">

**Author:** ![Myzhar](https://community.stereolabs.com/user_avatar/community.stereolabs.com/myzhar/32/11_2.png) [@Myzhar](https://community.stereolabs.com/u/Myzhar)\
**Post date:** [May 21, 2026, 11:44am UTC](https://community.stereolabs.com/t/zed-2i-calibration-and-skeleton-tracking/10648/10 "2026-05-21T11:44:41Z")

</div>

Hi @sachithwelarathna  
can you please test the new branch [fix\_data\_acquisition\_check](https://github.com/stereolabs/zed-opencv-calibration/tree/fix_data_acquisition_check) of the `zed-opencv-calibration` repository?

I improved the check of the quality of the acquired images and tuned the parameters in order to minimize the reprojection error.

Looking forward to your feedback.

---

<div class="post-metadata">

**Author:** ![sachithwelarathna](https://community.stereolabs.com/letter_avatar_proxy/v4/letter/s/b5ac83/32.png) [@sachithwelarathna](https://community.stereolabs.com/u/sachithwelarathna)\
**Post date:** [May 23, 2026, 2:36pm UTC](https://community.stereolabs.com/t/zed-2i-calibration-and-skeleton-tracking/10648/11 "2026-05-23T14:36:09Z")

</div>

Hi @Myzhar I will keep you updated on this (in coming days). I believe this approach would be highly beneficial. Before using this method, could you clarify something for me? I am planning to deploy this camera for skeleton tracking (and later possibly for face identification/re-identification), and I intend to use it up to 10 m for the first application and even farther for the second.

In that case, do I need to acquire calibration samples at such long distances (for example upto 10 m for first case, and upto 20m for the second case)? Is it important? Alternatively, could you recommend how I should position the checkerboard, especially considering that this mainly affects the SIZE.

Do you have any recommendations for using this camera for human identification tasks?

---

<div class="post-metadata">

**Author:** ![sachithwelarathna](https://community.stereolabs.com/letter_avatar_proxy/v4/letter/s/b5ac83/32.png) [@sachithwelarathna](https://community.stereolabs.com/u/sachithwelarathna)\
**Post date:** [May 24, 2026, 3:04pm UTC](https://community.stereolabs.com/t/zed-2i-calibration-and-skeleton-tracking/10648/12 "2026-05-24T15:04:52Z")

</div>

**Update on** `fix_data_acquisition_check`

I attempted to build the project and encountered a few minor issues, which I have now fix as follows:

**1. Duplicate case in** `main.cpp` **(around line 95) in** `stereo_reprojection_viewer`

The original code contained a duplicated case:

cpp

```cpp
case sl::MAT_TYPE::U8_C3: 
    cv_type = CV_8UC3; 
    break;
case sl::MAT_TYPE::U8_C3: // ← Duplicate
    cv_type = CV_8UC3; 
    break;

```

I updated it to:

cpp

```auto
case sl::MAT_TYPE::U8_C3: cv_type = CV_8UC3; break;
case sl::MAT_TYPE::U8_C4: cv_type = CV_8UC4; break;

```

**2. Undeclared identifier** `M_PI` **in** `calibration_checker.cpp`

I defined `PI` manually as:

cpp

```cpp
constexpr float PI = 3.1415926535f;

```

**3. Missing** `<iomanip>` **for** `std::setprecision`

I added the required include:

cpp

```cpp
#include <iomanip>

```

After applying these fixes, the project compiles successfully.

Note: Then I attempted the calibration, but the reprojection error is still above 0.5, even though all metrics reached a 100% score. For this test I used a 25 mm square size, but I will try other square sizes as well. 1) I appreciate your feedback on this. 2) Could you recommend an alternative threshold value to 0.5 (although it is commonly used/acceptable) that would be suitable for computer vision applications such as skeleton tracking and face identification.

---

<div class="post-metadata">

**Author:** ![Myzhar](https://community.stereolabs.com/user_avatar/community.stereolabs.com/myzhar/32/11_2.png) [@Myzhar](https://community.stereolabs.com/u/Myzhar)\
**Post date:** [May 25, 2026, 3:10pm UTC](https://community.stereolabs.com/t/zed-2i-calibration-and-skeleton-tracking/10648/13 "2026-05-25T15:10:01Z")

</div>

> [@sachithwelarathna](#):
>
> In that case, do I need to acquire calibration samples at such long distances (for example upto 10 m for first case, and upto 20m for the second case)?

No, this is not required. The calibration is used to estimate the camera parameters that remain the same at every distance.

---

<div class="post-metadata">

**Author:** ![sachithwelarathna](https://community.stereolabs.com/letter_avatar_proxy/v4/letter/s/b5ac83/32.png) [@sachithwelarathna](https://community.stereolabs.com/u/sachithwelarathna)\
**Post date:** [June 22, 2026, 7:31am UTC](https://community.stereolabs.com/t/zed-2i-calibration-and-skeleton-tracking/10648/14 "2026-06-22T07:31:33Z")

</div>

Hi @Myzhar I have successfully calibrated camera after tying the [fix\_data\_acquisition\_check](https://github.com/stereolabs/zed-opencv-calibration/tree/fix_data_acquisition_check). This is the calibration file I generated. Could you verify whether the layout of the calibration file is correct? (the layout is bit different from .conf file, so I want to verify it) I also performed depth‑estimation experiments and will post them soon.

YAML:1.0  
Size: [2208, 1242]  
K\_LEFT: !!opencv-matrix  
rows: 3  
cols: 3  
dt: d  
data: [ 1.0676519472817354e+03, 0., 1.0978201978812779e+03, 0.,  
1.0681617095488198e+03, 6.0064326133212433e+02, 0., 0., 1. ]  
K\_RIGHT: !!opencv-matrix  
rows: 3  
cols: 3  
dt: d  
data: [ 1.0704691555698448e+03, 0., 1.1088176130598445e+03, 0.,  
1.0710811388561497e+03, 6.1180592743498369e+02, 0., 0., 1. ]  
D\_LEFT: !!opencv-matrix  
rows: 1  
cols: 14  
dt: d  
data: [ -1.6795706746958645e+00, 2.0923969565900005e+00,  
-6.7634871941727869e-04, 4.9154385414694880e-04,  
4.5925190016088874e-01, -1.5947408018141676e+00,  
1.9194489863462014e+00, 6.2634679248367220e-01, 0., 0., 0., 0.,  
0., 0. ]  
D\_RIGHT: !!opencv-matrix  
rows: 1  
cols: 14  
dt: d  
data: [ -1.9451915416917565e+00, 3.6974781795013754e+00,  
-7.2739513483112894e-04, 9.6443399400409445e-06,  
2.1067530891981975e-01, -1.8672903536476264e+00,  
3.5495111301131432e+00, 4.0209503974962801e-01, 0., 0., 0., 0.,  
0., 0. ]  
R: !!opencv-matrix  
rows: 3  
cols: 1  
dt: d  
data: [ 4.9880168994476835e-03, 5.6520662727681543e-03,  
1.1367546543679379e-03 ]  
T: !!opencv-matrix  
rows: 3  
cols: 1  
dt: d  
data: [ -1.1986634649529667e+02, -2.9533694948442074e-01,  
3.8496285406407665e-01 ]

---

<div class="post-metadata">

**Author:** ![Myzhar](https://community.stereolabs.com/user_avatar/community.stereolabs.com/myzhar/32/11_2.png) [@Myzhar](https://community.stereolabs.com/u/Myzhar)\
**Post date:** [June 26, 2026, 4:11pm UTC](https://community.stereolabs.com/t/zed-2i-calibration-and-skeleton-tracking/10648/15 "2026-06-26T16:11:36Z")

</div>

Hi @sachithwelarathna  
This is an OpenCV YAML file.  
It’s not automatically loaded by the ZED SDK.

You must set the path to it in the InitParameters::[optional\_opencv\_calibration\_file](https://www.stereolabs.com/docs/api/structsl_1_1InitParameters.html#a9eab2753374ef3baec1d31960859ba19) to use it
