Need help in writing "config_dewarper_perspective.txt" in the correct format

Accelerated Computing Intelligent Video Analytics DeepStream SDK
,
1

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU): Nvidia Orin Jetson 64Gb
• DeepStream Version: 6.2
• JetPack Version (valid for Jetson only) : 5.1-b147
• TensorRT Version: 8.5.2-1+cuda11.4
• Issue Type( questions, new requirements, bugs): question

Hello,
I want to use “Gst-nvdewarper” to undistort my image from a USB RGB camera which has a 3.5mm lens and 1/1.7" sensor size. Following are my camera intrinsic and distortion parameters:

Image width: 1616
Image height: 1240

fx= 776.2205796983008
fy= 775.1831242733599
cx= 779.3577422263877
cy= 630.5392491612774

k1= -1.7550965901443998
k2= 0.42233583084391374
k3= 0.45023886093988386
k4= -1.534818381528613
k5= -0.04519751751606131
k6= 0.7163137290337043
p1= -0.00043710473030367426
p2= -0.0009160066845597751

camera_matrix = np.array([
[fx, 0, cx],
[0, fy, cy],
[0, 0, 1]
], dtype=np.float)

distortion = np.array([[k1], [k2], [p1], [p2], [k3], [k4], [k5], [k6]], dtype=np.float)

undistorted = cv2.undistort(image, camera_matrix, distortion, None, camera_matrix)

I have used the undistort-function of openCV in the past as shown above and it works well with these parameters but I want to utilize Deepstream for this.

I am having a tough time creating the “config_dewraper_perspective.txt” file. I have taken reference from this
config_dewarper_1280x720.txt (2.1 KB) but it’s giving me weird results. Can you please help me to figure out which parameter goes where in the file?

3

Please refer to Dewarp Deepstream 6 - Intelligent Video Analytics / DeepStream SDK - NVIDIA Developer Forums for the parameters calculation.

4

I am getting a “nvdewraper0 : Failed to parse config file” error if I am using focal-length=776.2205796983008/775.1831242733599/avg(776.2205796983008,775.1831242733599)
but it works if I use focal-length=776.2205796983008; 775.1831242733599
I am confused on what’s the correct way to use it.

Also if I have 8 distortion parameters instead of 5 (as I mentioned above), can I still use this pipeline?

5

There is no update from you for a period, assuming this is not an issue anymore. Hence we are closing this topic. If need further support, please open a new one. Thanks

This is the formula, please calculate the value and fill the value in the configuration file.

We use the alogorithm discribed in Camera calibration With OpenCV — OpenCV 2.4.13.7 documentation. If the 8 distortion parameters are compatible to it, you can use the corresponding 5 parameters.

7

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.