TDA3: Surround view execution

Masami-San,

I'm not sure I understand the question. Can you please reach out to Ti field and explain your queries to them

Regards

Shashank

Hi Shashank.
Thank you for responding.

I'm sorry for lack of explanation.

I want to simulate a new surround_view image on PC using the generated data(GAlignLUT table-data and InitialPerspectiveParams),
but the original images(front/rear/left/right) used at simulating are different from the images used at the "calibration_tool" and "surroundview.exe".
There are no markers in the original images.

Regards
Masami

Masami-San,

Are you able to use the manual calibration tool released with vision sdk? We provide only two tools for surroundview

1) A tool for performing manual calibration: This will take in 4 chart images , create the calibration data and generate the output

2) A mesh tool: This is used for specifying the views to create again using 4 images. The images can be of a real scene that you want to simulate. You can input the calmat.bin file from step-1 to account for calibration as long as the images being used are captured with the same pose of the cameras. This will help in creating a reference of how the final output will look like (Please note that this is done on GPU so the output won't be bit accurate as the embedded output from tda3x)

Both these tools are released with vision sdk. You should use these tools to create your simulation results.

Regards

Shashank

Hi Shashank.

Thank you for responding.

I'll try the contents of No.2.
I knew the existence of the tool, but I have never used it before.

Regards
Masami