Onboard camera is rotated

I am followed the directions provided in the README file for 11_camera_object_identification and I successed.

However, since the camera image is rotated 180 degrees. So some object may be misrecognized.

How do I do to obtain and use camera image that not rotated?

Please add below line in creating NvVideoConverter


Thank you. I solve this problem.

But I have another problem.

When I execute only camera, output image is not delay.
But if I execute this example, output image is delay about 2 second (when camera width=1920, height=1080).

How do I do to solve this problem?

It takes certain time in running deep learning model. You may to execute jetson_clocks.sh to get max performance.