search

ruhyadi / YOLO3D

YOLO 3D Object Detection for Autonomous Driving Vehicle
https://ruhyadi.github.io/project/computer-vision/yolo3d
247 stars 43 forks source link

Converting script #8

Closed pdd-vn closed 2 years ago

pdd-vn commented 2 years ago

Do you have any conventional script that can be used to convert raw kitti data (download with this script) to your training format? Currently I am using this format and it's fine. Btw why did you use a global calib_cam_to_cam.txt? I can see that there is one calib_cam_to_cam.txt for each "date" index.

ruhyadi commented 2 years ago

Sorry for the slow response. I didn't check the notification. The conversion code from kitti format to yolo3d format (the format in this repository) is already loaded in yolo3d/script/Dataset.py. You can check in line 93 and below. The reason I'm doing this is so we don't have to do the conversion up front. So, the conversion will take place during training.

As for calib_cam_to_cam.txt, I assume that the camera/lidar calibration value does not change from the beginning to the end of the training data. But if you want to use the calibration data of each image, you can comment line 45, and adjust the code.

Hope this helps,

pdd-vn commented 2 years ago

@ruhyadi My issue is resolved. Thank you very much.