search

mileyan / pseudo_lidar

(CVPR 2019) Pseudo-LiDAR from Visual Depth Estimation: Bridging the Gap in 3D Object Detection for Autonomous Driving
https://mileyan.github.io/pseudo_lidar/
MIT License
967 stars 215 forks source link

About the calibration problem between the true location and the Pseudo-LiDAR #51

Closed chenhao-1115 closed 3 years ago

chenhao-1115 commented 3 years ago

Hello, Firstly, thank you for the amazing work with this repo. May I ask how you calibrate between the true location and the Pseudo-LiDAR? Through the guidance of your program, I have realized the generation of Pseudo-LiDAR, but when I tried to mark the true location of the object, I found it was not correct.

As shown in the following experiment: The true location: ./KITTI/object/training/label_2/000000.txt Pedestrian 0.00 0 -0.20 712.40 143.00 810.73 307.92 1.89 0.48 1.20 1.84 1.47 8.41 0.01 The true location of the object in this annotation file is: 1.89 0.48 1.20

1

The experimental results are shown in the figure above. The red box represents the picture I tested; the blue box represents the generated Pseudo-LiDAR. It can be seen that the generated Pseudo-LiDAR is fine. The yellow box represents the true location I marked (1.89 0.48 1.20). From the figure, it is obvious that there is a difference between the Pseudo-LiDAR and the true location. How do you calibrate in actual operation?

chenhao-1115 commented 3 years ago

I have solved the problem.