How to generate video demo with predicted results?

[zaneoo]

Like this (https://github.com/weiyithu/SurroundOcc/blob/main/assets/demo1.gif)

[weiyithu]

Hi, you can refer to this uncleaned code

[zaneoo]

Hi, you can refer to this uncleaned code

I got it, thanks so much! By the way, whether you guys will upload the version with temporal module in the future?

[USTCLH]

Hi, you can refer to this uncleaned code

Hi, how can i get the pose.npy used in visual_video.py ?

[weiyithu]

Hi, you can refer to this uncleaned code

Hi, how can i get the pose.npy used in visual_video.py ?

Hi，pose.npy is the extrinsics from LiDAR coordinate to camera coordinate. For nuscenes, it is same with the 'lidar2cam' in pkl files.

[VietPT3502]

Hi, you can refer to this uncleaned code

Hi, how can i get the pose.npy used in visual_video.py ?

Hi，pose.npy is the extrinsics from LiDAR coordinate to camera coordinate. For nuscenes, it is same with the 'lidar2cam' in pkl files.

can you give me the code to visualize video with the in_the_wild data

[mlxgzx]

Hi, you can refer to this uncleaned code

Hi, how can i get the pose.npy used in visual_video.py ?

Hi，pose.npy is the extrinsics from LiDAR coordinate to camera coordinate. For nuscenes, it is same with the 'lidar2cam' in pkl files.

Hi, I can't seem to find 'lidar2cam' in my generated NuScenes_inference_val.pkl file, is there any other way I can obtain the pose.npy file?

[Lacorey02]

Hi, I've downloaded the code, can you tell me how to get the "visual_path" and "scene_token"?

[mlxgzx]

Hi, I've downloaded the code, can you tell me how to get the "visual_path" and "scene_token"?

Both can be found in either the nuscenes info val/train pkl files.

[Lacorey02]

Hi, I can't seem to find 'lidar2cam' in my generated NuScenes_inference_val.pkl file, is there any other way I can obtain the pose.npy file?

Have you solved this problem?

[DURUII]

Hi, you can refer to this uncleaned code

Using visual_video.py, in-the-wild-data, and in-the-wild pkl lidar2cam data, I encounter the following issue: images are white, and the focal location appears to be incorrect.

Could anyone help me out？😭

[mlxgzx]

Hi, you can refer to this uncleaned code

Using visual_video.py, in-the-wild-data, and in-the-wild pkl lidar2cam data, I encounter the following issue: images are white, and the focal location appears to be incorrect.

Could anyone help me out？😭

I encountered the same problem and was able to solved it by setting the pc_range x ,y , z axis to 0

[Lacorey02]

I encountered the same problem and was able to solved it by setting the pc_range x ,y , z axis to 0

Have you ever encounter this situation,the six views on the top right are strange,and the next generated "front" and "top" views are different from the demo given by the author.

[958984240]

pc_range x

I encountered the same problem and was able to solved it by setting the pc_range x ,y , z axis to 0

Have you ever encounter this situation,the six views on the top right are strange,and the next generated "front" and "top" views are different from the demo given by the author.

Did you solve this problem?