-
Notifications
You must be signed in to change notification settings - Fork 271
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
BEVDet4D特征对齐的问题 #236
Comments
align_after_view_transfromation是指是否在lss之后对齐,True的话,是把lss生成的bev_feat对齐,而False的话,就是在lss计算映射关系的时候直接映射到当前帧的ego空间,所以align_after_view_transfromation与你的这个问题无关。 |
非常感谢!
…---- 回复的原邮件 ----
| 发件人 | ***@***.***> |
| 日期 | 2023年05月16日 14:23 |
| 收件人 | ***@***.***> |
| 抄送至 | ***@***.***>***@***.***> |
| 主题 | Re: [HuangJunJie2017/BEVDet] BEVDet4D特征对齐的问题 (Issue #236) |
align_after_view_transfromation是指是否在lss之后对齐,True的话,是把lss生成的bev_feat对齐,而False的话,就是在lss计算映射关系的时候直接映射到当前帧的ego空间,所以align_after_view_transfromation与你的这个问题无关。
with_prev要为True,在训练时有个hook,SequentialControlHook,在某个epoch之后with_prev就会为True的。
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you authored the thread.Message ID: ***@***.***>
|
"而False的话,就是在lss计算映射关系的时候直接映射到当前帧的ego空间" 请问这一步是怎么实现的呢?我在LSSViewTransformer没有找到相关code,似乎LSSViewTransformer只是把img features project到对应时刻的bev坐标系下? |
计算img2ego转换矩阵的时候,算了 sweep_img2key_ego,LSS投影时就直接把任意时刻的视锥都投到当前帧的ego下了 |
你说的这部分实现是指get_inputs函数中调用的get_sensor_transforms部分吗?(https://github.com/HuangJunJie2017/BEVDet/blob/dev2.1/mmdet3d/datasets/pipelines/loading.py#L936) |
他说的应该是BEVDet4D里面的prepare_inputs吧? |
脑子抽了后知后觉才醒悟说的是什么ORZ。这部分大概是在https://github.com/HuangJunJie2017/BEVDet/blob/dev3.0/mmdet3d/models/detectors/bevdet.py#L408 实现的。希望能帮到所有同有疑问的朋友们:) |
作者您好!我在调试BEVDet4D-CBGS时发现每次生成bev特征时只用了第一帧的信息。请问我如果想用两帧的信息是否是把with_prev设置成True,同时把align_after_view_transfromation也设置成True。请问这样可以吗?
The text was updated successfully, but these errors were encountered: