JOP-Lee / READ

AAAI2023,implementation of "READ: Large-Scale Neural Scene Rendering for Autonomous Driving", the experimental results are significantly better than Nerf-based methods
https://github.com/JOP-Lee/READ-Large-Scale-Neural-Scene-Rendering-for-Autonomous-Driving
GNU General Public License v2.0
447 stars 55 forks source link

What's the setting of metashape Align photo on the kitti pointcloud data? #18

Closed Kanon777 closed 1 year ago

Kanon777 commented 1 year ago

I tried to run Align Photos on the kitti dataset to reproduce the result, but found that the pointcloud generated by metashape is much sparser and lower quality than the pointcloud provided in this repo. I guess this may caused by the difference of settings in the metashape software, so could you please provide the detail settings of metashape align photo that produce the point cloud?

Fllacsaa commented 1 year ago

Same question here @JOP-Lee

shuhangz commented 1 year ago

You can try 'build dense cloud' in the Metashape after the 'align photo' is finished. The step of 'align photo' just build a sparse point cloud with SfM. The 'build dense cloud' performs the dense image matching in MVS. Detailed instructions can be referenced here: https://agisoft.freshdesk.com/support/solutions/articles/31000152092

JOP-Lee commented 1 year ago

You can try 'build dense cloud' in the Metashape after the 'align photo' is finished. The step of 'align photo' just build a sparse point cloud with SfM. The 'build dense cloud' performs the dense image matching in MVS. Detailed instructions can be referenced here: https://agisoft.freshdesk.com/support/solutions/articles/31000152092

Exactly! @Fllacsaa @Kanon777