search

halajun / VDO_SLAM

VDO-SLAM: A Visual Dynamic Object-aware SLAM System
Other
734 stars 113 forks source link

how to generate the depth map? #16

Closed shqfeng closed 3 years ago

shqfeng commented 3 years ago

I try to use SGBM to calculate the depth, but it dose not work! Can you share the project of depth map generation?

shqfeng commented 3 years ago

thanks for your reply！ how to generate the "depth" folder of KITTI stereo dataset, I use the opencv SGBM, hower, the results is bad, can you share this work? thanks very much!

------------------ 原始邮件 ------------------ 发件人: @.>; 发送时间: 2021年3月29日(星期一) 晚上9:54 收件人: @.>; 抄送: @.>; @.>; 主题: Re: [halajun/VDO_SLAM] how to generate the depth map? (#16)

It's for monocular

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

halajun commented 3 years ago

Hi, @shqfeng. the method SGBM generates disparity map, not depth map. You need to compute the depth value from disparity by [depth = baseline*focal_lenth/(disparity/depth_factor)]. For your reference, I use the method of "Efficient Joint Segmentation, Occlusion Labeling, Stereo and Flow Estimation" to calculate the disparity map. Hope this helps.

shqfeng commented 3 years ago

thanks！

------------------ 原始邮件 ------------------ 发件人: "Jun Zhang @.>; 发送时间: 2021年4月1日(星期四) 上午6:58 收件人: @.>; 抄送: @.>; @.>; 主题: Re: [halajun/VDO_SLAM] how to generate the depth map? (#16)

Hi, @shqfeng. the method SGBM generates disparity map, not depth map. You need to compute the depth value from disparity by [depth = baseline*focal_lenth/(disparity/depth_factor)]. For your reference, I use the method of the paper "Efficient Joint Segmentation, Occlusion Labeling, Stereo and Flow Estimation" to calculate the disparity map. Hope this helps.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.