利用PSMNET生成kitti_odometry_00的深度图像，最大深度是128。然后利用vins-fusion进行该算法的测试，结果非常不好，不知道什么原因？

HKUST-Aerial-Robotics / DenseSurfelMapping

This is the open-source version of ICRA 2019 submission "Real-time Scalable Dense Surfel Mapping"

682 stars 140 forks source link

Closed yueshukun closed 4 years ago

yueshukun commented 4 years ago

具体测试截图如下： Screenshot from 2020-05-19 11-17-39 生成的深度图如下： 000000 求高手解答，谢谢！

WANG-KX commented 4 years ago

The result fro PSMNet is a disparity map. Convert to depth maps (meters) before any usage.

708949503 commented 4 years ago

您好，请问您之前提出的那个问题解决了吗？ OpenCV Error: Assertion failed (0 <= _rowRange.start && _rowRange.start <= _rowRange.end && _rowRange.end <= m.rows)

zheng-xiao commented 2 years ago

请问，您的视差图还保存着吗？我使用PMSNet生成的视差图精度不好

zheng-xiao commented 2 years ago

如果还保存，可以分享给我一份吗？我的邮箱是1287595254@qq.com。非常感谢