Mazhichao

It means can not find the node,are there some errors or warning when you "catkin_make" the source codes?

it looks not not be make well done.

looks like it is based on cuda rely on the nvidia gpu,so I can't find the .so files when you running on the cpu. the detection running on cpu depends...

sorry for my mistakes,if You want to run by CPU(Opencv DNN),the conflict of opencv version must be sloved，the ROS melodic opencv is 3.2,But this version dont have DNN,so you should...

这个log也看不出发生了什么呀，进程崩溃问题可能要费些时间check，通常有些数据维度，图像大小的问题吧

可以，之前有看到南洋理工有个工作，使用realsene 的L515做的，有一定效果；但是可以预见肯定有很多局限，应对旋转容易崩溃。ZED双目相机视场大，本身考视差获取深度效果还是会好一点 ------------------ 原始邮件 ------------------ 发件人: "Mazhichaoruya/Perception-of-Autonomous-mobile-robot" ***@***.***>; 发送时间: 2021年11月2日(星期二) 上午10:03 ***@***.***>; ***@***.***>; 主题: [Mazhichaoruya/Perception-of-Autonomous-mobile-robot] Apply issue? (Issue #1) 你好， 请问这种RGBD相机是否能够直接建立深度地图（深度作为输入跑类似LOAM），他的性能和和其他RGBD相机（Kinect， ZED， 奥比中广），有实际的明显差异吗？ — You are receiving this because you are subscribed to this...