pedropro / cape Goto Github PK

This is a great work, thanks for your sharing.
I have a structure sensor in my hand, I want to get the RGB and depth data from Structure sensor, then send them to your CAPE module to detect planes.
In the "Note", you wrote "The parameters are fine-tuned for detecting large surfaces with Kinect 1 and Structure sensor".
So I want to know how to get data from Structure sensor by coding. Could you please offer me some means or links to show how to do that?

Dear Developers,

I have only depth image. I don't have rgb image. Could you tell me how to modify your code in order to run on my dataset in that case?
I am trying to modify your code but no success((
Thanks for your attention!

According to the instruction, I target to the sequence folder, but the missing calibration file is weird. Since the .xml file is already there. My OS is ubuntu 14.04 64bit.

Good day! Thanks a lot for your work. I would like to understand the parameters of flat surfaces. I'm interested in how I can get the equation of a plane or its other parameters in 3D space relative to the camera. Do I understand correctly that each plane is defined by parameters mean[3] and normal[3]. I have these parameters defined from -1 to 1. As I understand these parameters are normalized. But I still have no idea how to restore the position and orientation of the plane in three-dimensional space from them. Could you help me with this?

Hi! I am trying to use the CAPE detector for plane detection. I have a problem when converting the input data from organized pcl::PointCloud to Eigen::MatrixXf.

Base on my understanding, the input data is a width*height x 3 matrix (2D), but I don't know the order of it. It will be great if you can give me some hints on it!

Hi,Thanks for your sharing,I have replaced the dataset with my own depth data which is 8 bit,but it did't work.I think the key point is at ‘d_img.convertTo(d_img,CV_32F)’,which cause nr_pt s< min_nr_pts.So counld you please tell me what can I do to make it right?Wish for your reply.