Image reconstruction from BearingAngleImage not working

Using the PCL BearingAngleImage class updates points in the input point cloud to contain grayscale values proportional to the bearing angles calculated. Reconstructing the image from these points does not work.

Your Environment

• Operating System and version: Ubuntu 16.04 LTS
• Compiler: gcc
• PCL Version:

Context

Trying to construct a 2D image from a 3D pointcloud

Expected Behavior

Output a 2D image

Current Behavior

3D points are provided but the ordering is unclear

Code to Reproduce

pcl::BearingAngleImage bearing_angle_image;
bearing_angle_image.generateBAImage(*pcl_cloud_xyz);
cv::Mat Im(bearing_angle_image.height, bearing_angle_image.width, CV_8UC1, cv::Scalar(255));
int count = 0;
for (int i=0; i< bearing_angle_image.height-1; i++)
{
for (int j=0; j< bearing_angle_image.width-1; j++)
{
Im.at(i,j) = (bearing_angle_image.points[count].rgba >> 8) & 0xff;
count++;
}
}
cv::namedWindow("image");
cv::imshow("image", Im);
cv::waitKey(0);
cv::destroyAllWindows();

Possible Solution

The function should return an image, or maybe provide documentation on how to reconstruct this

Start with the BearingAngleImage API and the supplied generateBAImage reproducer; the payload does not name an implementation file or test. Determine how generated points map to image rows and columns, then verify that reconstruction produces the expected 2D image or document the required ordering.