Hello, thanks for your great job.
In my project, for the reason of occlusion, the lidar has only 270 degree。 In theory the performance would obviously be a little poorer, especially for the SC based loop closure. In my test, the loop closure recall would drop from over 90% to about 30% if the loop frame direction to the previous frame is about 30 degree.
Have you or anyone else tested with different FOV lidar (full 360 degree or partial 270 degree)? How the overall mapping error differs with different FOV lidar? And do you have any suggestion for adjusting the great algorithm to only 270 degree FOV lidar?
Thanks for your attention and I am always keeping looking forward to your kind response.
Hello, thanks for your great job. In my project, for the reason of occlusion, the lidar has only 270 degree。 In theory the performance would obviously be a little poorer, especially for the SC based loop closure. In my test, the loop closure recall would drop from over 90% to about 30% if the loop frame direction to the previous frame is about 30 degree. Have you or anyone else tested with different FOV lidar (full 360 degree or partial 270 degree)? How the overall mapping error differs with different FOV lidar? And do you have any suggestion for adjusting the great algorithm to only 270 degree FOV lidar? Thanks for your attention and I am always keeping looking forward to your kind response.