Problem about KITTI experiment MODE

[xuwuzhou]

Thanks for your good work, and I have some questions, wish for your reply.
Question No.1:This experiment seems that not support the mode which only input lidar and visual measurements. I am sure of this, because I have tried a variety of matching environments for nearly 10 days(Since I think it is a problem with environment at the beginning). In particular, the environments I have tried are recorded as:
A. a desktop with intel i710700+ ubuntu18+ ceres1.14+eigen3.3.7+opencv2.14.3+pcl1.7
B. a desktop with intel i710700+ ubuntu18+ ceres2+eigen3.3.7+opencv2.14.3+pcl1.7
C. a desktop with intel i710700+ ubuntu18+ ceres1.14+eigen3.3.7+opencv3.4.5+pcl1.7
D. a desktop with intel i710700+ ubuntu18+ ceres1.14+eigen3.3.7+opencv3.4.5+pcl1.8
E. a desktop with intel i710700+ ubuntu18+ ceres1.14+eigen3.3.7+opencv2.14.3+pcl1.8
F. a desktop with intel i710700+ ubuntu18+ ceres1.14+eigen3.3.7+opencv4.X.X+pcl1.8
G. a laptop with inteli7-6700+ ubuntu 16+ ceres1.14+eigen3.3.7+opencv2.14.1+pcl1.7
H. a laptop with inteli7-6700+ ubuntu 16+ ceres1.14+eigen3.3.7+opencv2.14.1+pcl1.8
I. a laptop with inteli7-6700 + ubuntu 16 + ceres1.14+eigen3.3.7+opencv3.4.5+pcl1.7
J. a laptop with inteli7-6700 + ubuntu 16 + ceres1.14+eigen3.3.7+opencv3.4.5+pcl1.8
K. a laptop with inteli7-6700 + ubuntu 16 + ceres1.14+eigen3.3.7+opencv4.X+pcl1.8
None of the above environments can run Mode( lidar+Camera), and always collapse for the same reason(As shown in the figures below)

[xuwuzhou]

Question No.2: Last Year I opened issue6 #6 ，and find that the result generated by Mode(IMU+lidar+Camera) are
significantly worse than the result you showed in paper(Mode is IMU+GPS+lidar+Camera ).
The accuracy recorded by Mode(IMU+Lidar+Camera) is 16.22, and the accuracy recorded in your paper (Mode is GPS+IMU+Camera+Lidar) is 1.13.Why does this happen? What do you think is the reason? I guess it may be that the IMU data of KITTI is not accurate enough, leading to poor accuracy?

[xuwuzhou]

Sincerely look forward to your reply, best wishes!

[xuwuzhou]

@jypjypjypjyp