Help testing code

[sezarxray]

Hi,
I hardly try to get result with your algorithm but failed...
I dont have lidar.
your code successfully run to end of my bag file but all features are classified to outlier!
trajectory only show line!!!
I change nearest/far depth and reprojection error and ....
I think reprojection error and depth scale is problem.
my imu installed frd and my camera is install in a manner to see down.

could you plz help me with this.

Best Regards.

my sensor config from kalibr:
distortion: [-0.09830587 0.12908673 0.00000912 0.00030652] +- [ 0.00149047 0.00370425 0.00017228 0.0002321 ]
projection: [ 1051.87491107 1049.92208525 664.78113662 375.40768068] +- [ 0.88491547 0.84498195 0.83076271
0.66532587]
reprojection error: [0.000000, 0.000000] +- [0.212658, 0.206313]
Transformation (cam0):

T_ci: (imu0 to cam0):
[[-0.02743445 0.99960638 -0.00586749 0.00271632]
[-0.9996219 -0.02744476 -0.00168346 -0.05918756]
[-0.00184383 0.00581909 0.99998137 -0.01591447]
[ 0. 0. 0. 1. ]]
T_ic: (cam0 to imu0):
[[-0.02743445 -0.9996219 -0.00184383 -0.05912 ]
[ 0.99960638 -0.02744476 0.00581909 -0.00424703]
[-0.00586749 -0.00168346 0.99998137 0.01583048]
[ 0. 0. 0. 1. ]]
timeshift cam0 to imu0: [s] (t_imu = t_cam + shift)
-0.0160764751881

###################### yaml file #######################################################

Configuration file for LE-VINS

结果输出路径

outputpath: "/data/LE-VINS/workspace/output"
is_make_outputdir: true

是否开启可视化

is_use_visualization: true

ROS接口配置

ros:
imu_topic: "/imu0"
image_topic: "/cam0"
lidar_topic: ""
use_compressed_image: false # 使用压缩图像

# 读取ROS包
is_read_bag: true
bag_file: ""

IMU噪声建模参数

imu:
imudatarate: 204 # IMU原始数据频率, Hz
is_use_zupt: true # 使用零速检测和零速约束

arw: 0.2 # deg/sqrt(hr)
vrw: 0.5 # m/s/sqrt(hr)
gbstd: 10.0 # deg/hr
abstd: 150.0 # mGal
corrtime: 1.0 # hr

gb_prior_std: 1500 # deg/hr
ab_prior_std: 10000 # mGal

优化器

optimizer:
optimize_reprojection_std: 15 # 像素重投影误差
optimize_point_to_plane_std: 0.1 # 点到面观测协方差, m
optimize_window_size: 10 # 滑动窗口大小

optimize_estimate_cam_extrinsic: true # 是否估计相机和IMU的外参
optimize_estimate_cam_td: true # 是否估计相机和IMU之间的时间间隔
optimize_cam_extrinsic_accurate: true # 相机和IMU的外参初值准确

visual:
is_use_lidar_depth: false # 使用激光深度增强
track_max_features: 200 # 最大提取特征数量

# 内参 [fx, fy, cx, cy(, skew)]
# intrinsic:
#     [
#         8.1640221474060002e+02,
#         8.1738388562809996e+02,
#         6.0882658427579997e+02,
#         2.6668865652440002e+02,
#         -2.3882017757999998e+00,
#     ]

# # 畸变参数 [k1, k2, p1, p2(, k3)]
# distortion:
#     [
#         -5.0040485799999999e-02,
#         1.2001217170000000e-01,
#         -6.2598060000000004e-04,
#         -1.1825064000000000e-03,
#         -6.3505207999999994e-02,
#     ]
intrinsic: [1051.87491107, 1049.92208525, 664.78113662, 375.40768068]

# 畸变参数 [k1, k2, p1, p2(, k3)]
# Distortion parameters
distortion: [-0.09830587, 0.12908673, 0.00000912, 0.00030652]

# 图像分辨率
resolution: [1280, 720]

# 相机IMU外参 Pb = q_b_c * Pc + t_b_c
# q (x, y, z, w)
q_b_c: [-0.707116, -0.000057, 0.000454, 0.707097]
t_b_c: [0, -.05, 0]

# IMU和相机时间延时
# t_b = t_c + td
td_b_c: 0

# 激光相机外参 Pc = q_c_l * Pl + t_c_l
# q (x, y, z, w)
q_c_l: [0.5, 0.5, 0.5, 0.5]
t_c_l: [0, 0, 0]

# t_b = t_l + td
td_b_l: 0.0

lidar:
# 激光参数
lidar_type: 1 # Livox: 1; Velodyne: 2; Ouster: 3
scan_line: 1 # 扫描线数
nearest_distance: 1 # 扫描盲区, m
farthest_distance: 250.0 # 最远距离, m
frame_rate: 25 # 帧率, Hz
####################################################