kitti

[gzyabc]

When I run in kitii data set, why are the weighted feature full of red dots?I modified config with reference to the mono file.

%YAML:1.0

#common parameters
imu: 1
num_of_cam: 1

imu_topic: "/imu_raw"
image0_topic: "/kitti/camera_gray_left/image_raw"
image1_topic: "/kitti/camera_gray_left/image_raw"
output_path: "/home/gzy/dynavins_ws/src/output/"

cam0_calib: "cam04-12.yaml"
cam1_calib: "cam04-12.yaml"
image_width: 1241
image_height: 376

Extrinsic parameter between IMU and Camera.

estimate_extrinsic: 0 # 0 Have an accurate extrinsic parameters. We will trust the following imu^R_cam, imu^T_cam, don't change it.
# 1 Have an initial guess about extrinsic parameters. We will optimize around your initial guess.

body_T_cam0: !!opencv-matrix
rows: 4
cols: 4
dt: d
data: [1, 0, 0, 0,
0, 1, 0, 0,
0, 0, 1, 0,
0, 0, 0, 1]

body_T_cam1: !!opencv-matrix
rows: 4
cols: 4
dt: d
data: [1, 0, 0, 0.537150653267924,
0, 1, 0, 0,
0, 0, 1, 0,
0, 0, 0, 1]

#Multiple thread support
multiple_thread: 4

#feature traker paprameters
max_cnt: 150 # max feature number in feature tracking
min_dist: 15 # min distance between two features
freq: 10 # frequence (Hz) of publish tracking result. At least 10Hz for good estimation. If set 0, the frequence will be same as raw image
F_threshold: 2.0 # ransac threshold (pixel)
max_depth: 5.0 # max estimated depth (m)
show_track: 1 # publish tracking image as topic
show_image_feat_weight: 1
flow_back: 1 # perform forward and backward optical flow to improve feature tracking accuracy

#dynaVINS parameters
dyna_on: true # do not change it to false
regularization_lambda: 2.0
momentum_on: true
momentum_lambda: 0.2
alternating_converge: 0.9
margin_feature_thresh: 0.1

#optimization parameters
max_solver_time: 3 # max solver itration time (s), to guarantee real time
max_num_iterations: 10 # max solver itrations, to guarantee real time
keyframe_parallax: 10.0 # keyframe selection threshold (pixel)

#imu parameters The more accurate parameters you provide, the better performance
acc_n: 0.1 # accelerometer measurement noise standard deviation. #0.2 0.04
gyr_n: 0.01 # gyroscope measurement noise standard deviation. #0.05 0.004
acc_w: 0.001 # accelerometer bias random work noise standard deviation. #0.02
gyr_w: 1.0e-4 # gyroscope bias random work noise standard deviation. #4.0e-5
g_norm: 9.81007 # gravity magnitude

#unsynchronization parameters
estimate_td: 0 # online estimate time offset between camera and imu
td: 0.0 # initial value of time offset. unit: s. readed image clock + td = real image clock (IMU clock)

Look forward to your advice

[gzyabc]

used features: 112
used features: 112
COSTDIFF:0.291534
used features: 112
used features: 112
COSTDIFF:0.997322
used features: 96
used features: 96
COSTDIFF:0.320441
used features: 96
used features: 96
COSTDIFF:0.993633
used features: 89
used features: 89
COSTDIFF:0.234236
used features: 89
used features: 89
COSTDIFF:0.996468
used features: 72
used features: 72
COSTDIFF:0.148323
used features: 72
used features: 72
COSTDIFF:0.996312
used features: 70
used features: 70
COSTDIFF:0.177851
used features: 70
used features: 70
COSTDIFF:0.995406
used features: 57
used features: 57
COSTDIFF:0.160343
used features: 57
used features: 57
COSTDIFF:0.969409
used features: 43
used features: 43
COSTDIFF:0.125119
used features: 43
used features: 43
COSTDIFF:0.857258
used features: 43
used features: 43
COSTDIFF:0.866535
used features: 43
used features: 43
COSTDIFF:0.964017
used features: 36
used features: 36
COSTDIFF:0.0277237
used features: 36
used features: 36
COSTDIFF:0.830565
used features: 36
used features: 36
COSTDIFF:0.919677
used features: 31
used features: 31
COSTDIFF:0.0255904
used features: 31
used features: 31

This is the input to the terminal and it says that fewer and fewer feature points are utilized and the trajectory is poor

[gzyabc]

I think the poor trajectory may be the reason for the initialization. I used the same euroc data set to run vinsmono and vinsfusion, both of which could run normally, but the trajectory of dynavins ran out of control.