Visual and Visual-Inertial SLAM: State of the Art, Classification, and Experimental Benchmarking

Table 2

Comparative classification of main vSLAM and viSLAM methods. The algorithms adapted to pedestrian navigation applications are presented in bold.

Algorithm map gestion

Hardware requirements

Approach

Input treatment

Localis./Mapping

Memory loop

Monoc.

Stereo

Depth

IMU

Filter

Optim.

Direct

Indir.

2D-2D

3D-2D

IMU

closure

MonoSLAM [21]

X

X

Sparse

X

X

Monocular FastSLAM [22]

X

X

Sparse

X

X

PTAM [27]

X

X

Sparse

X

X

PTAM with edgelets [65]

X

X

Sparse

X

X

PTAM with DWO [79]

X

X

Sparse

X

X

X

Stereo PTAM [78]

X

X

Sparse

X

X

X

X

CD-SLAM [80]

X

X

Sparse

X

X

X

ORB-SLAM [37]

X

X

Sparse

X

X

X

X

X

ORB-SLAM2 [76]

X

(X)

(X)

X

Sparse

X

X

X

X

X

Edge-SLAM [81]

X

X

Sparse

X

X

X

X

DTAM [34]

X

X

Dense

X

(X)

X

MobileFusion [66]

X

(X)

X

Dense

X

X

(X)

X

Semidense visual odom. [5]

X

X

Semidense

X

X

X

LSD-SLAM [35]

X

X

Semidense

X

X

X

Semidirect VO (SVO) [67]

X

X

Sparse

X

X

X

X

X

Direct sparse odom. (DSO) [33]

X

X

Sparse

X

X

X

KinectFusion [68]

X

X

Dense

X

X

Kintinuous [82]

(X)

X

X

Dense

X

X

X

DVO SLAM [69]

X

X

X

Dense

X

X

X

ElasticFusion [70]

X

X

X

Dense

X

X

X

X

MSCKF [25]

X

X

X

None

X

X

X

X

MSCKF 2.0 [45]

X

X

X

None

X

X

X

X

ROVIO [26]

X

(X)

X

X

None

X

X

X

X

X

OKVIS [73]

(X)

X

X

X

Sparse

X

X

X

X

X

S-MSCKF [17]

X

X

X

None

X

X

X

X

Vins-Mono [74]

X

X

X

Sparse

X

X

X

X

X

Kimera [60]

(X)

X

X

X

X

Dense

X

X

X

X

X

X

SOFT-SLAM [72]

X

(X)

X

Dense

X

X

X

(X)

X

X

STCM-SLAM [77]

X

X

X

Sparse

X

X

X

X

X

VIORB [75]

X

X

X

Sparse

X

X

X

X

X