FAST-LIVO

Introduction

FAST-LIVO: a fast LiDAR-Inertial-Visual Odometry system, which builds on two tightly-coupled and direct odometry subsystems: a VIO subsystem and a LIO subsystem. The LIO subsystem registers raw points (instead of feature points on e.g., edges or planes) of a new scan to an incrementally-built point cloud map. The map points are additionally attached with image patches, which are then used in the VIO subsystem to align a new image by minimizing the direct photometric errors without extracting any visual features (e.g., ORB or FAST corner features). To further improve the VIO robustness and accuracy, a novel outlier rejection method is proposed to reject unstable map points that lie on edges or are occluded in the image view. The system supports both multi-line spinning LiDARs and emerging solid-state LiDARs with completely different scanning patterns, and can run in real-time on both Intel and ARM processors. Here are some reference links: code link. paper link.

Evaluation

Platforms	Sequences	Length(m)	ATE
A Handheld Device	Escalator	77.460	2.390
	MCR normal dark	76.499	4.560
	MCR aggressive 6dof light	100.871	35.731
	Parkway loop night	461.049	26.556
	Forest	130.937	X
A UGV	Elevator	39.336	X
	Indoor loop	270.674	7.130
	MCR hdr	193.918	0.0990
	Street day	2064.475	60.110
	Parkway loop night	461.051	11.455
A QR	Underground	98.312	0.790
	MCR hdr	85.08	0.656
	Forest	108.037	5.231
A UAV	MCR loop light	104.989	LiDARs not supported
A Car	Urban night loop	1807.884	2.831