Feature point tracking method based on event camera

Abstract

The invention discloses a feature point tracking method based on an event camera, and aims to improve the feature point tracking precision. According to the technical scheme, the feature point tracking system based on the event camera is composed of a data acquisition module, an initialization module, an event set selection module, a matching module, a feature point updating module and a templateedge updating module. The initialization module extracts feature points and an edge graph from the image frame. The event set selection module selects an event set S of the feature points from the event flow around the feature point. The matching module matches the S with the template edge around the feature points to solve an optical flow set Gk of n feature points at the moment tk. The feature point updating module calculates a position set FDk + 1 of the n feature points at the tk + 1 moment according to the Gk, and the template edge updating module updates the positions in the PBDk by using IMU data to obtain a position set PBDk + 1 of template edges corresponding to the n feature points at the tk + 1 moment. By adopting the method, the precision of tracking the feature points on the event stream can be improved, and the average tracking time of the feature points can be prolonged.

Description

technical field [0001] The invention relates to the field of computer image processing, in particular to using an event camera to track feature points in an image. Background technique [0002] SLAM (full name Simultaneous Localization and Mapping), as an important branch in the field of robotics, has been extensively studied in recent years. SLAM tries to solve such a problem: a robot moves in an unknown environment, how to determine its own trajectory through observation of the environment, and at the same time construct a map of the environment. SLAM technology is the sum of many technologies involved in achieving this goal. A complete SLAM system mainly includes the front-end visual odometry part and the back-end optimization part. The visual odometry part estimates the state of the robot, and the estimation methods are mainly divided into two methods: feature point method and direct method. The feature point method is currently the mainstream method for robot state e...

AI Technical Summary

Problems solved by technology

Using this method, the position of the feature point is updated every time an event is received, which increases the amount of calculation, and this method will introduce more errors, resulting in poor tracking accuracy of feature points, and long-term tracking cannot be achieved.

[0006] Therefore, the existing methods of using event cameras for feature point tracking still have the disadvantages of low tracking accuracy and short tracking time.

Images