Nonlinear estimation method and system based on optical flow and IMU fusion

Abstract

The embodiment of the invention discloses a nonlinear estimation method and system based on optical flow and IMU fusion. The method comprises the steps: obtaining an optical flow, and obtaining a ratio of a linear speed to a distance of equipment from the optical flow; decomposing the optical flow, and fusing the decomposed optical flow with a measurement value of an airborne accelerometer to obtain a plane distance; obtaining the linear speed of the equipment through the plane distance; calculating a target estimation error convergence value; and performing scale estimation according to the target estimation error convergence value and the linear velocity to generate a scale estimation result. Compared with a classic EKF, the method and the system have the advantage that error convergenceis faster and more controllable. A nonlinear filter can generate an estimation error with a completely characterized convergence behavior, transient response can be applied by properly adjusting an estimation gain and equipment motion, and finally, compared with the EKF, a nonlinear scheme has remarkable superiority in the aspects of a convergence rate and predictability of estimation.

Description

technical field [0001] The invention relates to the technical field of optical communication, in particular to a nonlinear estimation method and system based on fusion of optical flow and IMU. Background technique [0002] In recent years, the visual-SLAM algorithm has been widely used in state estimation, providing a closed-form solution to calculate the metric scale factor by observing a single feature over time, and then completing the mapping map and location feature estimation. However, vision requires the problem of extracting and correlating data across multiple frames. As in most typical cases relying on monocular cameras must solve the problem of fusion of visual information and other on-board sensor data. In addition, visual-SLAM methods depend on the possibility of continuously tracking features for a longer period of time, and the visual system may be affected by unexpected occlusions or the need for intensive calculations. [0003] In the prior art, the state ...

AI Technical Summary

Problems solved by technology

[0005] In view of the above technical problems, the embodiment of the present invention provides a nonlinear estimation method and system based on optical flow and IMU fusion, which can solve the conventional classic EKF framework for state estimation of source vision and IMU data fusion in the prior art, but the EKF The scheme has two disadvantages that cannot be ignored: the linearization of nonlinear system dynamics is required, such as when dealing with visual-inertial estimation problems; the technical problem of making explicit description of estimation error behavior difficult

Images