Small target motion state estimation method in active sonar echo diagram

Abstract

The invention discloses a small target motion state estimation method in an active sonar echo diagram, and the method comprises the steps: firstly describing the edge constancy of the small target inthe sonar echo graph through Gaussian Laplacian of transform (LoG), so as to solve a problem that the target echo is not stable because of the non-rigid characteristics of the small target, then, forthe wake flow and cavitation noise problems generated by ship motion, introducing adaptive regularization parameters based on contextual information to regularize non-local items in a target function,and finally, carrying out warping processing on the acoustic flow field image and the edge detection image which are subjected to up-sampling to obtain an acoustic flow field image and an edge detection image, and obtaining a small target motion state estimation result in the continuous sonar echo image. According to the method, the influence of the non-rigid characteristic of the underwater moving small target on the motion estimation precision is effectively reduced, and meanwhile, the acoustic streaming estimation precision is improved when ship wake noise and cavitation noise exist in sonar echoes.

Description

technical field [0001] The invention belongs to the field of image processing, and in particular relates to a method for estimating an object's motion state. Background technique [0002] For a long time, motion estimation methods using motion flow fields such as optical flow have been widely used in many fields such as intelligent monitoring, remote sensing, and computer vision. In the last 30 years, optical flow and similar motion flow field estimation techniques have been increasingly used in the motion estimation of underwater targets. Among them, the literature that can be found includes: "Robust tracking of multiple objects in sector scansonar image sequences using optical flow motion estimation" published in "IEEE Journal of Oceanic Engineering" in 1998. The 2D sports field is used as prior information for target tracking and recognition; and, "Determining underwater vehicle movement from sonar data in relatively featurelessseafloor tracking missions" published in "I...

AI Technical Summary

Problems solved by technology

[0004] In order to overcome the deficiencies in the prior art, the present invention provides a method for estimating the motion state of a small target in an active sonar echo image. First, the small target in the sonar echo image is described by Laplacian of Gaussian (LoG) transform In order to solve the problem of unstable target echo caused by the non-rigid characteristics of small targets; then, for the wake and cavitation noise problems caused by ship motion, an adaptive regularization parameter based on context information is introduced Regularize the non-local terms in the objective function, and finally warping the up-sampled acoustic flow field image and the edge detection image to obtain the motion state estimation result of the small target in the continuous sonar echo image

Images