Visible light visual navigation method in smog environment

Abstract

The invention discloses a visible light visual navigation method in a smog environment. The method comprises the steps of calibrating a camera; collecting a continuous frame data image through the camera, and judging whether the image is foggy or not; if it is judged that fog exists, estimating atmospheric scattering coefficients, and carrying out defogging processing on the image; calculating a relative pose transformation relation between the defogged current image and the image; determining a matching point according to the relative pose transformation relation, and fusing the atmospheric scattering coefficients obtained by the image at the matching point to obtain an atmospheric scattering coefficient estimated value of the matching point; and according to the atmospheric scattering coefficient estimated value, carrying out defogging processing on the next image of the image again, repeating the above steps, when the atmospheric scattering coefficient estimated value satisfies a preset condition, stopping executing the above steps, and outputting a pose estimation result obtained at the last time. The technical problem that an existing visual navigation scheme is not applicable in the smog environment in the prior art is solved.

Description

technical field [0001] The application of this application relates to the technical fields of computer vision and navigation. Background technique [0002] Visual navigation has important applications in the fields of high-precision guided weapons and automatic driving. Natural environment interference such as smoke will seriously affect the visual navigation effect. [0003] There are many methods for image dehazing. For example, the dark channel prior dehazing method based on the atmospheric scattering model can achieve good results. However, image defogging does not consider the continuity of the navigation process, and directly joins the visual navigation process, which makes it difficult to guarantee the robustness and accuracy of navigation. Imaging with an infrared camera can eliminate the influence of smoke, but the resolution of the infrared camera is low and the image quality is poor. [0004] These image defogging methods focus on the visual effects of defogging...

AI Technical Summary

Problems solved by technology

[0007] For this reason, the first purpose of this application is to propose a visible light visual navigation method in a smoky environment, which solves the problem of image defogging in existing methods without considering the continuity of the navigation process, directly adding the visual navigation process, and improving navigation robustness. and accuracy are difficult to guarantee, by fully considering the continuity of visual navigation, the accuracy of visual navigation can be greatly improved in misty environments

Images