Camouflaged target recognition method based on k-means clustering and polarization information extraction

Abstract

The invention discloses a camouflaged target identification method based on K-means cluster and polarization information extraction. Target polarization parameter image data can be acquired, and then a polarization degree image, a polarization angle image, and an ellipticity angle image can be acquired, and then the clustering of the acquired multi-dimensional images I (pi, theta, epsilon) can be carried out by adopting the K-means, and then the new cluster center Cmi can be calculated, and in addition, the new cluster center can be calculated repeatedly until the change of the cluster center Cmi is stopped, and the iteration is terminated, and the final cluster fusion image can be acquired. The camouflaged target identification method is advantageous in that the subjective vision and the objective evaluation can be improved greatly, and because the cluster fusion result is provided with the polarization degree information, the polarization angle information, and the ellipticity angle information in an integrated manner, the polarization information of the image can be more completed.

Description

technical field [0001] The invention belongs to the technical field of polarization imaging, in particular to a camouflage target recognition method based on K-means clustering and polarization information extraction. Background technique [0002] Artificial target detection in natural background is a hotspot in the field of target recognition, and detecting camouflaged targets in natural background has always been a difficult problem in this field. The research on polarization imaging detection at home and abroad has found that the polarization state information of the reflected or radiated light waves of objects can effectively distinguish scatterers with different materials, different surface shapes, and different conductivity. Under certain conditions, there is a large difference between the polarization information of the camouflaged target obtained by polarization imaging and the polarization information of the natural background. The surface of the artificial camoufl...

AI Technical Summary

Problems solved by technology

[0009] The purpose of the present invention is to provide a camouflage target recognition method based on K-means clustering and polarization information extraction, which improves the problem of partial polarization information missing in the existing polarization image fusion process, and the polarization information of the fused image is more complete, and the details With richer features, camouflage targets can be better highlighted

Images