A feature extraction method of one-dimensional range profile for non-linear discriminative learning of true and false targets

Abstract

The invention discloses a one-dimensional range image feature extraction method for nonlinear discrimination learning true and false targets, belonging to the technical field of radar target recognition. The present invention first utilizes the nonlinear function to map the one-dimensional range image to the high-dimensional feature space, then obtains the projection transformation matrix through discriminant learning in the high-dimensional feature space, and then obtains the one-dimensional range image of any feature to be extracted based on the projection transformation matrix eigenvectors of . The feature extraction method of the present invention can well represent the nonlinearity in the sample data distribution, thereby improving the target recognition performance, overcoming the disadvantage that the existing conventional subspace method is not suitable for nonlinear data distribution, and effectively improving the Classification performance for real and fake radar targets.

Description

technical field [0001] The invention belongs to the technical field of radar target recognition, and in particular relates to a one-dimensional distance image feature extraction method for nonlinear discrimination learning true and false targets which can be used for radar target recognition. Background technique [0002] Radar target recognition needs to extract the relevant information signs and stable features (target features) of the target from the radar echo of the target and determine its attributes. It identifies targets based on their electromagnetic backscatter. Using the characteristics of the scattered field generated by the target in the far zone of the radar, information for target identification (target information) can be obtained. The acquired target information is processed by computer and compared with the characteristics of existing targets, so as to achieve the purpose of automatic target identification. Radar target recognition includes two parts: fea...

AI Technical Summary

Problems solved by technology

[0008] However, the conventional subspace method is a linear method, which is only suitable for the case where the sample data distribution is linear. In practice, the sample data distribution may appear nonlinear, resulting in a significant decline in the recognition performance of the conventional subspace method.

There is room for further improvement in the recognition performance of existing conventional subspace methods

Images