A Locally Optimal Subspace Recognition Method for One-Dimensional Range Profiles of Radar Targets
A local optimal, radar target technology, applied to radio wave measurement systems, instruments, etc., can solve problems such as local non-optimal, classification information loss, etc., to achieve the effect of improving recognition performance and classification performance
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Publication Date
- 2016-12-07
- Estimated Expiration
- Not applicable · inactive patent
Smart Images
Figure 1 Figure 2 Figure 3
Abstract
Description
technical field
[0001] The invention belongs to the technical field of radar target recognition, and relates to a local optimal subspace recognition method of a one-dimensional range image of a radar target. Background technique
[0002] The subspace method is a classic pattern recognition method, which is widely used in image recognition, face recognition, and has many applications in radar target recognition. The common characteristic subspace method and canonical subspace method have achieved good recognition results in radar target recognition. Among them, the feature subspace can maintain the energy of the original data in the low-dimensional feature space, but it is not optimal in terms of classification performance. Compared with the feature subspace, the regular subspace extracts target features by minimizing the intra-class distance and maximizing the inter-class distance, which improves the performance of target recognition to a certain extent.
[0003] However, ...
Examples
Embodiment Construction
[0025] The present invention will be described in further detail below in combination with specific implementation methods and accompanying drawings.
[0026] Locally optimal transformation subspace:
[0027] let x ij (n-dimensional column vector) is the i-th th The jth of the class target th training one-dimensional range images, i=1,2,…,g; j=1,2,…,N i , N 1 +N 2 +…+N g = N, where N i for i th The number of training one-dimensional range image samples for class targets, and N is the total number of training one-dimensional range image samples. define x ij The nearest in-class one-dimensional range image sample and the nearest out-of-class one-dimensional range image sample are
[0028] x ij W = arg min { x ir } | | ...