Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

A method for object recognition in sar images based on two-dimensional nonlinear projection features

A target recognition and non-linear technology, applied in character and pattern recognition, instruments, computer parts, etc., can solve the problems of reducing the final classification result, dimensionality disaster, and increasing the amount of calculation, so as to improve the accuracy of classification and improve Classification efficiency, effect of reducing feature dimension

Inactive Publication Date: 2017-03-29
UNIV OF ELECTRONICS SCI & TECH OF CHINA
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] SAR images have complex characteristics, and there is inevitably a certain correlation between the features extracted from the same target echo, and this correlation is often imperceptible. Redundant features will not only increase the amount of computation, but may also reduce the final classification. result
C.J.Enderli et al. used nonlinear KLDA to recognize targets in SAR images, and converted the two-dimensional SAR image matrix into one-dimensional vectors for processing, which will lose the spatial structure information of the target and is prone to the disaster of dimensionality.
Zhang et al. used two-dimensional LDA to extract the features of SAR image targets, but the nonlinear features of the image could not be obtained.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A method for object recognition in sar images based on two-dimensional nonlinear projection features
  • A method for object recognition in sar images based on two-dimensional nonlinear projection features
  • A method for object recognition in sar images based on two-dimensional nonlinear projection features

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0036] The specific implementation manners of the present invention will be described below in conjunction with the accompanying drawings.

[0037] Such as figure 1 Shown, the implementation process of the present invention is as follows:

[0038] Step 1. Select N pieces of m×n training sample SAR image A according to the target data 1 ,A 2 ,...,A N , belonging to c categories, N 1 ,N 2 ,...,N c Indicates the sample size of each class, N 1 +N 2 +...+N c =N.

[0039] Step 2.1, express Al in columns as l=1,2,...,N, means A l The kth column of , k=1,2,...,n. to A l After nonlinear mapping φ, the image samples in the kernel space H are obtained as Then the inter-class discrete matrix and the intra-class discrete matrix in the high-dimensional space are

[0040]

[0041]

[0042] It can be seen from formulas (1) and (2) that its essence is to take corresponding operations on the columns of the image. Select an appropriate kernel function, for a certain col...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention belongs to the technical field of Synthetic Aperture Radar (SAR, Synthetic Aperture Radar) automatic target recognition, in particular to SAR image target recognition of two-dimensional nonlinear projection features. The specific steps of the present invention are as follows: S1, determine the SAR image training sample matrix; S2, determine the kernel function and the kernel sample matrix, determine the scatter matrix between the nuclear classes and the scatter matrix in the kernel through the kernel sample matrix, and pass the kernel class Scatter matrix and kernel inner scatter matrix construct objective criterion function, obtain projection matrix and projection subspace; S3, determine the nonlinear projection characteristic subset of input SAR image, determine the distance of nonlinear projection characteristic subimage and projection subspace, determine Enter the category to which the SAR image target belongs. The present invention uses a certain column vector of all training samples to construct a kernel vector, cleverly constructs a kernel sample matrix, and adopts a projection feature extraction method in a high-dimensional space, which improves the classification efficiency and improves the classification accuracy while reducing the dependence on samples. decreased.

Description

technical field [0001] The invention belongs to the technical field of Synthetic Aperture Radar (SAR, Synthetic Aperture Radar) automatic target recognition, in particular to SAR image target recognition of two-dimensional nonlinear projection features. Background technique [0002] The principle of SAR image target recognition is to establish a feature library based on the known training sample target category information, perform feature extraction on the test sample, and select the type of training sample corresponding to the highest similarity in the library as the classification result of the test sample. [0003] The rapid development of SAR technology has greatly improved the resolution of the resulting image, and the target information in the SAR image has also shown explosive growth, which has brought about a substantial increase in the amount of corresponding data. Facing the huge amount of data, Key technologies in object detection and recognition must be improved...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/66G06K9/46
Inventor 周代英田兵兵谭敏洁谭发曾贾继超余为知
Owner UNIV OF ELECTRONICS SCI & TECH OF CHINA
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products