Unlock instant, AI-driven research and patent intelligence for your innovation.

Through-wall radar imaging method based on convolutional neural network

A technology of convolutional neural network and wall-penetrating radar, which is applied in the field of wall-penetrating radar imaging based on convolutional neural network, can solve the problems of increased false alarm rate of target detection, difficulty in target detection and recognition, and affecting imaging quality, etc. Intuitive handling, enhanced image quality, and enhanced image quality effects

Inactive Publication Date: 2020-08-14
XI'AN PETROLEUM UNIVERSITY
View PDF9 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, due to the existence of the wall, the electromagnetic wave propagation path changes, the signal energy attenuation is serious, and there is strong wall clutter in the echo signal. In addition, the imaging algorithm of the traditional imaging algorithm will generate strong The grating lobe effect, the strong grating lobe seriously affects the imaging quality, causing the target to be submerged in the strong grating lobe, which brings difficulties to the subsequent target detection and recognition, and increases the false alarm rate of target detection

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
  • Through-wall radar imaging method based on convolutional neural network
  • Through-wall radar imaging method based on convolutional neural network
  • Through-wall radar imaging method based on convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0052] The present invention will be further described in detail below with reference to the drawings and specific embodiments, but the embodiments of the present invention are not limited thereto.

[0053] In order to solve the problem of low imaging quality of through-wall radar, the present invention designs a through-wall radar imaging method based on convolutional neural network, such as figure 1 Shown.

[0054] Step 1: Training data generation.

[0055] The training data is the prior knowledge of the imaging model, which includes two parts: input echo data and output reference data. In the present invention, figure 2 The imaging geometric model shown is an object, and corresponding training data is generated.

[0056] Step 1-1: Input echo data generation.

[0057] The invention adopts a stepped frequency synthetic aperture system, the radar transceiver co-located antenna moves along a track parallel to the wall surface at a constant speed, the number of radar measurement positi...

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 discloses a through-wall radar imaging method based on a convolutional neural network. The method comprises the steps of generating training data which comprise two parts of input echodata and output reference data; the convolutional neural network being a multi-layer network and comprising two activation functions, namely, designing a convolutional layer activation function in a convolutional layer, and designing an output layer activation function in an output layer; constructing a loss function according to the output reference data and the activation function; constructinga convolutional neural network structure for through-wall imaging through the training data, the activation function and the loss function in combination with the convolutional neural network technology; and inputting the training data into the convolutional neural network structure, training the convolutional neural network structure, then inputting echo data to be imaged, and carrying out an imaging test. According to the method, the deep learning network is introduced into the imaging process instead of image post-processing, so that the imaging quality is improved, and grating lobes are suppressed.

Description

Technical field [0001] The invention relates to the technical field of through-wall radar imaging, in particular to a through-wall radar imaging method based on a convolutional neural network. Background technique [0002] In the field of through-wall radar technology, traditional imaging algorithms, such as backward projection algorithms, can be used to obtain high-resolution target images. However, due to the existence of the wall, the electromagnetic wave propagation path changes, the signal energy is attenuated seriously, and there is strong wall clutter in the echo signal. In addition, the imaging algorithm of the traditional imaging algorithm will produce strong energy during the process of energy accumulation. The strong grating lobes seriously affect the imaging quality, causing the target to be submerged in the strong grating lobes, which will cause difficulties in subsequent target detection and recognition, and increase the false alarm rate of target detection. Summar...

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 Applications(China)
IPC IPC(8): G01S13/88G01S13/90G01S7/41
CPCG01S13/888G01S13/9021G01S13/9094G01S13/006G01S7/417G01S7/418G01S7/414G01S7/411
Inventor 雒明世方阳冯建利段沛沛闫效莺
Owner XI'AN PETROLEUM UNIVERSITY