Image extraction method for ultra-wideband step-frequency continuous wave underground target detection system

Abstract

The invention discloses an ultra-wideband step-frequency continuous wave underground target detection system image extraction method, which adopts an underground target detection system, the underground target detection system comprises an upper computer, a multi-port vector network analyzer and a directional antenna, a step-frequency continuous wave signal system is utilized, a single-transmitting multi-receiving signal acquisition mode is adopted, and a multi-port vector network analyzer and a directional antenna are combined to obtain an image of an ultra-wideband step-frequency continuous wave underground target detection system. And obtaining a hidden target echo signal, designing an image edge detection algorithm based on improved phase consistency, and extracting a final edge image. According to the invention, the robustness of a test system to noise and image amplitude change is effectively improved, high signal-to-noise ratio data acquisition is realized, and a technical means is provided for high-resolution and low-noise system sampling. The hyperbolic characteristic of the underground target test image is specifically utilized, the target identification and detection performance in different scenes is improved, high-precision rapid imaging of the buried target in a complex environment can be realized, and technical support is provided for complex city environment perception and hidden target detection.

Description

Technical Field

[0001] The present invention belongs to the technical field of underground target detection, and in particular relates to an image extraction method for an ultra-wideband stepped-frequency continuous-wave underground target detection system. Background Art

[0002] Underground target detection systems detect underground areas by transmitting high-frequency electromagnetic waves to the surface and receiving return signals. Due to the varying electromagnetic properties of underground targets and transmission media, the system's echoes exhibit distinct characteristics. Traditional methods for identifying and interpreting underground target images are largely limited to the experience of professionals, hindering the advancement of large-scale underground target detection capabilities.

[0003] To achieve high-resolution, rapid identification and detection of buried targets, researchers both domestically and internationally have conducted extensive research. However, several challenges have arisen. First, far-field target detection capabilities are insufficient. Underground media vary in complexity, with a wide variety of propagation media and varying degrees of electromagnetic wave loss. Electromagnetic waves also reflect and scatter in the inhomogeneous underground medium, making target identification and imaging reliant on experience rather than a systematic approach. Second, the echo signal's signal-to-noise ratio is low. When the electromagnetic characteristics of the target and its surroundings are similar, the target signal is easily submerged in clutter and noise. Furthermore, because the fluctuations of underground stratification are random and uncertain, accurate modeling is impossible. Third, image matching and detection are challenging. Common underground targets have few image features, making common matching frameworks difficult to apply to underground target detection images and resulting in low matching success rates.

[0004] The present invention aims to develop an image extraction method for an ultra-wideband stepped-frequency continuous wave underground target detection system, solve the problem of high-precision imaging of hidden targets underground in complex environments, and provide technical support for complex urban environment perception and hidden target detection. Summary of the Invention

[0005] Aiming at the demand for high-resolution and rapid imaging of hidden targets, the present invention proposes an image extraction method for an ultra-wideband stepped-frequency continuous wave underground target detection system. The method has a reasonable design, solves the shortcomings of the existing technology, and has good effects.

[0006] An image extraction method for an ultra-wideband stepped-frequency continuous wave underground target detection system is proposed. The underground target detection system includes a host computer, a multi-port vector network analyzer, and a directional antenna. The stepped-frequency continuous wave signal system is used to obtain the hidden target echo signal through a single-transmit multiple-receive signal acquisition mode. An image edge detection algorithm based on improved phase consistency is designed to extract the final edge image.

[0007] Furthermore, the host computer sends commands to the transmitter and receiver of the multi-port vector network analyzer. When the transmitter receives the command from the host computer, it further sends commands to the transmitting antenna to transmit broadband electromagnetic waves underground.

[0008] After receiving the instructions from the receiver, the receiving antenna performs data acquisition. After receiving the echo signal of the target, it transmits the echo signal to the signal processing unit of the host. The signal processing unit will display the image of the detected imaging system according to the data type and format received.

[0009] Furthermore, the image edge detection algorithm based on improved phase consistency includes a preprocessing stage, a multi-scale phase consistency calculation stage, a fusion optimization stage and a post-processing stage;

[0010] The preprocessing stage performs noise reduction and contrast enhancement on the original image to construct high-quality input for frequency domain decomposition;

[0011] The multi-scale phase consistency calculation stage extracts sub-bands in each direction in the frequency domain through the Log-Gabor filter, calculates the local energy in combination with the Hilbert transform, and normalizes it to generate a preliminary phase consistency map;

[0012] In the fusion optimization stage, directional entropy is used to dynamically allocate weights to obtain spatial three-dimensional detection results, and broken edges are connected through morphological closing operations. At the same time, adaptive thresholds are generated based on the Otsu algorithm to complete binarization.

[0013] The post-processing stage imposes geometric constraints based on the horizontal scanning characteristics of the image.

[0014] Furthermore, the geometric constraints include suppressing short edges in the vertical direction and combining a contextual verification mechanism of neighborhood energy mutation to distinguish between real reflection interfaces and random interference.

[0015] Beneficial technical effects brought about by the present invention:

[0016] The present invention proposes an image extraction method for an ultra-wideband stepped-frequency continuous wave underground target detection system, which improves the system's far-field target testing capability and solves the problems of rapid image matching and target recognition under low signal-to-noise ratio.

[0017] (1) Effectively improve the robustness of the test system to noise and image amplitude changes, achieve high signal-to-noise ratio data acquisition, and provide technical means for high-resolution, low-noise system sampling.

[0018] (2) Targeted use of the hyperbolic characteristics of underground target test images can improve the target recognition and detection performance in different scenarios, and can achieve high-precision and rapid imaging of buried targets in complex environments, providing technical support for complex urban environment perception and hidden target detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 Schematic diagram of the structure of the underground target detection system in the present invention;

[0020] Figure 2 This is a flow chart of the image edge detection algorithm for improving phase consistency in the present invention; DETAILED DESCRIPTION

[0021] The specific implementation of the present invention will be further described below with reference to specific embodiments:

[0022] An image extraction method for an ultra-wideband stepped-frequency continuous wave underground target detection system, such as Figure 1 As shown, an underground target detection system is adopted. The underground target detection system includes a host computer, a multi-port vector network analyzer and a directional antenna. The stepped frequency continuous wave signal system is used to obtain the hidden target echo signal through the single-transmitter and multiple-receiver signal acquisition mode. An image edge detection algorithm based on improved phase consistency is designed to extract the final edge image.

[0023] The host computer sends commands to the transmitter and receiver of the multi-port vector network analyzer. When the transmitter receives the command from the host computer, it further issues instructions to the transmitting antenna to emit broadband electromagnetic waves underground.

[0024] Due to the complex and diverse underground environment and media, the path and speed of electromagnetic wave propagation are affected. Signal transmission and acquisition are performed simultaneously. The test system continuously sends commands to the receiver. After receiving the receiver's instructions, the receiving antenna performs data acquisition. After receiving the target's echo signal, it transmits the echo signal to the host's signal processing unit. The signal processing unit displays the image of the detected imaging system based on the received data type and format.

[0025] Because electromagnetic wave propagation exhibits different electromagnetic properties in different underground media, and because electromagnetic properties vary between different media, these differences can be exploited to further analyze the spatial distribution and morphological characteristics of target signals based on their physical properties and the radar image displayed by the echo. The target image of an underground target detection system exhibits a hyperbolic shape, with the apex of the hyperbola corresponding to the projected position of the target center. The radius of curvature of the hyperbola is determined by the burial depth and the relative dielectric constant of the medium. This physical property allows the inversion of target depth and size parameters by fitting the hyperbola equation, forming the basis for quantitative analysis of underground target test data.

[0026] This image edge detection algorithm based on improved phase congruency integrates frequency domain analysis with adaptive optimization strategies, focusing on addressing the traditional method's sensitivity to low-contrast targets and strong noise interference. Its core concept is to leverage phase congruency, combined with multi-scale and multi-directional decomposition techniques, to accurately extract local features of underground structures from a frequency domain perspective. Traditional phase congruency algorithms suffer from insufficient directional coverage, weak noise suppression, and strong threshold dependence in complex scenarios. This improved method leverages dynamic weight fusion and nonlinear optimization to effectively suppress spurious responses while preserving weak edge features.

[0027] like Figure 2 As shown in the figure, the specific process of the algorithm includes the preprocessing stage, the multi-scale phase consistency calculation stage, the fusion optimization stage and the post-processing stage;

[0028] In the preprocessing stage, the original image is subjected to noise reduction and contrast enhancement to construct high-quality input for frequency domain decomposition. Wavelet denoising is used for noise reduction, and CLAHE histogram equalization is used for contrast enhancement.

[0029] In the multi-scale phase consistency calculation stage, the Log-Gabor filter is used to extract the sub-bands in each direction in the frequency domain, and the local energy is calculated by combining the Hilbert transform, and then normalized to generate a preliminary phase consistency map.

[0030] In the fusion optimization stage, directional entropy is used to dynamically assign weights to obtain spatial three-dimensional detection results. The broken edges are connected through morphological closing operations, and an adaptive threshold is generated based on the Otsu algorithm to complete the binarization.

[0031] The post-processing stage imposes geometric constraints based on the horizontal scanning characteristics of the image.

[0032] The geometric constraints include suppressing short vertical edges and combining a contextual verification mechanism based on neighborhood energy mutation to distinguish real reflection interfaces from random interference.

[0033] The key technical improvements of the algorithm are reflected in three aspects. First, multi-scale decomposition is achieved through a Log-Gabor filter bank with higher frequency domain coverage. The filter parameters are designed to cover 3 to 5 scale layers and 8 evenly distributed directions. This not only adapts to the multi-scale characteristics of reflected waves from targets such as underground pipelines and rock formations, but also captures the structural trend characteristics at different angles. Second, a local energy correction strategy is introduced in the phase consistency calculation. The weights of each frequency band are dynamically adjusted according to the real-time estimated signal-to-noise ratio. For example, an adaptive threshold is set by the background noise power spectrum, which significantly improves the robustness to high-frequency random noise. Third, a directional consistency optimization mechanism is used to perform joint probability analysis on multi-directional phase responses. Geometric constraints are used to remove artifacts, ensuring edge continuity while avoiding fractures. In the preprocessing stage, the algorithm integrates wavelet threshold denoising and contrast-constrained histogram equalization technology to eliminate typical high-frequency oscillation noise. The latter enhances the difference in weak signals between the layered interface and the pipeline through local histogram stretching.

[0034] Of course, the above description is not a limitation of the present invention, and the present invention is not limited to the above examples. Changes, modifications, additions or substitutions made by technicians in this technical field within the essential scope of the present invention should also fall within the scope of protection of the present invention.

Claims

1. An image extraction method for an ultra-wideband stepped-frequency continuous wave underground target detection system, characterized in that: An underground target detection system is adopted, which includes a host computer, a multi-port vector network analyzer and a directional antenna. The stepped frequency continuous wave signal system is used to obtain the hidden target echo signal through the single-transmitter multiple-receiver signal acquisition mode. An image edge detection algorithm based on improved phase consistency is designed to extract the final edge image.

2. The image extraction method of an ultra-wideband stepped-frequency continuous wave underground target detection system according to claim 1, characterized in that: The host computer sends commands to the transmitter and receiver of the multi-port vector network analyzer. When the transmitter receives the command from the host computer, it further issues instructions to the transmitting antenna to emit broadband electromagnetic waves underground. After receiving the instructions from the receiver, the receiving antenna performs data acquisition. After receiving the echo signal of the target, it transmits the echo signal to the signal processing unit of the host. The signal processing unit will display the image of the detected imaging system according to the data type and format received.

3. The image extraction method of an ultra-wideband stepped-frequency continuous wave underground target detection system according to claim 2, characterized in that: The image edge detection algorithm based on improved phase consistency includes a preprocessing stage, a multi-scale phase consistency calculation stage, a fusion optimization stage, and a post-processing stage; The preprocessing stage performs noise reduction and contrast enhancement on the original image to construct high-quality input for frequency domain decomposition; The multi-scale phase consistency calculation stage extracts sub-bands in each direction in the frequency domain through the Log-Gabor filter, calculates the local energy in combination with the Hilbert transform, and normalizes it to generate a preliminary phase consistency map; In the fusion optimization stage, directional entropy is used to dynamically allocate weights to obtain spatial three-dimensional detection results, and broken edges are connected through morphological closing operations. At the same time, adaptive thresholds are generated based on the Otsu algorithm to complete binarization. The post-processing stage imposes geometric constraints based on the horizontal scanning characteristics of the image.

4. The image extraction method of an ultra-wideband stepped-frequency continuous wave underground target detection system according to claim 3, characterized in that: The geometric constraints include suppressing short vertical edges and combining a contextual verification mechanism based on neighborhood energy mutation to distinguish real reflection interfaces from random interference.

Images