Orientation sensitive target detection method based on sub-aperture color image saturation characteristics

Abstract

The invention discloses an SAR image orientation sensitive target detection method based on sub-aperture image saturation characteristics, and belongs to the technical field of synthetic aperture radar image target detection. The azimuth sensitive target detection is realized through the following steps: 1) sub-aperture image generation: dividing an azimuth frequency spectrum into a plurality of sub-bands, and generating a plurality of sub-aperture images through inverse Fourier transform; 2) color synthesis and feature extraction: allocating different hues to each sub-aperture image by using an HSV color space, synthesizing an RGB color image, converting the RGB color image to the HSV color space, and extracting a saturation channel in the RGB color space as an azimuth sensitivity feature map; and 3) target detection: carrying out threshold segmentation on the saturation feature map, and identifying a pixel region with a high saturation value, namely, an orientation sensitive target. According to the method, the color saturation change caused by the scattering difference of the target in different sub-apertures is utilized, effective detection of azimuth sensitive targets such as artificial buildings and vehicles is achieved, and the method has the advantages of being simple in calculation and high in robustness.

Description

Technical Field

[0001] This invention belongs to the field of SAR image processing and target detection technology, specifically a method for azimuth-sensitive target detection based on the saturation features of sub-aperture color images. Background Technology

[0002] Synthetic Aperture Radar (SAR) is an all-weather, all-day microwave imaging system. In SAR images, different targets exhibit varying scattering characteristics. Man-made targets, such as buildings, bridges, and vehicles, often display strong azimuth sensitivity due to their regular geometric structures. This means that even a small change in the radar's observation angle can cause significant fluctuations in the backscattered energy of these targets. In traditional intensity images, the average scattering intensity of these targets may be similar to that of surrounding ground features, leading to detection difficulties and a high false alarm rate. Furthermore, traditional SAR image target detection algorithms are largely based on intensity and texture information, making it difficult to effectively utilize the target's azimuth sensitivity to improve the detection capability of specific targets. Summary of the Invention

[0003] The purpose of this invention is to provide a location-sensitive target detection method based on the saturation features of sub-aperture color images.

[0004] To achieve the above objectives, the present invention employs the following steps: A method for detecting orientation-sensitive targets based on the saturation features of sub-aperture color images, comprising:

[0005] Step 1: Use single-channel SAR image data to perform sub-aperture segmentation and sub-band image generation;

[0006] Step 2: Based on the color mapping matrix, weighted composite of all sub-aperture images into a single RGB color image;

[0007] Step 3: Convert the RGB color image to the HSV color space, and perform saturation feature extraction and target detection based on the HSV color space.

[0008] Preferably, the specific method for sub-aperture division and sub-band image generation using single-channel SAR image data is as follows:

[0009] A Fourier transform is performed on the single-channel SAR image data along the azimuth direction to obtain the full aperture spectrum;

[0010] The full aperture spectrum is uniformly divided into Individual belt;

[0011] Perform an inverse Fourier transform on the spectrum of each sub-band to generate multiple sub-aperture images.

[0012] Preferably, the specific method for weighted synthesis of all sub-aperture images into a single RGB color image based on the color mapping matrix is ​​as follows:

[0013] Generate an HSV color mapping matrix using MATLAB's built-in hsv function. The HSV color mapping matrix contains A different color;

[0014] The sub-aperture images are normalized, and the normalized sub-aperture images are weighted and combined into an RGB color image.

[0015] Preferably, the RGB color image specifically comprises:

[0016]

[0017] in Corresponding to R, G, and B channels, This is the normalized sub-aperture image. For sub-aperture index, For the number of children, This is the HSV color mapping matrix.

[0018] Preferably, the RGB color image is converted to the HSV color space, and the specific method for saturation feature extraction and target detection based on the HSV color space is as follows:

[0019] synthesized RGB image Converting to the HSV color space yields three components: hue (H), saturation (S), and lightness (V).

[0020] The saturation map is processed using an adaptive thresholding method to determine the final binary detection mask image as follows:

[0021]

[0022] In this scenario, all pixels in the saturation map that are greater than a threshold are marked as 1, and the rest as 0. , Saturation quantiles: In a binary detection mask, the set of pixels with a value of 1 represents the orientation-sensitive target region.

[0023] Compared with the prior art, the significant advantage of this invention is that it does not rely on complex models and identifies targets by analyzing the color saturation of the synthesized sub-aperture images. Attached Figure Description

[0024] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0025] The following details the implementation steps of the present invention:

[0026] An azimuth-sensitive target detection method based on the saturation features of sub-aperture color images separates the scattering response of a target at different azimuth angles by dividing the spectrum into multiple sub-bands and imaging them separately. When these significantly different sub-aperture responses are mapped to different color channels and synthesized into a single color image, the azimuth-sensitive target will exhibit vibrant, highly saturated colors, while the background region with uniform azimuth scattering will appear as a low-saturation gray tone. Therefore, the saturation component of the image can serve as a powerful feature for distinguishing azimuth-sensitive targets from the background. The specific steps are as follows:

[0027] Step 1: Sub-aperture segmentation and sub-aperture image generation are performed using single-channel SAR image data. This step aims to decompose the azimuth scattering information into multiple sub-aperture images. The specific steps are as follows:

[0028] First, perform a Fourier transform on the single-channel SAR image data along the azimuth direction to obtain the full aperture spectrum:

[0029]

[0030] in, This indicates that a Fourier transform is performed in the azimuth direction. Full aperture spectrum, To input single-channel complex SAR data, The azimuth length, The distance is the length;

[0031] The full aperture spectrum is uniformly divided into There are 1 sub-band, and the length of each sub-band is:

[0032]

[0033] In the formula, For the effective spectrum length, , This is the oversampling rate, typically around 1.2. For spectral trimming length, Take the integer part;

[0034] The frequency domain interval for each sub-band is:

[0035]

[0036] in, For sub-aperture index;

[0037] Finally, the spectrum of each sub-band Perform inverse Fourier transform to generate multiple sub-aperture images:

[0038]

[0039] in, Indicates the inverse Fourier transform. These are the pixel coordinates of the image.

[0040] Step 2: Based on the color mapping matrix, weighted composite of all sub-aperture images into a single RGB color image, specifically:

[0041] Orientation-sensitive targets in different sub-aperture images The saturation of objects within a scene can change, while the saturation of background objects remains relatively stable. This invention utilizes this difference to convert it into saturation features through color synthesis.

[0042] First, use MATLAB's built-in hsv function to generate an HSV color mapping matrix. The HSV color mapping matrix contains A few different colors:

[0043]

[0044] in It is The matrix, , is the sub-aperture index. Corresponding to the R, G, and B channels. Each row corresponds to RGB values ​​evenly distributed in the HSV color space.

[0045] The sub-aperture images are normalized, and then the normalized sub-aperture images are weighted and combined into a single RGB color image. :

[0046]

[0047] in Corresponding to R, G, and B channels, This is a normalized sub-aperture image. , which is the sub-aperture index.

[0048] Step 3: Saturation Feature Extraction and Target Detection

[0049] synthesized RGB image Converting to the HSV color space yields three components: hue (H), saturation (S), and lightness (V).

[0050]

[0051] High saturation values ​​correspond to strong orientation sensitivity. An adaptive thresholding method is used to process the saturation map. The final binary detection mask image is as follows:

[0052]

[0053] In this scenario, all pixels in the saturation map that are greater than a threshold are marked as 1, and the rest as 0. , Saturation quantiles. The set of pixels with a value of 1 in this mask image represents the orientation-sensitive target area, which can be superimposed on the SAR image in pseudo-color to visually display the target detection results.

[0054] Example

[0055] A Position-Sensitive Target Detection Method Based on Sub-Aperture Color Image Saturation Features

[0056] Step 1: Sub-aperture image generation. Input is a single-channel complex SAR image data matrix. Its dimensions are ,in Represents the length in the directional direction. Represents the distance in the direction of travel. (For) Perform a Fourier transform along the azimuth direction to obtain the full aperture spectrum, and then apply the given oversampling rate. (1.2) and number of sub-apertures (For example, 5) Calculate the effective spectral length, and then obtain the length of each sub-band. Process each sub-band separately and perform an inverse Fourier transform to generate multiple sub-aperture images, each corresponding to the scattering response of the target in different azimuth sub-intervals.

[0057] Step 2: Color Mapping and Color Image Composition. A 5×3 color mapping matrix `cmap` is constructed using the `hsv` function to assign unique colors to the five sub-aperture images. Then, the five sub-aperture images are weighted and summed to generate a single RGB color image. Background regions with uniform azimuth scattering characteristics exhibit similar responses across all sub-apertures, resulting in a composite color that tends towards gray. In contrast, azimuth-sensitive man-made targets display highly saturated, vibrant colors after composite processing.

[0058] Step 3: Saturation Feature Extraction. Extract the RGB color image from Step 2. Converting to the HSV color space, where the saturation component directly describes the vibrancy of the color, allows us to extract the second channel of the converted HSV image, thus obtaining the saturation channel image. .

[0059] Step 4: Target Detection and Output. Based on saturation quantiles. (For example, 99.5%), calculate the 99.5th percentile of all pixel values ​​in the saturated image, and use it as the final segmentation threshold. Traverse the saturation map For each pixel, if the value is greater than the threshold, it is assigned a value of 1; otherwise, it is assigned a value of 0. The final output is the processed binary detection mask image.

Claims

1. A method for detecting orientation-sensitive targets based on the saturation features of sub-aperture color images, characterized in that, include: Step 1: Use single-channel SAR image data to perform sub-aperture segmentation and sub-band image generation; Step 2: Based on the color mapping matrix, weighted composite of all sub-aperture images into a single RGB color image; Step 3: Convert the RGB color image to the HSV color space, and perform saturation feature extraction and target detection based on the HSV color space.

2. The orientation-sensitive target detection method based on sub-aperture color image saturation features according to claim 1, characterized in that, The specific method for sub-aperture segmentation and sub-band image generation using single-channel SAR image data is as follows: A Fourier transform is performed on the single-channel SAR image data along the azimuth direction to obtain the full aperture spectrum; The full aperture spectrum is uniformly divided into Individual belt; Perform an inverse Fourier transform on the spectrum of each sub-band to generate multiple sub-aperture images.

3. The orientation-sensitive target detection method based on sub-aperture color image saturation features according to claim 1, characterized in that, The specific method for weighted synthesis of all sub-aperture images into a single RGB color image based on a color mapping matrix is ​​as follows: Generate an HSV color mapping matrix using MATLAB's built-in hsv function. The HSV color mapping matrix contains A different color; The sub-aperture images are normalized, and the normalized sub-aperture images are weighted and combined into an RGB color image.

4. The orientation-sensitive target detection method based on sub-aperture color image saturation features according to claim 3, characterized in that, The RGB color image specifically refers to: ， in, Corresponding to R, G, and B channels, This is the normalized sub-aperture image. For sub-aperture index, For the number of children, This is the HSV color mapping matrix.

5. The orientation-sensitive target detection method based on sub-aperture color image saturation features according to claim 1, characterized in that, The specific method for converting RGB color images to the HSV color space and performing saturation feature extraction and object detection based on the HSV color space is as follows: synthesized RGB image Converting to the HSV color space yields three components: hue (H), saturation (S), and lightness (V). The saturation map is processed using an adaptive thresholding method to determine the final binary detection mask image as follows: ， In this scenario, all pixels in the saturation map that are greater than a threshold are marked as 1, and the rest as 0. , Saturation quantiles: In a binary detection mask, the set of pixels with a value of 1 represents the orientation-sensitive target region.

Images