Ship ISAR image recognition method based on optical image assistance

Abstract

The invention discloses a ship ISAR image recognition method based on optical image assistance, and relates to a ship ISAR image recognition method. The invention aims to solve the problem of low shipISAR image recognition accuracy in the prior art. The method comprises the following steps: 1, preprocessing ship ISAR images with known ship types in a database; 2, obtaining a trained derivative network Pix2pix; 3, generating a ship optical image from the ship ISAR image by using the trained derivation network Pix2pix, and merging the ship optical image and the ship ISAR image; obtaining a trained convolutional neural network; 4, preprocessing the ship ISAR image to be tested; 5, combining the obtained ship optical image with the ship ISAR image by using the trained derivation network Pix2pix; and 6, inputting the obtained composite image into the trained convolutional neural network to obtain a category. The method is applied to the field of radar target recognition.

Description

Technical field

[0001] The invention belongs to the field of radar target recognition, and specifically relates to a ship ISAR image recognition method. Background technique

[0002] Inverse synthetic aperture radar (ISAR) completes target imaging by transmitting signals to the detection target and receiving target echoes. The most commonly used imaging algorithm is the range-Doppler imaging algorithm. The essence of the algorithm is to perform a two-dimensional Fourier transform of the received echo in the azimuth and range. The resolution of the range is determined by the carrier frequency. , The resolution of the azimuth depends on the angle that the target has turned in the coherent accumulation time. The classification and recognition of ISAR images started from flying targets in Botha E C.Classification of aerospacetargets using superresolution ISAR images[C].Communications and SignalProcessing,1994.COMSIG-94,Proceedings of the 1994IEEE South African Sympoisumon.1994:138-1...

Examples

specific Embodiment approach 1

[0028] Specific embodiment one: combination figure 1 To explain this embodiment, the specific process of a ship ISAR image recognition method based on optical image assistance in this embodiment is:

[0029] Step 1. Pre-process the ISAR images of ships of known ship types (frigates, destroyers, etc.) in the database. The pre-processing includes denoising processing and image normalization processing;

[0030] Step 2: Use the GAN derivative network Pix2pix to complete the mapping of the preprocessed ship ISAR image domain to the ship optical image (known in the database) domain in step 1, and obtain the trained derivative network Pix2pix;

[0031] Step 3. Use the derived network Pix2pix trained in step 2 to generate the ship optical image from the preprocessed ship ISAR image, and merge the ship optical image with the ship ISAR image. The composite image contains both ship ISAR image information and Optical image information; use synthetic images to construct a training set, train a c...

specific Embodiment approach 2

[0035] Second embodiment: This embodiment is different from the first embodiment in that the ISAR image is preprocessed in the first step, and the preprocessing includes denoising processing and image normalization processing; the specific process is:

[0036] Step 1: Use the global threshold to denoise the ISAR image to obtain the denoised ISAR image of the ship;

[0037] Step one and two: normalize the denoised ISAR image of the ship obtained in step one.

[0038] The other steps and parameters are the same as in the first embodiment.

specific Embodiment approach 3

[0039] Specific embodiment three: this embodiment is different from specific embodiment one or two in that in the step one, a global threshold is used to perform the ISAR image (such as Image 6 ) To perform denoising processing to obtain the denoised ISAR image of the ship (such as Figure 7 ); The process is: see the flowchart figure 2 :

[0040] Step one by one, set the difference threshold t, which is used to determine the global threshold;

[0041] Step one, two, set the initial threshold to T:

[0042]

[0043] In the formula, I(i,j) max Represents the brightest pixel value in the ISAR image of the ship, I(i,j) min Represents the pixel value of the darkest point in the ship ISAR image; i represents the ith row of the image matrix, and j represents the jth column of the image matrix;

[0044] Step 13: Compare the intensity (pixel value) of each pixel in the ship's ISAR image with the initial threshold threshold T, and group the pixels whose intensity is greater than or equal to...

Technical field

[0001] The invention belongs to the field of radar target recognition, and specifically relates to a ship ISAR image recognition method. Background technique

[0002] Inverse synthetic aperture radar (ISAR) completes target imaging by transmitting signals to the detection target and receiving target echoes. The most commonly used imaging algorithm is the range-Doppler imaging algorithm. The essence of the algorithm is to perform a two-dimensional Fourier transform of the received echo in the azimuth and range. The resolution of the range is determined by the carrier frequency. , The resolution of the azimuth depends on the angle that the target has turned in the coherent accumulation time. The classification and recognition of ISAR images started from flying targets in Botha E C.Classification of aerospacetargets using superresolution ISAR images[C].Communications and SignalProcessing,1994.COMSIG-94,Proceedings of the 1994IEEE South African Sympoisumon.1994:138-1...

Examples

specific Embodiment approach 1

[0028] Specific embodiment one: combination figure 1 To explain this embodiment, the specific process of a ship ISAR image recognition method based on optical image assistance in this embodiment is:

[0029] Step 1. Pre-process the ISAR images of ships of known ship types (frigates, destroyers, etc.) in the database. The pre-processing includes denoising processing and image normalization processing;

[0030] Step 2: Use the GAN derivative network Pix2pix to complete the mapping of the preprocessed ship ISAR image domain to the ship optical image (known in the database) domain in step 1, and obtain the trained derivative network Pix2pix;

[0031] Step 3. Use the derived network Pix2pix trained in step 2 to generate the ship optical image from the preprocessed ship ISAR image, and merge the ship optical image with the ship ISAR image. The composite image contains both ship ISAR image information and Optical image information; use synthetic images to construct a training set, train a c...

specific Embodiment approach 2

[0035] Second embodiment: This embodiment is different from the first embodiment in that the ISAR image is preprocessed in the first step, and the preprocessing includes denoising processing and image normalization processing; the specific process is:

[0036] Step 1: Use the global threshold to denoise the ISAR image to obtain the denoised ISAR image of the ship;

[0037] Step one and two: normalize the denoised ISAR image of the ship obtained in step one.

[0038] The other steps and parameters are the same as in the first embodiment.

specific Embodiment approach 3

[0039] Specific embodiment three: this embodiment is different from specific embodiment one or two in that in the step one, a global threshold is used to perform the ISAR image (such as Image 6 ) To perform denoising processing to obtain the denoised ISAR image of the ship (such as Figure 7 ); The process is: see the flowchart figure 2 :

[0040] Step one by one, set the difference threshold t, which is used to determine the global threshold;

[0041] Step one, two, set the initial threshold to T:

[0042]

[0043] In the formula, I(i,j) max Represents the brightest pixel value in the ISAR image of the ship, I(i,j) min Represents the pixel value of the darkest point in the ship ISAR image; i represents the ith row of the image matrix, and j represents the jth column of the image matrix;

[0044] Step 13: Compare the intensity (pixel value) of each pixel in the ship's ISAR image with the initial threshold threshold T, and group the pixels whose intensity is greater than or equal to...

Images