SAR image condition generation method and device based on statistical diffusion model

Abstract

The invention discloses an SAR image condition generation method and device based on a statistical diffusion model, and relates to the technical field of image generation, and the method comprises the steps: a forward process: obtaining to-be-processed data, gradually adding noise to the to-be-processed data, and generating to-be-processed noise data; in the reverse process, the noise data to be processed are sampled step by step, noise in the noise data to be processed is estimated and removed step by step, and new samples with conditions are generated; wherein a preset condition is added to generate a new sample with the condition. According to the invention, a high-resolution, high-quality and clear SAR image can be generated.

Description

Technical Field

[0001] The invention belongs to the technical field of image generation, and in particular relates to a SAR image condition generation method and device based on a statistical diffusion model. Background Art

[0002] Synthetic Aperture Radar (SAR) is a high-resolution imaging radar. SAR actively transmits electromagnetic waves to the ground, and then receives the reflected electromagnetic waves from the scene to interact with it, thereby achieving a description of the ground or target observation. SAR has the advantages of multi-polarization, all-day, multi-band and all-weather, and is not affected by environmental factors such as weather and light. Therefore, SAR has been widely used in meteorological detection, land survey, resource exploration and military fields, and has attracted widespread attention and research from scholars at home and abroad. With the continuous development of SAR, the research on SAR technology in various countries has also increased. At present, SAR is mainly used in image recognition and target detection. However, image recognition and target detection require a large number of samples for training to ensure good accuracy. However, SAR images are costly and limited in number, and it is difficult to obtain a large number of SAR images. To this end, proposing an effective and low-cost SAR image generation method has become a research hotspot for many scholars.

[0003] In traditional SAR generation model methods, most methods reconstruct images through SAR image feature information, but the simulated SAR images are not very similar to the original images, mainly because the information on azimuth, target outline, and shadow is not real. At the same time, the cost of constructing SAR images by traditional methods is high, and the constructed images lack diversity. Therefore, traditional methods have many difficulties in the field of SAR image generation and have developed slowly. Summary of the invention

[0004] In order to solve the above problems existing in the prior art, the present invention provides a method and device for generating SAR image conditions based on a statistical diffusion model. The technical problem to be solved by the present invention is achieved through the following technical solutions:

[0005] In a first aspect, the present invention provides a SAR image condition generation method based on a statistical diffusion model, comprising:

[0006] In the forward process, the data to be processed is obtained, and noise is gradually added to the data to be processed to generate noise data to be processed;

[0007] The reverse process is to gradually sample the noise data to be processed, gradually estimate and remove the noise in the noise data to be processed, and generate new samples with conditions;

[0008] Among them, preset conditions are added to generate new samples with conditions.

[0009] In a second aspect, the present invention further provides a SAR image condition generation device based on a statistical diffusion model, comprising: a forward processing module, used to obtain data to be processed, gradually add noise to the data to be processed, and generate noise data to be processed;

[0010] A reverse processing module is used to gradually sample the noise data to be processed, gradually estimate and remove the noise in the noise data to be processed, and generate new samples with conditions;

[0011] Among them, preset conditions are added to generate new samples with conditions.

[0012] Beneficial effects of the present invention:

[0013] The present invention provides a SAR image condition generation method and device based on a statistical diffusion model, which generates high-resolution, high-quality, and clear SAR images to provide high-quality images for SAR target detection and recognition. It solves the problem that the training of deep neural networks often faces serious overfitting problems due to the insufficient number of SAR samples, which reduces the generalization ability of the network and other related problems.

[0014] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a schematic diagram of the probability diffusion model provided by the prior art;

[0016] Figure 2 It is a flow chart of a SAR image condition generation method based on a statistical diffusion model provided by an embodiment of the present invention;

[0017] Figure 3 It is a schematic diagram of a SAR image condition generation method based on a statistical diffusion model provided by an embodiment of the present invention;

[0018] Figure 4 is another schematic diagram of a SAR image condition generation method based on a statistical diffusion model provided by an embodiment of the present invention;

[0019] Figure 5 It is a schematic diagram of a simulation experiment provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0020] The present invention is further described in detail below with reference to specific embodiments, but the embodiments of the present invention are not limited thereto.

[0021] In the existing technology, with the rapid development of deep generative networks, image generation methods based on deep generative networks have become the mainstream of current development. Variational autoencoders (VAE) and generative adversarial networks (GAN) are two widely used models for image generation; among them, the VAE model, as a generative model, can effectively generate SAR images with statistical characteristics and physical consistency by constructing a latent variable space; the GAN model can generate high-resolution, visualized SAR images through game learning between the generator and the discriminator, solving the shortcomings of traditional generation methods in removing detail textures and target features.

[0022] The above two methods have solved the problem of insufficient number of SAR image samples to a certain extent, and the interpretation effect of SAR images has been significantly improved. Among them, although the VAE model has a complete Bayesian explanation, the VAE model encodes the image into a Gaussian distribution latent vector z with independent components, which is not suitable for fitting any complex distribution; therefore, no matter how the hyperparameters are adjusted, KL(p(x,z),q(x,z)) cannot be zero. This formula represents the difference between the distribution learned by the encoder (p(x,z) and the prior distribution q(x,z), so the VAE model can only get a rough and average result, which is also one of the reasons why the images generated by the VAE model are blurred; in addition, the VAE model pays too much attention to the reconstruction loss and ignores the smoothness of the latent space, resulting in poor quality of generated samples. Although the GAN model can generate high-quality and real data, it is difficult to train and easy to crash because the GAN model is relatively complex in design and needs to achieve a balance between the generator and the discriminator during training; in addition, due to the design of the generator and discriminator of the GAN model, the generated images may be insufficient in diversity, that is, the generated samples are too similar and lack diversity.

[0023] In view of this, the present invention provides a SAR image conditional generation method based on a statistical diffusion model, which adopts multiple conditional generation models to generate SAR images, thereby achieving the generation of high-quality SAR images under specific conditions, and solving the problems in the prior art such as insufficient number of SAR image samples, blurred generated images, overly similar generated samples, lack of diversity, etc.

[0024] Before introducing the technical solution adopted by the present invention, it is necessary to first understand the Denoising Diffusion Probabilistic Models (DDMP). DDMP includes a forward process and a reverse process. The forward process is also called a diffusion process. Both the forward process and the reverse process are parameterized Markov chains. The reverse process is used to generate data, such as Figure 1 As shown, Figure 1 It is a schematic diagram of the probability diffusion model provided by the prior art.

[0025] like Figure 1 As shown in Figure 2, the DDPM forward process is a process of continuously adding noise, where x0 is the original image, ε represents the noise sampled from the standard normal distribution, and α t Represents linear interpolation from large to small, which is used to control the degree of noise addition of the model. The forward process is expressed as:

[0026]

[0027] Among them, x t represents the image after t steps of noise addition, x t-1 represents the image after t-1 steps of noise addition, i represents the index of the step, and N(0,I) represents the normal distribution.

[0028] The original image is continuously denoised, and after a finite number of steps T, the image x T It becomes a standard normal distribution for the reverse process to perform denoising operations.

[0029] The DDPM reverse process is a gradual denoising process, such as Figure 1 As shown in the figure, DDPM continuously removes noise from the standard normal distribution and finally generates an image; according to the Bayesian formula, given x0 and x t In the case of, formula (1.1) can be expressed as:

[0030]

[0031] According to formula (1.2), p(x t |x0) satisfies the Gaussian distribution, so we can reversely infer that p(x t-1 |x t ,x0) satisfies:

[0032]

[0033] That is, given x0 and x t In the case of xt-1 It obeys a Gaussian distribution, and its mean is related to x0, but in generating x t-1 Therefore, DDPM uses a neural network to predict x0, that is, the following formula (1.5), ε θ (x t ,t) represents the noise predicted by the neural network, Represents the predicted x0, which is expressed as:

[0034]

[0035] From the above equations (1.4) and (1.5), it can be actually expressed as:

[0036]

[0037] Thus, by repeatedly selecting from the Gaussian distribution x t Sample x t-1 , and finally generate an image, due to the prediction The prediction is made through the trained neural network, and the noise prediction also relies on the trained neural network. DDPM retains a strong generation capability.

[0038] The existing DDPM training process is the forward process, training the model to predict noise; the use process is its reverse process, which has been described above. The present invention is an extension based on DDPM. DDPM generates unconditional noise, while the present invention provides four conditional generation schemes based on DDPM.

[0039] See also Figure 2 , Figure 2 1 is a flow chart of a SAR image condition generation method based on a statistical diffusion model provided by an embodiment of the present invention. The SAR image condition generation method based on a statistical diffusion model provided by the present invention includes:

[0040] S101, forward process, obtaining data to be processed, and gradually adding noise to the data to be processed to generate noise data to be processed.

[0041] S102, reverse process, step by step sampling the noise data to be processed, and step by step estimating and removing the noise in the noise data to be processed, to generate new samples with conditions;

[0042] Among them, preset conditions are added to generate new samples with conditions.

[0043] The present invention provides a SAR image conditional generation method based on a statistical diffusion model, which generates high-resolution, high-quality, and clear SAR images to provide high-quality images for SAR target detection and recognition. It solves the problem that the training of deep neural networks often faces serious overfitting problems due to the insufficient number of SAR samples, which reduces the generalization ability of the network and other related problems.

[0044] In an optional embodiment of the present invention, see Figure 3 , Figure 3 : is a schematic diagram of a SAR image condition generation method based on a statistical diffusion model provided by an embodiment of the present invention. The main contents of the scheme 1 include:

[0045] Step by step, sample the noise data to be processed, and gradually estimate and remove the noise in the noise data to be processed, and generate new samples with conditions, including:

[0046] Step by step, sample the noise data to be processed and input it into the trained convolutional neural network, gradually estimate and remove the noise in the noise data to be processed, and generate unconditional new samples;

[0047] Use the trained classifier to classify the unconditional new samples and generate new samples with conditions.

[0048] In an optional embodiment of the present invention, the main contents of Solution 2 include:

[0049] Stepwise sampling of the noise data to be processed includes:

[0050] In the process of sampling the noise data to be processed, a gradient offset is added so that the mean of the sampled distribution is close to the mean of the distribution of the required conditional category; the expression of the gradient offset is:

[0051]

[0052] Among them, s represents the gradient scaling parameter, β t-1 represents the noise intensity at time step t-1 in the forward process, β t represents the noise intensity at time step t in the forward process, often also called the diffusion coefficient, x t represents the time step, and y represents the category information.

[0053] It should be noted that the reverse process satisfies the Gaussian distribution, and the mean is the center point of the Gaussian distribution, which is also the core of reverse generation. The mean of the required generated category is the center point of the target distribution. The closer the two center points are, the better the generation effect.

[0054] In an optional embodiment of the present invention, stepwise sampling of the noise data to be processed includes:

[0055] Given the known time step x t and category information y, get the data state x of the previous time step t-1 The expression of the conditional probability distribution of is:

[0056]

[0057] Among them, x t By x t-1 Add noise to get p(y|x t-1 ,x t )=p(y|x t-1 ), the previous state x t-1 The conditional probability distribution of is updated as:

[0058]

[0059] according to Expressed as a first-order approximation from a Taylor expansion, p(y|x t )About x t Expand so that you can use x t Approximate description x t-1 The distribution around it is:

[0060]

[0061] The above formula is expressed as a Gaussian distribution, defined from x t Sample x t-1 The distribution process of x t represents the data state of the current time step, σ t Indicates that the noise intensity associated with the time step in the reverse process is correlated with the noise intensity in the forward process;

[0062] Get the data state x of the previous time step t-1 The conditional probability distribution of is expressed as:

[0063]

[0064] in, is to convert x t After inputting into the preset classifier, x t The derivative shows that Ι represents the unit matrix, which is used for the covariance matrix, indicating that each dimension is independent and the noise distribution is uniform.

[0065] In an optional embodiment of the present invention, the method further includes: training the forward process and the reverse process; wherein,

[0066] Training the forward process includes taking the preset data as the original data, gradually adding noise to the original data, and generating the noise data of the standard normal distribution;

[0067] Training the reverse process involves taking the noisy data as initial data, training the initial convolutional neural network with the goal of minimizing the difference between the predicted noise and the actual added noise, and gradually removing the noise to reconstruct the original data.

[0068] In an optional embodiment of the present invention, see Figure 4 , Figure 4 It is another schematic diagram of a SAR image condition generation method based on a statistical diffusion model provided by an embodiment of the present invention. The main contents of the scheme three include:

[0069] Training the forward process also includes:

[0070] Encode the time step and the category information at the same time;

[0071] The initial convolutional neural network is trained using the encoded time steps and the encoded category information.

[0072] In an optional embodiment of the present invention, the data state x of the previous time step is obtained. t-1 The conditional probability distribution of is expressed as:

[0073]

[0074]

[0075] in, Indicates the data status of the current time step preset after adding category information.

[0076] Specifically, in this embodiment, the time step and category information are encoded so that they are input together into the convolutional neural network to be trained;

[0077] From x T Start iterating the above process to x0;

[0078] Use a convolutional neural network to predict the noise at the current time step t, denoise it based on the noise prediction and sample the next time step x t-1 ;

[0079] The above process is subject to the constraints of category information. At time step t=0, the denoising process is completed, and due to the constraints of category information, an image that meets the conditions is generated.

[0080] It should be noted that in scheme three, the category information is encoded at the beginning of the forward process, and it carries the category information throughout the entire diffusion process (forward and backward). In the forward process, the model does not rely on the category information, and only relies on the category information for guidance and image generation in the backward diffusion process.

[0081] In an optional embodiment of the present invention, the main contents of scheme 4 include:

[0082] Training the reverse process further includes:

[0083] The initial convolutional neural network is trained using the preset data set and the noise data, with the goal of minimizing the difference between the predicted noise and the actual added noise, and the noise is gradually removed to reconstruct the original data, and the convolutional neural network parameters are obtained to construct a trained convolutional neural network.

[0084] Specifically, in this embodiment, first, the DDPM model is used as a pre-training model, and the forward process (noise adding process) is frozen, which means that the parameters of the forward process will not be updated during subsequent training; secondly, at the beginning of the reverse process, a specific data set (with specific category information) is input into the convolutional neural network to be trained together with the noise data; finally, in the denoising process, the reverse process network is optimized according to the current noise state and category information, and images that conform to specific categories are gradually denoised.

[0085] It should be noted that Solution 3 does not freeze the forward process. The training model is optimized by training the forward and backward processes together. It requires greater computing resources and a longer training time. However, if the training resources are sufficient, it has stronger control over details. Solution 4 only needs to train the backward process, and the forward process is the forward process of DDPM. In this way, computing resources and time costs can be saved.

[0086] In an optional embodiment of the present invention, training the initial convolutional neural network includes:

[0087] A threshold is set in the convolutional neural network to filter the output of the convolutional neural network and remove images with poor generation effect.

[0088] Based on the same inventive concept, the present invention further provides a SAR image condition generation device based on a statistical diffusion model, which is used to implement the SAR image condition generation method based on a statistical diffusion model provided in the above embodiment of the present invention. The method embodiment is referred to above and will not be described in detail here. The device comprises:

[0089] A forward processing module is used to obtain data to be processed, gradually add noise to the data to be processed, and generate noise data to be processed;

[0090] A reverse processing module is used to gradually sample the noise data to be processed, and gradually estimate and remove the noise in the noise data to be processed, so as to generate new samples with conditions;

[0091] Among them, preset conditions are added to generate new samples with conditions.

[0092] In an optional embodiment of the present invention, the effect of the SAR image condition generation method based on the statistical diffusion model provided in the above embodiment is verified through simulation experiments, specifically:

[0093] 1. FID scores of the four conditional generative models.

[0094] Commonly used evaluation indicators in the generation field include Inception Score (IS) and Fréchet Inception Distance (FID). However, since IS uses the Inception model pre-trained on the natural dataset ImageNet to calculate the conditional distribution of the generated image, and SAR images are not natural images, IS is not suitable for the evaluation of SAR image models in this article. In contrast, the FID indicator compares the feature vectors obtained by inputting the original image and the generated image into the Inception model respectively to obtain the FID score. Therefore, most of the current SAR image generation models use the FID indicator for evaluation.

[0095] However, the FID index also has the following defects: the FID score can only measure the distance between samples well when the number of samples is extremely large or the number of samples is the same. When the number of samples is small, the covariance matrix fluctuates too much and cannot well represent the distance between samples.

[0096] In order to ensure the same number of samples, we selected 428 samples (the smallest number of categories in the MSTAR dataset) to evaluate the four conditional generation schemes. Since the Classifier-Free scheme requires large GPU memory, the image is cropped to (96x96), while the other three schemes crop the image to (128x128), so we only evaluated the performance of the classifier conditional generation scheme, the classifier-guided conditional generation scheme, and the pre-trained conditional generation scheme. The FID scores we obtained are shown in Table 1, where the black part indicates the model with the lowest FID score, that is, the best performance.

[0097] Table 1 FID scores of three conditional generation models

[0098] Category Name Classifier pre-trained Classifier-Guidance 2S1 108.326 113.9692 124.4123 BMP2 93.2471 98.29862 98.65277 BRDM_2 73.9459 74.69507 90.906 BTR60 135.2398 129.214 152.8392 BTR70 112.9692 111.26 120.9701 D7 85.1499 92.71053 99.25594 T62 92.4519 96.06115 98.98671 T72 92.157 100.1628 118.0358 ZIL131 91.414 97.41465 105.1324 ZSU_23_4 87.9022 96.10545 93.20193

[0099] 2. Evaluation of generative models based on AlexNet.

[0100] SAR target detection and recognition are important research topics in SAR image interpretation. In order to ensure accuracy, target detection and recognition require a large number of samples for training. This section will add the SAR images generated by Diffusion to the classifier training, and verify the research significance of this invention by whether the classifier performance is improved. At the same time, the three conditional generation schemes other than the Classifier-Free scheme are indirectly evaluated by the performance improvement.

[0101] Since the data generated by Diffusion in the present invention is trained by the SOC dataset in the MSTAR dataset, we will use the EOC dataset in the MSTAR dataset as the training sample of the classifier. Figure 5 As shown in Figure 1, we divide the EOC dataset into training and test sets, and merge the SOC dataset and the EOC training set into a new training set. The images generated by the three conditional generation schemes are injected into the training set at a certain ratio to determine whether the injected generated images can improve the classifier performance. The injection ratio is based on the amount of data in the SOC dataset, and is interpolated according to the linear interpolation from 0 to 1 with a step size of 0.1.

[0102] Here, AlexNet is used as the classifier network, with a batch size of 32 and an initial learning rate of 0.01. The training rate is reduced to one tenth of the original value every 20 rounds, and a total of 50 rounds are iterated. Table 2 shows the experimental results. From Table 2, it can be seen that when there is no injected image, the classifier accuracy is 0.9358. The injection of images generated by three conditions can greatly improve the performance of the classifier. Among them, the classifier conditional generation without threshold is the use of classifier conditional generation, in which the threshold is set to the mean of the ten-dimensional vector after the image passes through the classifier. It can be seen that a reasonably set threshold can improve the generation quality of the classifier conditional generation model. The pre-trained conditional generation model also achieved good results. Therefore, from the evaluation results of the classifier model, the classifier conditional generation with a threshold achieved the best effect, followed by the pre-trained conditional generation, and the classifier-guaidance classifier is finally the conditional generation without a threshold. At the same time, the injection of images generated by the three schemes can significantly improve the performance of the classifier. Therefore, using Diffusion to expand SAR images is helpful for the development of SAR image interpretation.

[0103] Table 2 Classifier accuracy after injecting generated images

[0104]

[0105] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. Moreover, the term "include", "comprise" or any other variant is intended to cover non-exclusive inclusion, so that the article or device including a series of elements includes not only those elements, but also other elements that are not explicitly listed. In the absence of more restrictions, the elements defined by the sentence "including one..." do not exclude the existence of other identical elements in the article or device including the elements. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. The orientation or position relationship indicated by "up", "down", "left", "right", etc. is based on the orientation or position relationship shown in the accompanying drawings, which is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation of the present invention.

[0106] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification.

[0107] The above contents are further detailed descriptions of the present invention in combination with specific preferred embodiments, and it cannot be determined that the specific implementation of the present invention is limited to these descriptions. For ordinary technicians in the technical field to which the present invention belongs, several simple deductions or substitutions can be made without departing from the concept of the present invention, which should be regarded as falling within the protection scope of the present invention.

Claims

1. A SAR image conditional generation method based on a statistical diffusion model, characterized in that: include: In the forward process, the data to be processed is obtained, and noise is gradually added to the data to be processed to generate noise data to be processed; In the reverse process, the noise data to be processed is sampled step by step, the noise in the noise data to be processed is estimated and removed step by step, and new samples with conditions are generated; Among them, preset conditions are added to generate new samples with conditions.

2. The SAR image condition generation method based on the statistical diffusion model according to claim 1, characterized in that: The step of sampling the noise data to be processed step by step, estimating and removing the noise in the noise data to be processed step by step, and generating new samples with conditions includes: Step by step, the noise data to be processed is sampled and input into a trained convolutional neural network, and the noise in the noise data to be processed is estimated and removed step by step to generate unconditional new samples; The unconditional new samples are classified using the trained classifier to generate conditional new samples.

3. The SAR image condition generation method based on the statistical diffusion model according to claim 1, characterized in that: The stepwise sampling of the noise data to be processed comprises: In the process of sampling the noise data to be processed, a gradient offset is added so that the mean of the sampled distribution is close to the mean of the distribution of the required conditional category; wherein the expression of the gradient offset is: Among them, s represents the gradient scaling parameter, β t-1 represents the noise intensity at time step t-1 in the forward process, β t represents the noise intensity at time step t in the forward process, x t represents the time step, and y represents the category information.

4. The SAR image condition generation method based on the statistical diffusion model according to claim 3 is characterized in that: The stepwise sampling of the noise data to be processed comprises: Given the known time step x t and category information y, get the data state x of the previous time step t-1 The expression of the conditional probability distribution of is: Among them, x t By x t-1 Add noise to get p(y|x t-1 ,x t )=p(y|x t-1 ), the data state x of the previous time step t-1 The conditional probability distribution of is updated as: according to get: Among them, x t represents the data state of the current time step, σ t represents the noise intensity associated with the time step in the reverse process; Get the data state x of the previous time step t-1 The distribution of the conditional probability is expressed as: in, is to convert x t After inputting into the preset classifier, x t Taking the derivative, we get Ι, which represents the identity matrix.

5. The SAR image condition generation method based on the statistical diffusion model according to claim 1, characterized in that: Also includes: The forward process and the reverse process are trained; wherein, Training the forward process includes taking preset data as original data, gradually adding noise to the original data, and generating noise data of standard normal distribution; Training the reverse process includes taking the noise data as initial data, training the initial convolutional neural network to minimize the difference between the predicted noise and the actual added noise as the goal, and gradually removing the noise to reconstruct the original data.

6. The SAR image condition generation method based on the statistical diffusion model according to claim 5, characterized in that: Training the forward process further includes: Encode the time step and the category information at the same time; The initial convolutional neural network is trained using the encoded time steps and the encoded category information.

7. The SAR image condition generation method based on the statistical diffusion model according to claim 6, characterized in that: Get the data state x of the previous time step t-1 The conditional probability distribution of is expressed as: in, Indicates the data status of the current time step preset after adding category information.

8. The SAR image condition generation method based on the statistical diffusion model according to claim 5, characterized in that: Training the reverse process further includes: The initial convolutional neural network is trained using the preset data set and the noise data, with the goal of minimizing the difference between the predicted noise and the actual added noise, and the noise is gradually removed to reconstruct the original data, and the convolutional neural network parameters are obtained to construct a trained convolutional neural network.

9. The SAR image condition generation method based on the statistical diffusion model according to claim 5, characterized in that: Training the initial convolutional neural network includes: A threshold is set in the convolutional neural network to filter the output of the convolutional neural network and remove images with poor generation effect.

10. A SAR image condition generation device based on a statistical diffusion model, characterized in that: include: A forward processing module is used to obtain data to be processed, and gradually add noise to the data to be processed to generate noise data to be processed; A reverse processing module, used for gradually sampling the noise data to be processed, gradually estimating and removing the noise in the noise data to be processed, and generating new samples with conditions; Among them, preset conditions are added to generate new samples with conditions.

Images