Mars high-resolution image functional analysis stripe removing method and system and terminal equipment

Abstract

The invention discloses a Mars high-resolution image functional analysis stripe removing method and system and terminal equipment, and relates to the technical field of planet remote sensing image processing. The method comprises the steps of receiving Mars high-resolution image data containing stripe noise as original data input, and performing preprocessing; a high-resolution image degradation model is constructed, wavelet decomposition is carried out on original noisy image data based on a wavelet transform method, the number of layers of wavelet decomposition is n, and the wavelet decomposition is stopped until no noise appears in each sub-image after decomposition. The method combines the advantages and characteristics of multi-resolution of wavelet transform and automation in energy functional analysis, combines the advantages of the two methods, performs multilayer wavelet decomposition on an original image, and then operates sub-images of corresponding layers, so that strip noise is removed to the greatest extent, and space details of ground features are reserved to the greatest extent; and the fidelity of the denoised image is high.

Description

Technical Field

[0001] The present invention relates to the technical field of planetary remote sensing image processing, and in particular to a method, system and terminal device for removing stripes by functional analysis of high-resolution Mars images. Background Art

[0002] At present, humans have carried out many deep space exploration missions to extraterrestrial bodies such as the moon and Mars, and have made sufficient progress. High-resolution images of Mars observation imaging have important scientific significance for the study of scientific issues such as the surface morphology and geological evolution of Mars, and are also a relatively scarce data resource in current Mars research. However, stripe noise is a common degradation problem in remote sensing satellite imaging systems with multiple CCD splicing. It not only obscures the real detailed information of the surface, but also reduces the quality and application value of the image. Many factors can cause stripe noise in the image, such as sensor aging, non-uniform response between CCD detectors, temperature changes of the detector, mechanical movement, etc.

[0003] The principles of stripe noise removal of remote sensing images can be roughly summarized as follows: digital filtering-based methods, statistics-based methods, and optimization-based methods. Digital filtering is a common and effective method for image denoising. Spatial filtering often loses image detail information. A common solution is to use a low-pass filter in the frequency domain to remove stripe noise, achieving a good image denoising effect. However, most of the current methods based on image processing in the frequency domain still have problems and shortcomings. Most filters cannot effectively separate the signal and noise parts, or the threshold and threshold function selection of wavelet transform is very difficult. In addition, the stripe noise of Mars high-resolution remote sensing images often exists irregularly in the image. Selecting an inappropriate filter or threshold processing will produce blurring and ringing phenomena, resulting in unclean noise processing and misprocessing of effective information. Summary of the invention

[0004] The purpose of the present invention is to provide a method, system and terminal device for functional analysis of Mars high-resolution images, which reduces the process of traditional human intervention, achieves high fidelity of denoised image information, and also improves the quality of denoised images.

[0005] According to a first aspect of the present invention, in order to achieve the above-mentioned purpose, the present invention provides the following technical solution: a method for removing stripes by functional analysis of Mars high-resolution images, comprising the following steps: Receiving high-resolution Mars image data containing stripe noise as raw data input and performing preprocessing; A high-resolution image degradation model is constructed. The original noisy image data is decomposed by wavelet transform method. The number of wavelet decomposition layers is n. The decomposition is terminated when there is no noise in each sub-image after decomposition. The noisy sub-images containing stripe noise are screened out and the sub-images without noise are filtered out. For all noisy subgraphs after wavelet decomposition, the energy functional of each noisy subgraph is established, and a one-way total variation model is constructed for the noisy subgraph. The fidelity term and regularization term are applied, and the alternating direction multiplier method is used to solve the functional minimum. The tolerance is set to end the iterative convergence, and the processed noisy subgraph is obtained. The inverse wavelet transform is used to reconstruct and restore the processed denoised sub-image and the sub-image without noise to obtain the entire denoised image; A quantitative evaluation is performed on the entire denoised image, and repeated denoising is performed if it does not meet the standard.

[0006] Furthermore, the high-resolution image data of Mars containing stripe noise is acquired through satellite collection, and the raw data is preprocessed, specifically including: (21) Determine the image format to determine whether it is a single-band image or a multi-band image; (22) If it is a multi-band image, it is processed band by band and finally synthesized into one image; (23) Determine whether the direction of the stripe noise in the image is horizontal, vertical or diagonal.

[0007] Furthermore, a high-resolution image degradation model is constructed as follows: For two-dimensional The degradation model of high-resolution images containing stripe noise is expressed as: In the formula , represents an image with stripe noise, represents an image without stripe noise, Indicates banding noise.

[0008] Furthermore, the original noisy image data is decomposed by wavelet transform method. The number of wavelet decomposition layers is n, and the decomposition is terminated when there is no noise in each sub-image after decomposition. The noisy sub-image containing stripe noise is screened out and the sub-image without noise is filtered out. The details are as follows: For the original noisy image For , its two-dimensional discrete wavelet transform is expressed as: in is the component of each sub-graph, To decompose the scale, Represents the domain space, assuming that the original image contains noise Recorded as , then its multi-layer decomposition is: In the formula and yes and The conjugate matrix of is the number of decomposition layers, , , and Respectively represent The low-frequency component, horizontal component, vertical component and diagonal component after layer decomposition.

[0009] Furthermore, for all noisy subgraphs after wavelet decomposition, the energy functional of each noisy subgraph is established, and a one-way total variation model is constructed for the noisy subgraph. The fidelity term and regularization term are applied, and the alternating direction multiplier method is used to solve the functional minimum. The tolerance is set to end the iterative convergence, and the processed noisy subgraph is obtained, as follows: (51) The energy functional of each noisy subgraph is established and expressed as: In the formula For the fidelity item; is the regularization term; (52) For noisy images In a bounded space The energy total variation model is constructed as: In the formula For the fidelity item; is the regularization term, The total variation of ; is the Lagrangian factor, which is used to adjust the effect of the balance fidelity term and the regularization term; Therefore, image denoising becomes solving the energy functional Minimum problem: in For the denoised image, solving the functional minimum is transformed into solving the Euler-Lagrange equation problem: In the formula, represents an image with stripe noise, represents an image without stripe noise, yes The gradient of is the Lagrangian factor.

[0010] Furthermore, the inverse wavelet transform is used to reconstruct and restore the processed denoised sub-image and the noise-free sub-image to obtain the entire denoised image, as follows: In the formula, and yes and The conjugate matrix of is the number of decomposition layers, , , and Respectively represent The low-frequency component, horizontal component, vertical component and diagonal component after layer decomposition.

[0011] Furthermore, a quantitative evaluation is performed on the entire denoised image, using three quantitative indicators: peak signal-to-noise ratio, structural similarity, and relative dimensionless global comprehensive error, as shown below: (71) The peak signal-to-noise ratio is calculated as follows: In the formula, is the maximum gray value of the image, and are the number of rows and columns of image pixels, and They are respectively the image before and after denoising. Line The pixel value of the column; (72) The structural similarity value ranges from 0 to 1 and is defined as follows: In the formula, and These are the two images before and after denoising. and The average value of and Represent the standard deviation of the two images before and after denoising, and are two positive integers; (73) Relative dimensionless global comprehensive error, specifically defined as follows: In the formula, Represents the ratio of the spatial resolution of the original image to that of the denoised image, Indicates the number of bands of the image, Represents the original image and the denoised image The root mean square error of the bands, The original image The mean of the bands; (74) Set SSIM ≥ 0.95 and ERGAS ≤ 10 for ideal processing effect, otherwise perform repeated denoising.

[0012] According to a second aspect of the present invention, the present invention provides a Mars high-resolution image functional analysis striping system, which is used to implement the above-mentioned Mars high-resolution image functional analysis striping method, comprising: A preprocessing module is used to receive high-resolution Mars image data containing stripe noise as raw data input and perform preprocessing; The construction module is used to construct a high-resolution image degradation model. The original noisy image data is decomposed by wavelet based on the wavelet transform method. The number of wavelet decomposition layers is n. The decomposition is terminated when there is no noise in each sub-image after decomposition. The noisy sub-image containing stripe noise is screened out and the sub-image without noise is filtered out. The noisy subgraph processing module is used to establish the energy functional of each noisy subgraph after wavelet decomposition, and construct a one-way total variation model for the noisy subgraph, apply the fidelity term and the regularization term, use the alternating direction multiplier method to solve the functional minimum, set the tolerance to end the iterative convergence, and obtain the processed noisy subgraph; An image reconstruction module is used to reconstruct and restore the processed denoised sub-image and the noise-free sub-image by using an inverse wavelet transform to obtain the entire denoised image; The evaluation module is used to perform quantitative evaluation on the entire denoised image, and repeat the denoising if it does not meet the standards.

[0013] According to a third aspect of the present invention, the present invention provides a terminal device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the memory stores a computer program capable of running on the processor, and when the processor loads and executes the computer program, the above-mentioned Mars high-resolution image functional analysis de-banding method is adopted.

[0014] According to a fourth aspect of the present invention, the present invention provides a storage medium comprising computer executable instructions, which, when executed by a computer processor, are used to execute the above-mentioned Mars high-resolution image functional analysis de-banding method.

[0015] The present invention has at least the following beneficial effects: (1) The present invention combines the advantages of wavelet transform in multi-resolution and energy functional analysis, combines the advantages of the two methods, performs multi-layer wavelet decomposition on the original image, and then operates the sub-images of the corresponding layers, thereby removing the stripe noise to the greatest extent and retaining the spatial details of the ground objects to the greatest extent. The denoised image has high fidelity; (2) While maintaining the same image restoration quality, the present invention minimizes human involvement, reduces the process of setting parameters such as traditional Fourier transform, reduces the intervention process and human intervention error factors, saves a lot of manpower and material resources, and is suitable for automated batch processing of large amounts of Mars high-resolution image data.

[0016] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic diagram of the process of the method of the present invention; Figure 2 It is a schematic diagram of the principle of the method of the present invention; Figure 3 Schematic diagram of three-layer wavelet decomposition in an embodiment of the present invention; Figure 4 Schematic diagram of the direction of a stripe noise image in an embodiment of the present invention, where a is horizontal, b is vertical, and c is oblique; Figure 5 This is a schematic diagram of HiRIC raw data in an embodiment of the present invention; Figure 6 Schematic diagram of three-scale wavelet decomposition of HiRIC raw data in an embodiment of the present invention, wherein C1 represents a low-frequency sub-image, H1 represents a horizontal sub-image, V1 represents a vertical sub-image, and D1 represents a diagonal sub-image; Figure 7 : Gradient maps of HiRIC raw data in various directions in an embodiment of the present invention, where (a) is the horizontal gradient and (b) is the vertical gradient; Figure 8 Schematic diagram of an image after stripe noise removal in an embodiment of the present invention; Fig. 9 Comparison of the image column mean of the denoising result of the present invention with the original image column mean (columns 250-400). DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present disclosure to clearly and completely describe the technical solutions in the embodiments of the present disclosure. Obviously, the described embodiments are only part of the embodiments of the present disclosure, rather than all the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0019] See also Figure 1 and Figure 2 The present invention provides a technical solution: a method for removing stripes by functional analysis of high-resolution images of Mars, comprising the following steps: S1. Receive Mars high-resolution image data containing stripe noise as raw data input (raw high-resolution image) and perform preprocessing; The high-resolution image data of Mars containing stripe noise is acquired through satellite collection, and the raw data is preprocessed, including: (21) Determine the image format to determine whether it is a single-band image or a multi-band image; (22) If it is a multi-band image, it is processed band by band and finally synthesized into one image; (23) Determine the directionality of the stripe noise in the image as horizontal, vertical or oblique, in preparation for subsequent operations; like Figure 4 As shown in the figure, the stripe noise in the acquired image includes horizontal, vertical and oblique directions. Figure 4 In the table, a, b and c are used to represent them respectively; S2. Construct a high-resolution image degradation model, and perform wavelet decomposition on the original noisy image data based on the wavelet transform method. The wavelet decomposition layer number is n, and the process ends when there is no noise in each sub-image after decomposition. The noisy sub-images containing stripe noise are screened out, and the sub-images without noise are filtered out. The details are as follows: S21. Construct a high-resolution image degradation model, as follows: For two-dimensional The degradation model of high-resolution images containing stripe noise is expressed as: In the formula , represents an image with stripe noise, represents an image without stripe noise, represents stripe noise; S22. Based on the wavelet transform method, the original noisy image data is decomposed by wavelet, such as Figure 3 As shown, the details are as follows: For the original noisy image For , its two-dimensional discrete wavelet transform is expressed as: in For each sub-layer component, To decompose the scale, Represents the domain space, assuming that the original image contains noise Recorded as , then its multi-layer decomposition is: In the formula and yes and The conjugate matrix of is the number of decomposition layers, , , and Respectively represent The low-frequency component, horizontal component, vertical component and diagonal component after layer decomposition; S3. Consider an image as an energy system. For all noisy sub-images after wavelet decomposition, establish the energy functional of each sub-image. According to the directional characteristics of stripe noise, construct a unidirectional total variation model (UTV). Only process in the noise direction without affecting the information in other directions. The regularization term under the unidirectional variation is mainly placed in the direction where the stripe is located. Use the alternating direction multiplier method to solve the functional minimum and set the tolerance. End the iteration convergence and obtain the processed noisy subgraph; S31. Establish the energy functional of each noisy subgraph, expressed as: in It is a fidelity term, which is mainly used to constrain the denoised image to retain the original information as much as possible. It is a regularization term, and its main function is to ensure that the image is smooth enough and remove noise in the image; (52) For noisy images In a bounded space The energy total variation model is constructed as: In the formula For the fidelity item; is the regularization term, The total variation of ; is the Lagrangian factor, which is used to adjust the effect of the balance fidelity term and the regularization term; Therefore, image denoising becomes solving the energy functional Minimum problem: in For the denoised image, solving the functional minimum is transformed into solving the Euler-Lagrange equation problem: In the formula, represents an image with stripe noise, represents an image without stripe noise, yes The gradient of is the Lagrangian factor; S4. Use inverse wavelet transform to reconstruct and restore the processed denoised sub-image and the noise-free sub-image (wavelet reconstruction) to obtain the entire denoised image, as follows: In the formula, and yes and The conjugate matrix of is the number of decomposition layers, , , and Respectively represent The low-frequency component, horizontal component, vertical component and diagonal component after layer decomposition; S5. Perform quantitative evaluation on the entire denoised image, and repeat denoising if it does not meet the standard; Three quantitative evaluation indicators, Peak Signal to Noise Ratio (PSNR), Structural Similarity Index (SSIM), and Relative Dimensionless Global Adimensionnelle de Synthese (ERGAS), are used for quantitative evaluation, as follows: S51. Peak signal-to-noise ratio (PSNR) is used to measure the ratio of the maximum signal to the noise intensity. The larger the value, the better the image restoration quality. The calculation formula is as follows: In the formula, is the maximum gray value of the image, and are the number of rows and columns of image pixels, and They are respectively the image before and after denoising. Line The pixel value of the column; S52. Structural similarity SSIM considers that images are structured and extracts features from brightness, contrast and structure to measure the structural similarity between two images. The value range is between 0 and 1. The larger the value, the better the quality of the denoised image and the more complete the original structural information retained. The definition is as follows: In the formula, and These are the two images before and after denoising. and The average value of and Represent the standard deviation of the two images before and after denoising, and are two positive integers; S53. Relative dimensionless global comprehensive error ERGAS can measure the global quality of the denoised image and reflect the distortion of the image radiation value. Ideally, the value of ERGAS is 0, but in the real reconstructed image, its value is higher. The lower the value of ERGAS, the better the image quality. It is defined as follows: In the formula, Represents the ratio of the spatial resolution of the original image to that of the denoised image, Indicates the number of bands of the image, Represents the original image and the denoised image The root mean square error of the bands, The original image The mean of the bands; S54. For high-resolution Mars images, set SSIM ≥ 0.95 and ERGAS ≤ 10 for ideal processing results. Otherwise, repeat steps S2 to S4 for denoising again. For different image sources, the parameters can be fine-tuned according to usage requirements.

[0020] The following is a detailed description of an embodiment of the present invention in conjunction with the accompanying drawings. Taking the image data obtained by the HiRIC high-resolution imaging camera of the Tianwen-1 orbiter for imaging Mars as an example, this embodiment is carried out on the basis of the technical solution of the present invention, and has achieved good results on the high-resolution image data of the Tianwen-1 orbiter. While minimizing manual participation, it can effectively remove the stripe noise of the high-resolution image of Mars, and can provide technical support for deep space exploration missions. This embodiment includes the following main steps: Step 1: Figure 5 As shown, the HiRIC original image data is obtained. The image data is full-color band data. The experimental data (400×400) selected in this embodiment has a spatial resolution of about 0.7 m / pixel. The stripe noise is an obvious longitudinal stripe, and the distribution is irregular. The most obvious stripe is located at Fig. 9 The DN value (column mean) of the 47th column of the original image is low and appears as a black strip, such as Fig. 9 The DN values ​​in the 264th and 346th columns of the original image are higher and appear as white stripes; Step 2: Perform multi-layer wavelet decomposition on the original data to determine the noise content in all layers. According to the characteristics of the vertical stripe noise in this case, the low-frequency sub-images ( ) and vertical subplots ( ), continue to decompose and observe, the noise in the third layer of sub-image disappears, stop decomposition, and the number of wavelet decomposition layers is 3, such as Figure 6 As shown, C1 represents the low-frequency sub-image, H1 represents the horizontal sub-image, V1 represents the vertical sub-image, and D1 represents the diagonal sub-image; Step 3: Select all noisy low-frequency sub-images ( ) and vertical subplots ( ), establish the energy system of the image, perform energy functional analysis, solve the functional minimum using the alternating direction multiplier method, and set the tolerance End the iteration convergence, such as Figure 7 As shown, (a) represents the horizontal gradient, and (b) represents the vertical gradient; Step 4: Perform inverse wavelet transform on each processed sub-image and the original retained sub-image to restore the image, such as Figure 8 As shown; Step 5: Calculate the quantitative evaluation index and evaluate the denoising results, as shown in Table 1: Table 1 Quantitative index results like Fig. 9 As shown, the column mean of the denoised image is compared with the column mean of the original image (columns 250-400); The results show that the present invention has an obvious denoising effect while ensuring that the original image information is not lost to the greatest extent, and has achieved a good denoising effect.

[0021] Embodiment 2: This embodiment provides a Mars high-resolution image functional analysis and striping removal system, which is used to implement the Mars high-resolution image functional analysis and striping removal method described in Embodiment 1, including: A preprocessing module is used to receive high-resolution Mars image data containing stripe noise as raw data input and perform preprocessing; The construction module is used to construct a high-resolution image degradation model. The original noisy image data is decomposed by wavelet based on the wavelet transform method. The number of wavelet decomposition layers is n. The decomposition is terminated when there is no noise in each sub-image after decomposition. The noisy sub-image containing stripe noise is screened out and the sub-image without noise is filtered out. The noisy subgraph processing module is used to establish the energy functional of each noisy subgraph after wavelet decomposition, and construct a one-way total variation model for the noisy subgraph, apply the fidelity term and the regularization term, use the alternating direction multiplier method to solve the functional minimum, set the tolerance to end the iterative convergence, and obtain the processed noisy subgraph; An image reconstruction module is used to reconstruct and restore the processed denoised sub-image and the noise-free sub-image by using an inverse wavelet transform to obtain the entire denoised image; The evaluation module is used to perform quantitative evaluation on the entire denoised image, and repeat the denoising if it does not meet the standards.

[0022] Specifically, the above-mentioned preprocessing module, construction module, noisy sub-image processing module, image reconstruction module and evaluation module can be embedded in a computer processing system. The computer calls the above-mentioned modules to complete the task of fault diagnosis of the star catalog intelligent system based on the above-mentioned Mars high-resolution image functional analysis de-banding method; the above-mentioned preprocessing module, construction module, noisy sub-image processing module, image reconstruction module and evaluation module can perform operations according to the specific steps given in the Mars high-resolution image functional analysis de-banding method.

[0023] It should be noted that it should be understood that the division of the various modules of the above system is only the division of logical functions. In actual implementation, they can be fully or partially integrated into one physical entity, or they can be physically separated, and these modules can all be implemented in the form of software calling through processing elements; they can also be all implemented in the form of hardware; some modules can also be implemented in the form of software calling through processing elements, and some modules can be implemented in the form of hardware. For example, the preprocessing module can be a separately established processing element, or it can be integrated in a certain chip of the above device. In addition, it can also be stored in the memory of the above device in the form of program code, and called and executed by a certain processing element of the above device. The implementation of other modules is similar. In addition, these modules can be fully or partially integrated together, or they can be implemented independently. The processing element described here can be an integrated circuit with signal processing capabilities. In the implementation process, each step of the above method or each module above can be completed by the hardware integrated logic circuit in the processor element or the instructions in the form of software.

[0024] For example, the above modules may be one or more integrated circuits configured to implement the above methods, such as one or more application specific integrated circuits (ASIC), or one or more digital singnal processors (DSP), or one or more field programmable gate arrays (FPGA). For another example, when a module is implemented in the form of a processing element scheduling program code, the processing element may be a general-purpose processor, such as a central processing unit (CPU) or other processors that can call program code. For another example, these modules may be integrated together and implemented in the form of a system-on-a-chip (SOC).

[0025] Embodiment three: The present invention provides a terminal device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor. The memory stores a computer program capable of running on the processor, and when the processor loads and executes the computer program, the above-mentioned Mars high-resolution image functional analysis de-banding method is adopted.

[0026] It should be noted that the terminal device can be a computer device such as a desktop computer, a laptop computer or a cloud server, and the terminal device includes but is not limited to a processor and a memory. For example, the terminal device can also include input and output devices, network access devices and buses.

[0027] Furthermore, the processor may adopt a central processing unit (CPU). Of course, according to actual usage, other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. may also be adopted. The general-purpose processor may adopt a microprocessor or any conventional processor, etc., and this application does not impose any restrictions on this.

[0028] Embodiment 4: The present invention provides a storage medium containing computer executable instructions, and the computer executable instructions are used to execute the above-mentioned Mars high-resolution image functional analysis de-banding method when executed by a computer processor.

[0029] Among them, the computer program can be stored in a computer-readable medium, the computer program includes computer program code, the computer program code can be in the form of source code, object code, executable file or certain middleware, etc. The computer-readable medium includes any entity or device that can carry the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the computer-readable medium includes but is not limited to the above-mentioned components.

[0030] 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 any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0031] For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to the specific circumstances. When an element is referred to as being "assembled on", "installed on", "fixed on" or "set on" another element, it can be directly on the other element or there can also be a centered element. When an element is considered to be "connected" to another element, it can be directly connected to the other element or there may be a centered element at the same time. The terms "vertical", "horizontal", "up", "down", "left", "right" and similar expressions used herein are only for illustrative purposes and are not intended to be the only implementation method.

[0032] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

[0033] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present disclosure. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

Claims

1. A functional analysis method for removing banding from Mars high-resolution images, characterized in that: The following steps are involved: Receiving high-resolution Mars image data containing stripe noise as raw data input and performing preprocessing; A high-resolution image degradation model is constructed. The original noisy image data is decomposed by wavelet transform method. The number of wavelet decomposition layers is n. The decomposition is terminated when there is no noise in each sub-image after decomposition. The noisy sub-images containing stripe noise are screened out and the sub-images without noise are filtered out. For all noisy subgraphs after wavelet decomposition, the energy functional of each noisy subgraph is established, and a one-way total variation model is constructed for the noisy subgraph. The fidelity term and regularization term are applied, and the alternating direction multiplier method is used to solve the functional minimum. The tolerance is set to end the iterative convergence, and the processed noisy subgraph is obtained. The inverse wavelet transform is used to reconstruct and restore the processed denoised sub-image and the sub-image without noise to obtain the entire denoised image; A quantitative evaluation is performed on the entire denoised image, and repeated denoising is performed if it does not meet the standard.

2. The method for removing banding by functional analysis of high-resolution Mars images according to claim 1, characterized in that: The high-resolution image data of Mars containing stripe noise is acquired through satellite collection, and the raw data is preprocessed, including: (21) Determine the image format to determine whether it is a single-band image or a multi-band image; (22) If it is a multi-band image, it is processed band by band and finally synthesized into one image; (23) Determine whether the direction of the stripe noise in the image is horizontal, vertical or diagonal.

3. The method for removing banding by functional analysis of high-resolution Mars images according to claim 2, characterized in that: Construct a high-resolution image degradation model as follows: For two-dimensional The degradation model of high-resolution images containing stripe noise is expressed as: In the formula , represents an image with stripe noise, represents an image without stripe noise, Indicates banding noise.

4. The method for removing banding by functional analysis of high-resolution Mars images according to claim 3, characterized in that: The original noisy image data is decomposed by wavelet transform method. The number of wavelet decomposition layers is n. The decomposition is terminated when there is no noise in each sub-image after decomposition. The noisy sub-image containing stripe noise is screened out and the sub-image without noise is filtered out. The details are as follows: For the original noisy image For , its two-dimensional discrete wavelet transform is expressed as: in is the component of each sub-graph, To decompose the scale, Represents the domain space, assuming that the original image contains noise Recorded as , then its multi-layer decomposition is: In the formula and yes and The conjugate matrix of is the number of decomposition layers, , , and Respectively represent The low-frequency component, horizontal component, vertical component and diagonal component after layer decomposition.

5. The method for removing banding by functional analysis of high-resolution Mars images according to claim 4, characterized in that: For all noisy subgraphs after wavelet decomposition, the energy functional of each noisy subgraph is established, and a one-way total variation model is constructed for the noisy subgraph. The fidelity term and regularization term are applied, and the alternating direction multiplier method is used to solve the functional minimum. The tolerance is set to end the iterative convergence, and the processed noisy subgraph is obtained, as follows: (51) The energy functional of each noisy subgraph is established and expressed as: In the formula For the fidelity item; is the regularization term; (52) For noisy images In a bounded space The energy total variation model is constructed as: In the formula For the fidelity item; is the regularization term, The total variation of ; is the Lagrangian factor, which is used to adjust the effect of the balance fidelity term and the regularization term; Therefore, image denoising becomes solving the energy functional Minimum problem: in For the denoised image, solving the functional minimum is transformed into solving the Euler-Lagrange equation problem: In the formula, represents an image with stripe noise, represents an image without stripe noise, yes The gradient of is the Lagrangian factor.

6. The method for removing banding by functional analysis of high-resolution Mars images according to claim 5, characterized in that: The inverse wavelet transform is used to reconstruct and restore the processed denoised sub-image and the noise-free sub-image to obtain the entire denoised image, as follows: In the formula, and yes and The conjugate matrix of is the number of decomposition layers, , , and Respectively represent The low-frequency component, horizontal component, vertical component and diagonal component after layer decomposition.

7. The method for removing banding by functional analysis of high-resolution Mars images according to claim 6, characterized in that: The entire denoised image is quantitatively evaluated using three quantitative indicators: peak signal-to-noise ratio, structural similarity, and relative dimensionless global comprehensive error, as shown below: (71) The peak signal-to-noise ratio is calculated as follows: In the formula, is the maximum gray value of the image, and are the number of rows and columns of image pixels, and They are respectively the image before and after denoising. Line The pixel value of the column; (72) The structural similarity value ranges from 0 to 1 and is defined as follows: In the formula, and These are the two images before and after denoising. and The average value of and Represent the standard deviation of the two images before and after denoising, and are two positive integers; (73) Relative dimensionless global comprehensive error, specifically defined as follows: In the formula, Represents the ratio of the spatial resolution of the original image to that of the denoised image, Indicates the number of bands of the image, Represents the original image and the denoised image The root mean square error of the bands, The original image The mean of the bands; (74) Set SSIM ≥ 0.95 and ERGAS ≤ 10 for ideal processing effect, otherwise perform repeated denoising.

8. A Mars high-resolution image functional analysis and striping system, used to implement the Mars high-resolution image functional analysis and striping method according to any one of claims 1 to 7, characterized in that: include: A preprocessing module is used to receive high-resolution Mars image data containing stripe noise as raw data input and perform preprocessing; The construction module is used to construct a high-resolution image degradation model. The original noisy image data is decomposed by wavelet based on the wavelet transform method. The number of wavelet decomposition layers is n. The decomposition is terminated when there is no noise in each sub-image after decomposition. The noisy sub-image containing stripe noise is screened out and the sub-image without noise is filtered out. The noisy subgraph processing module is used to establish the energy functional of each noisy subgraph after wavelet decomposition, and construct a one-way total variation model for the noisy subgraph, apply the fidelity term and the regularization term, use the alternating direction multiplier method to solve the functional minimum, set the tolerance to end the iterative convergence, and obtain the processed noisy subgraph; An image reconstruction module is used to reconstruct and restore the processed denoised sub-image and the noise-free sub-image by using an inverse wavelet transform to obtain the entire denoised image; The evaluation module is used to perform quantitative evaluation on the entire denoised image, and repeat the denoising if it does not meet the standards.

9. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that: The memory stores a computer program that can be run on the processor. When the processor loads and executes the computer program, the Mars high-resolution image functional analysis de-banding method described in any one of claims 1 to 7 is adopted.

10. A storage medium containing computer executable instructions, characterized in that: When the computer executable instructions are executed by a computer processor, they are used to perform the Mars high-resolution image functional analysis de-banding method according to any one of claims 1 to 7.

Images