Transformer infrared image enhancement method and system

Abstract

The invention discloses a transformer infrared image enhancement method and system, and the method comprises the steps: S1, inputting an original transformer infrared image, and carrying out the processing of the original transformer infrared image through an improved guided filtering algorithm, and obtaining a basic layer image; s2, obtaining a detail layer image by using a difference method; s3, performing nonlinear enhancement on the detail layer image obtained in the step S2, and outputting the enhanced detail layer image; s4, segmenting the image into a foreground layer image and a background layer image by adopting an Ostu algorithm; s5, enhancing the foreground layer image segmented in the step S4 by adopting a linear enhancement mode, and outputting the enhanced foreground layer image; s6, enhancing the background layer image segmented in the step S4 by adopting a histogram equalization algorithm, and outputting the enhanced background layer image; and S7, carrying out superposition based on the images processed in the steps S3, S5 and S6, and obtaining a final enhanced transformer infrared image. According to the method, the defect that noise is amplified while detail textures are enhanced in a traditional algorithm can be overcome.

Description

Technical Field

[0001] This invention belongs to the field of image processing technology for power equipment, and in particular relates to a method and system for enhancing infrared images of transformers. Background Technology

[0002] Transformers are crucial equipment in power systems. Statistics show that 70% of power transformer failures nationwide are caused by the gradual development of thermal defects. Infrared thermal imagers can convert transformer thermal radiation signals into thermal distribution images, and analysis of these images can detect early-stage thermal defects in transformers. However, because thermal radiation signals are typically weak and limited by detector imaging mechanisms and manufacturing processes, coupled with the complex electromagnetic environment of substations, infrared images suffer from drawbacks compared to visible light images, such as blurred details, poor contrast, and a narrower grayscale range. This not only affects the visual quality of the image but can also lead to misdiagnosis of subsequent faults. Therefore, image enhancement is necessary to improve the contrast and clarity of infrared images.

[0003] Traditional infrared image enhancement algorithms, such as histogram equalization and gamma transform, primarily enhance images by adjusting the grayscale histogram distribution or grayscale values. However, these methods often neglect the spatial correlation between pixels when processing infrared images, leading to problems such as loss of detail and texture, over-enhancement, and amplified noise in the enhanced image.

[0004] Prior art 1 discloses an infrared image enhancement method based on adaptive filtering layering (CN117422631A). This application discloses three steps: SSR image layering based on adaptive bilateral filtering, base layer image enhancement, and detail layer image enhancement. The base layer image uses the OTU algorithm to calculate segmentation points, and a maximum entropy-based dual histogram equalization method is used to enhance the base layer image. Finally, the enhanced base layer image and detail layer image are fused. However, prior art 1 has a drawback: the adaptive bilateral filtering algorithm is prone to gradient inversion during image enhancement, affecting the visual effect. This paper uses an improved guided filtering method to avoid the gradient inversion problem.

[0005] Prior art 2 discloses (CN118552462A) a method that uses guided filtering to obtain a base layer image, interpolation to obtain a detail layer image, nonlinear function enhancement of the detail layer image, and finally superimposition of the base layer enhancement image and the detail layer enhancement image. However, prior art 2 has the following shortcomings: This method uses a traditional guided filtering algorithm for image layering, making it difficult to achieve different smoothing levels for smooth regions and edge texture regions, resulting in weak preservation of image edge details; furthermore, it does not differentiate between the base layer image and the detail layer image after layering, leading to a lack of prominent foreground devices in the enhanced image; prior art 2 directly uses an arctan nonlinear function for noise suppression and detail enhancement in the detail layer decomposed from the infrared image, without considering the spatial detail features of different images, making it difficult to achieve good noise suppression and detail enhancement effects for various infrared images in the field.

[0006] To improve infrared image contrast, filter noise in infrared images of power equipment, and enhance image details, there is an urgent need to propose an enhancement algorithm based on improved guided filtering layering and OTU threshold segmentation. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention provides a transformer infrared image enhancement method and system, which solves problems such as neglecting pixel spatial correlation, loss of detail and texture, over-enhancement, and amplification of noise in traditional infrared image enhancement algorithms.

[0008] The present invention adopts the following technical solution.

[0009] The first aspect of this invention provides a transformer image enhancement method, comprising:

[0010] Step S1: Input the original transformer infrared image and process it using an improved guided filtering algorithm to obtain the base layer image;

[0011] Step S2: Based on the base layer image from step S1, obtain the detail layer image using the interpolation method;

[0012] Step S3: Based on the coefficients of the improved guided filtering algorithm in step S1 as a mask, perform nonlinear enhancement on the detail layer image obtained in step S2, and output the enhanced detail layer image.

[0013] Step S4: For the base layer image from step S1, the Ostu algorithm is used to segment it into a foreground layer image and a background layer image;

[0014] Step S5: The foreground layer image segmented in step S4 is enhanced using a linear enhancement method, and the enhanced foreground layer image is output.

[0015] Step S6: The background layer image segmented in step S4 is enhanced using a histogram equalization algorithm, and the enhanced background layer image is output.

[0016] Step S7: The images processed in steps S3, S5 and S6 are superimposed to obtain the final enhanced infrared image of the transformer.

[0017] Preferably, in step S1, the improved guided filtering algorithm is as follows:

[0018]

[0019] Where, q i改 To output the i-th pixel of the base layer image in the improved guided filtering algorithm, I represents the improved transform coefficient corresponding to pixel i in the original image. i The i-th pixel of the original infrared image; These are the transformation coefficients corresponding to pixel i in the original image.

[0020] Preferably, the improved transformation coefficients corresponding to pixel i in the original image are:

[0021]

[0022] Among them, a k改 For pixel block window w k The improved linear coefficients, where w is the number of pixels contained in the filter window, w k This is a filtering window centered at pixel k and containing pixel i.

[0023] Preferably, the pixel block window w k The improved linear coefficients are:

[0024]

[0025] in, For the k-th window w in the original infrared image k The variance of the inner pixels, p i μ is the pixel value of the i-th pixel in the original image. k To guide image I in window w k The average pixel value within the range, Let Γ be the mean of the original infrared image within the k-th window, ε be the smoothing coefficient, and Γ be the mean of the original infrared image within the k-th window. G (i) represents the edge weighting factor.

[0026] Preferably, the edge weight factor is:

[0027]

[0028] in, To guide the image, a window w centered on pixel i k within variance, To guide each pixel i in the image, a central window w is used. k The variance within the image, where N is the total number of pixels in the image, and γ σ This is the regularization factor.

[0029] Preferably, the formula for calculating the detail layer image using the interpolation method is as follows:

[0030] O = pq

[0031] In the formula, O is the detail layer image, p is the original infrared image, and q is the base layer image obtained by guided filtering.

[0032] Preferably, the enhanced detail layer image is as follows:

[0033]

[0034] O e This is the enhanced detail layer image, where O represents the detail layer image. These are the improved transformation coefficients corresponding to pixel i in the original image.

[0035] Preferably, in step S4, the optimal segmentation threshold of the Ostu algorithm is:

[0036] S'=argmax(l≤S≤L){δ 2 (S)}

[0037] Where S' is the optimal segmentation threshold, L is the gray level, l is 0, and δ 2 (S) represents the inter-class variance of the image.

[0038] Preferably, in step S5, the enhancement formula is:

[0039]

[0040] Where, x new To determine the corresponding grayscale value after enhancement, x min x max x represents the minimum and maximum gray values ​​of the foreground layer image, respectively. d The grayscale values ​​of the pixels in the image to be enhanced.

[0041] A second aspect of the present invention provides an infrared image enhancement system employing the above-described transformer infrared image enhancement method, comprising:

[0042] Improvements were made to the guided filtering processing module, the interpolation processing module, the nonlinear enhancement processing module, the Ostu algorithm segmentation module, the foreground layer image enhancement module, the background layer image enhancement module, and the image overlay module;

[0043] Among them, the improved guided filtering processing module is used to process the original transformer infrared image to obtain the base layer image by improving the guided filtering algorithm;

[0044] The interpolation processing module is used to obtain detail layer images based on base layer images using the interpolation method.

[0045] The nonlinear enhancement processing module is used to perform nonlinear enhancement on the detail layer of infrared images and output the enhanced detail layer image.

[0046] The Ostu algorithm segmentation module is used to segment the base layer image into a foreground layer image and a background layer image using the Ostu algorithm.

[0047] The foreground image enhancement module is used to enhance the foreground image using a linear enhancement method and output the enhanced foreground image.

[0048] The background image enhancement module is used to enhance the background image using a histogram equalization algorithm and output the enhanced background image.

[0049] The image overlay module is used to overlay the images processed by the nonlinear enhancement module, the foreground layer image enhancement module, and the background layer image enhancement module to obtain the final enhanced infrared image of the transformer.

[0050] The beneficial effects of this invention are that, compared with the prior art,

[0051] (1) This invention introduces an improved guided filtering processing module, which introduces an edge weight factor into the original guided filtering algorithm. The improved algorithm can adaptively adjust according to the smoothness of the infrared image, further enhancing the algorithm's ability to preserve image edge details and remove noise in smooth areas.

[0052] (2) The present invention uses the difference method to divide the original infrared image into two parts: the base layer and the detail layer. It performs enhancement processing on the different image information contained in the base layer and the detail layer respectively, so as to avoid amplifying noise information while enhancing image details.

[0053] (3) By performing nonlinear enhancement processing on the detail layer, the present invention suppresses the noise information of the detail layer while preserving the edge texture information of the detail layer, thereby achieving the purpose of further enhancing image details and suppressing noise.

[0054] (4) The present invention uses the Ostu algorithm to segment the infrared image of the transformer into a foreground layer image containing the main body of the electrical equipment and a background layer image containing environmental background information, and enhances the foreground layer image and the background layer image respectively to highlight the main body of the electrical equipment.

[0055] (5) The present invention uses an improved guided filter to divide the image into layers. High-frequency information such as noise and edge texture is filtered and then enters the detail layer. The detail layer is subjected to non-linear enhancement processing to enhance the details while suppressing noise information. The low-frequency information of the image enters the base layer and is processed separately by using OTU to highlight the visual effect of the foreground device in the infrared image. The two layers are processed separately to avoid the disadvantage of traditional algorithms that amplify noise while enhancing details and texture. Attached Figure Description

[0056] Figure 1 This is a schematic diagram of the transformer infrared image enhancement method in this invention. Detailed Implementation

[0057] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this application are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of this invention.

[0058] like Figure 1 As shown, this embodiment of the invention provides a method for enhancing infrared images of transformers, comprising the following steps:

[0059] Step S1: Input the original transformer infrared image and process it using an improved guided filtering algorithm to obtain the base layer image q.

[0060] The traditional guided filter formula is:

[0061]

[0062] Where, q i To output the i-th pixel of the base layer image, I i w is the i-th pixel of the original infrared image; k Let w be a 3×3 filtering window centered at pixel k and containing pixel i, where w is the number of pixels contained within the filtering window. and These are the transformation coefficients corresponding to pixel i in the original image. Since pixel i is surrounded by multiple w... k The included, therefore its transformation coefficients and Composed of multiple w containing pixel i k The corresponding a k and b k Find the average value, a k and b k For pixel block window wk The linear coefficients in p i μ is the pixel value of the i-th pixel in the original image. k To guide image I in window w k The average pixel value within the image is used; the original image is often used as the guide image. For the k-th window w in the original infrared image k The variance of the inner pixels, ε is the mean of the original infrared image in the k-th window, and ε is the smoothing coefficient. The larger ε is, the more obvious the smoothing effect.

[0063] In traditional guided filtering algorithms, ε is a fixed, manually specified constant. To achieve different smoothing levels in smooth regions and edge texture regions of the image, an edge weight factor Γ is introduced. G The ε is adaptively adjusted according to the smoothness of the image, which further enhances the algorithm's ability to preserve image edge details and remove noise in smooth areas.

[0064] The improved guided filter formula is as follows:

[0065]

[0066] Where, q i改 To output the i-th pixel of the base layer image in the improved guided filtering algorithm, Let a be the improved transform coefficient corresponding to pixel i in the original image. k改 For pixel block window w k The improved linear coefficients. Γ G With the improved coefficient a k改 As shown in the following formula:

[0067]

[0068] In the formula: To guide the image, a window w centered on pixel i k within variance, To guide each pixel i in the image, a central window w is used. k The variance within, Γ G (i) represents the edge weighting factor, which is the average of the variance of the current window divided by the variance of all windows in the entire image, where N is the total number of pixels in the entire image, and γ σ is the regularization factor, with a value of (0.001×L). 2 L represents the distribution range of image gray levels, p i μ is the pixel value of the i-th pixel in the original image. k To guide image I in window w k The average pixel value within the range.

[0069] If the pixel is located in a region of rapidly changing edge detail texture, then the variance of the window centered on that pixel will be large. G (i) will tend to be greater than 1, such that a k改 A value close to 1 better preserves edge details, while a value closer to 1 smooths out the edges. G (i) values ​​tend to be less than 1, thus making a k改 Further reduction can effectively suppress noise in the original infrared image, resulting in a smoother output.

[0070] Step S2: Based on the base layer image q obtained by the improved guided filtering in step S1, the detail layer image is obtained using the difference method;

[0071] The formula for calculating the detail layer image using the interpolation method is as follows:

[0072] O = pq

[0073] In the formula, O is the detail layer image, p is the original infrared image, and q is the base layer image obtained by guided filtering.

[0074] The detail layer image is obtained by subtracting the guided filter from the original infrared image of the transformer. The detail layer contains positive and negative data distributed around zero, which contains rich detail texture information and a lot of noise.

[0075] Step S3, based on the results obtained in step S1 The coefficients are used as a mask to perform nonlinear enhancement on the detail layer image obtained in step S2, highlighting image details while suppressing image noise, and outputting the enhanced detail layer image O. e .

[0076] In guided filtering The coefficients reflect the spatial details of an image to a certain extent; therefore, they can be used... This method masks noise in the detail layers, avoiding increased computation while achieving good denoising results. The operation is as follows:

[0077]

[0078] O e This is the enhanced detail layer image, where O represents the detail layer image.

[0079] This invention uses an improved guided filter. To construct a noise mask, an exponential operation is used; the exponent term is... Perform non-linear stretching to increase the size of detailed and flat areas. The difference in the value of the difference further enhances the suppression of noise in the detail layer and the enhancement of detail texture.

[0080] Step S4: For the base layer image obtained in step S1, the Ostu algorithm (Otsu's method, maximum inter-class variance method) is used to segment it into a foreground layer image and a background layer image.

[0081] Since the base layer image q generated after the original transformer infrared image is improved by guided filtering in step S1 mainly contains the main body of the transformer equipment and environmental background information, when the temperature is low, the gray values ​​of the electrical equipment and the background are similar, which affects the visual effect of the image. In order to further enhance the visual effect of the image and highlight the main body information of the transformer equipment, the Otsu algorithm is used to divide the infrared image into two parts: a background layer and a foreground layer. The foreground layer is F and the background layer is B. The foreground layer mainly contains the target body of the electrical equipment, while the background layer mainly contains environmental information. Different methods are used to enhance the two parts to highlight the main body of the electrical equipment.

[0082] The Ostu algorithm is briefly described below:

[0083] In a preferred but non-limiting embodiment of the present invention, the transformer infrared image is an image of size M×N, with a gray level of L and a range of [0, L-1]. Class P0 is defined as all pixels with a gray level of [0, S], and class P1 is defined as all pixels with a gray level of [S, L-1]. The probabilities of occurrence of classes P0 and P1 are P0(S) and P1(S), respectively, and the average gray levels are a0(S) and a1(S), respectively. Then the following equation holds:

[0084]

[0085] Among them, P i Let i be the probability of gray level i appearing in the image.

[0086] The average gray value of an image can be expressed as:

[0087] a = P0(S)a0(S) + P1(S)a1(S)

[0088] The inter-class variance of the image is:

[0089]

[0090] The maximum value of the inter-class variance is the optimal segmentation threshold.

[0091] S'=argmax(l≤S≤L){δ 2 (S)}

[0092] Where S' is the optimal segmentation threshold, L is the gray level, l is 0, and δ 2 (S) represents the inter-class variance of the image.

[0093] Step S5: The foreground image F segmented in step S4 is enhanced using a linear enhancement method, and the enhanced foreground image F' is output.

[0094] For the foreground image F extracted by Otsu, since the grayscale range of this part is small, a linear enhancement method is used to stretch the grayscale values ​​of this part. Let the grayscale range of the extracted foreground image be [x...]. min ,x max The range of grayscale values ​​after stretching is [x]. min

[255] , the grayscale value of the edge sub-band is stretched to the maximum extent. The grayscale enhancement formula is:

[0095]

[0096] Where, x new To determine the corresponding grayscale value after enhancement, x min x max x represents the minimum and maximum gray values ​​of the foreground layer image, respectively. d The grayscale values ​​of the pixels in the image to be enhanced.

[0097] Even after linear enhancement, the edge sub-band images retain their original image distribution characteristics while improving the visual effect of the target object. This enhances the recognizability and detail distribution of power equipment in infrared images, and the lower limit of grayscale values ​​in the enhanced edge portions remains unchanged. min This remains unchanged, which prevents the power equipment from blending into the background area, helps maintain the gray distribution pattern of the original power infrared image, and improves the visual effect of the power equipment area. The enhanced foreground layer image is named F'.

[0098] Step S6: The background layer image B segmented in step S4 is enhanced using the histogram equalization algorithm, and the enhanced background layer image B' is output.

[0099] For the background layer image B separated by Otsu, which belongs to the non-interest region in the full infrared image and mainly contains the surrounding environment, in order to solve the problem that the gray values ​​of electrical equipment and background are similar when the temperature is low, the histogram equalization algorithm (HE) is used to enhance the background layer image. HE adjusts the gray value of the background part of the image, so that the contrast between the main body of the electrical equipment and the background is increased after linear enhancement, and the visual difference is obvious, thereby reducing the interference of the background part on the foreground. The enhanced background layer image is named B'.

[0100] Step S7: The images processed in steps S3, S5, and S6 are superimposed to obtain the final enhanced infrared image of the transformer.

[0101] The formula is as follows:

[0102] I' = B' + F' + Oe

[0103] Where I' is the final enhanced infrared image of the transformer, B' is the enhanced background layer image, F' is the enhanced foreground layer image, and O... e This is the enhanced detail layer image.

[0104] To demonstrate the effectiveness of our proposed infrared image enhancement method for transformers, this paper compares it with histogram equalization, traditional guided filtering enhancement algorithms, and our method. Two metrics are used for evaluation: information entropy (IE) and average contrast (AC). Information entropy describes the amount of information in an image; a higher value indicates more information components and more detailed information. Average contrast reflects the contrast of an image; a higher value indicates higher image contrast and richer image information. The comparison table is as follows:

[0105]

[0106] The experimental results show that the algorithm presented in this paper outperforms traditional guided filtering enhancement algorithms and histogram equalization methods in both AC and IE metrics, while enhancing the image and preserving more image detail and texture information.

[0107] This invention also provides an infrared image enhancement system, based on the above-described method for enhancing infrared images of a transformer, comprising:

[0108] Improvements were made to the guided filtering processing module, the interpolation processing module, the nonlinear enhancement processing module, the Ostu algorithm segmentation module, the foreground layer image enhancement module, the background layer image enhancement module, and the image overlay module;

[0109] Among them, the improved guided filtering processing module is used to process the original transformer infrared image to obtain the base layer image by improving the guided filtering algorithm;

[0110] The interpolation processing module is used to obtain detail layer images based on base layer images using the interpolation method.

[0111] The nonlinear enhancement processing module is used to perform nonlinear enhancement on the detail layer of infrared images and output the enhanced detail layer image.

[0112] The Ostu algorithm segmentation module is used to segment the base layer image into a foreground layer image and a background layer image using the Ostu algorithm.

[0113] The foreground image enhancement module is used to enhance the foreground image using a linear enhancement method and output the enhanced foreground image.

[0114] The background image enhancement module is used to enhance the background image using a histogram equalization algorithm and output the enhanced background image.

[0115] The image overlay module is used to overlay the images processed by the nonlinear enhancement module, the foreground layer image enhancement module, and the background layer image enhancement module to obtain the final enhanced infrared image of the transformer.

[0116] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.

[0117] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination of the foregoing. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0118] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0119] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0120] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A method for enhancing infrared images of a transformer, characterized in that, include: Step S1: Input the original transformer infrared image and process it using an improved guided filtering algorithm to obtain the base layer image; Step S2: Based on the base layer image from step S1, obtain the detail layer image using the interpolation method; Step S3: Based on the coefficients of the improved guided filtering algorithm in step S1 as a mask, perform nonlinear enhancement on the detail layer image obtained in step S2, and output the enhanced detail layer image. Step S4: For the base layer image from step S1, the Ostu algorithm is used to segment it into a foreground layer image and a background layer image; Step S5: The foreground layer image segmented in step S4 is enhanced using a linear enhancement method, and the enhanced foreground layer image is output. Step S6: The background layer image segmented in step S4 is enhanced using a histogram equalization algorithm, and the enhanced background layer image is output. Step S7: The images processed in steps S3, S5 and S6 are superimposed to obtain the final enhanced infrared image of the transformer.

2. The transformer infrared image enhancement method according to claim 1, characterized in that: In step S1, the improved guided filtering algorithm is as follows: Where, q i改 To output the i-th pixel of the base layer image in the improved guided filtering algorithm, I represents the improved transform coefficient corresponding to pixel i in the original image. i The i-th pixel of the original infrared image; These are the transformation coefficients corresponding to pixel i in the original image.

3. The transformer infrared image enhancement method according to claim 2, characterized in that: The improved transformation coefficients corresponding to pixel i in the original image are: Among them, a k改 For pixel block window w k The improved linear coefficients, where w is the number of pixels contained in the filter window, w k This is a filtering window centered at pixel k and containing pixel i.

4. The transformer infrared image enhancement method according to claim 3, characterized in that: Pixel block window w k The improved linear coefficients are: in, For the k-th window w in the original infrared image k The variance of the inner pixels, p i μ is the pixel value of the i-th pixel in the original image. k To guide image I in window w k The average pixel value within the range, Let Γ be the mean of the original infrared image within the k-th window, ε be the smoothing coefficient, and Γ be the mean of the original infrared image within the k-th window. G (i) represents the edge weighting factor.

5. The transformer infrared image enhancement method according to claim 4, characterized in that: The edge weighting factor is: in, To guide the image, a window w centered on pixel i k within variance, To guide each pixel of the image The central window w k The variance within the image, where N is the total number of pixels in the image, and γ σ This is the regularization factor.

6. The transformer infrared image enhancement method according to claim 1, characterized in that: The formula for calculating the detail layer image using the interpolation method is as follows: O = pq In the formula, O is the detail layer image, p is the original infrared image, and q is the base layer image obtained by guided filtering.

7. The transformer infrared image enhancement method according to claim 1, characterized in that: The enhanced detail layer image is as follows: O e This is the enhanced detail layer image, where O represents the detail layer image. These are the improved transformation coefficients corresponding to pixel i in the original image.

8. The method for enhancing infrared images of a transformer according to claim 1, characterized in that: In step S4, the optimal segmentation threshold of the Ostu algorithm is: S'=argmax(l≤S≤L){δ 2 (S)} Where S' is the optimal segmentation threshold, L is the gray level, l is 0, and δ 2 (S) represents the inter-class variance of the image.

9. The method for enhancing infrared images of a transformer according to claim 1, characterized in that: In step S5, the enhancement formula is: Where, x new To determine the corresponding grayscale value after enhancement, x min x max x represents the minimum and maximum gray values ​​of the foreground layer image, respectively. d The grayscale values ​​of the pixels in the image to be enhanced.

10. An infrared image enhancement system employing the transformer infrared image enhancement method according to any one of claims 1-9, characterized in that, include: Improvements were made to the guided filtering processing module, the interpolation processing module, the nonlinear enhancement processing module, the Ostu algorithm segmentation module, the foreground layer image enhancement module, the background layer image enhancement module, and the image overlay module; Among them, the improved guided filtering processing module is used to process the original transformer infrared image to obtain the base layer image by improving the guided filtering algorithm; The interpolation processing module is used to obtain detail layer images based on base layer images using the interpolation method. The nonlinear enhancement processing module is used to perform nonlinear enhancement on the detail layer of infrared images and output the enhanced detail layer image. The Ostu algorithm segmentation module is used to segment the base layer image into a foreground layer image and a background layer image using the Ostu algorithm. The foreground image enhancement module is used to enhance the foreground image using a linear enhancement method and output the enhanced foreground image. The background image enhancement module is used to enhance the background image using a histogram equalization algorithm and output the enhanced background image. The image overlay module is used to overlay the images processed by the nonlinear enhancement module, the foreground layer image enhancement module, and the background layer image enhancement module to obtain the final enhanced infrared image of the transformer.

Images