Single mine image detail enhancement algorithm based on fractal dimension maximization

Abstract

The invention relates to the technical field of image enhancement, in particular to a single mine image detail enhancement algorithm based on fractal dimension maximization, and the algorithm comprises the steps: obtaining an original image, and calculating the gradient of the original image through a difference method; performing fractal analysis on the original image by taking the gradient of the image as a measure to obtain the local fractal length and the local fractal dimension data of the original image; partitioning the local forming dimension matrix, and comparing local forming dimensions in blocks to obtain maximum values in the blocks; performing pixel-by-pixel gradient reconstruction according to a fractal theory, and combining to obtain a theoretically enhanced image gradient; and calculating an enhanced image pixel value according to the reconstructed image gradient, the original image and the gradient of the original image, and finally reconstructing the image. Fractal analysis is carried out on the mine image by utilizing multi-scale calculation, many detail information in the image can be well captured, and by adopting a mode of maximizing fractal dimensions, not only can important details in the image be enhanced, but also the naturalness of the details and the sense of reality of the image can be maintained.

Description

Technical Field

[0001] The present application relates to the technical field of image enhancement, and particularly to a single-frame mine image detail enhancement algorithm based on maximizing the fractal dimension. Background Art

[0002] In the construction of intelligent mines, the application of digital technology has become an important means to promote the modernization of the mining industry, especially the extensive application of image processing and analysis technologies in mine monitoring, resource exploration, equipment maintenance, etc. With the gradual realization of "unmanned" and "intelligent" in intelligent mines, efficient image detail enhancement technology will provide stronger support for mine safety management, production monitoring, environmental monitoring, etc. Applying the image detail enhancement algorithm to the construction of intelligent mines can greatly improve the accuracy and efficiency of image processing, and bring innovative solutions to mine construction and management at multiple levels.

[0003] In various applications of intelligent mines, such as real-time monitoring of the ore mining process, fault diagnosis of mine equipment, and monitoring of the mining area environment, high-quality image data is required. The mine area is often in complex terrain and environmental conditions, and the problems of unclear details and large interference in images are relatively common. When traditional image enhancement methods (such as histogram equalization, image sharpening, etc.) enhance the image contrast, they are prone to over-amplify noise or generate serious artifacts, affecting the actual usability of the image. The images in intelligent mines are highly diverse, involving multiple aspects such as geological exploration, equipment monitoring, and personnel safety. The detail features and noise characteristics of each type of image are different. Mine images contain complex backgrounds, lighting changes, etc., which makes it difficult to optimize the image processing algorithm, and it is often difficult to distinguish between noise and details.

[0004] Existing technologies often tend to suppress noise or retain details, and it is difficult to achieve both at the same time. Although deep learning methods are effective, data annotation in the mine environment is difficult and requires a large amount of computing resources, especially in scenarios with high real-time requirements, which is a bottleneck. Summary of the Invention

[0005] The technical problem to be solved by the present invention is that in the existing image recognition algorithms, the effect of image detail enhancement processing is poor.

[0006] To this end, the present invention provides a single-frame mine image detail enhancement algorithm based on maximizing the fractal dimension.

[0007] The technical solution adopted by the present invention to solve its technical problems is:

[0008] A single-frame mine image detail enhancement algorithm based on maximizing the fractal dimension, comprising the following steps,

[0009] Step 1: Obtain the original image and calculate the gradients in the vertical and horizontal directions of the original image using the difference method;

[0010] Step 2: Conduct fractal analysis on the original image, using the image gradient as a metric to obtain the local fractal length and local fractal dimension data of the original image;

[0011] Step 3: Divide the original image into blocks, compare the magnitudes of the local fractal dimensions within the blocks, and obtain the maximum value of the local fractal dimension;

[0012] Step 4: Perform pixel-by-pixel gradient reconstruction and combine to obtain the enhanced image gradient;

[0013] Step 5: Calculate the pixel values of the enhanced image from the reconstructed image gradient, the original image, and its gradient, and finally reconstruct the image.

[0014] Further, in Step 1, use the pixel-by-pixel difference method to calculate the gradients in the vertical and horizontal directions of the original image where and represent the directional difference values in the X and Y directions respectively, and G r (x) represents the gradient of the pixel point in the image.

[0015] Further, in Step 2, regard a pixel point as a fractal set, consider the pixel point gradient as a metric of the pixel point, the dimension of the fractal set is α, and the density function corresponding to any point x in the fractal set is: where I(x) is the density function. In a digital image, I(x) is a constant, r represents the scale size of the fractal analysis, Kb(x) is a Gaussian kernel with a standard deviation of σ and a radius of b, and f(G r (x)) represents the measure value calculated using the image gradient as a metric, and z is the integration element introduced for the convolution integral.

[0016] Further, in Step 2, the local fractal dimension α(x) and the local fractal length S(x) at the pixel point x satisfy

[0017] Further, in Step 3, the reconstructed image satisfies: ln(f enhanced (G r (x)))=α m (x)ln2r + S(x), where f enhanced (G r (x)) represents the gradient value at the pixel point x of the reconstructed image, and α m (x) is the maximum value of the local fractal dimension in the local area.

[0018] Further, in step four, according to the gradient of the reconstructed image and the measure value calculated using the image gradient as a metric, the theoretically enhanced gradient value f enhanced (G r (x)):

[0019] Further, in step four, after magnifying the detail layer of the image, it is superimposed on the original image to enhance the image details.

[0020] Further, in step four, the difference between the enhanced gradient value and the original image gradient value is regarded as the detail layer, and then multiplied by an enhancement factor and superimposed on the original image to obtain the image with enhanced details: H enhanced = H + β(f enhanced (G r (x)) - f(G r (x))), where H represents the original image, and H enhanced represents the image with enhanced details, and β is the enhancement factor.

[0021] A computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps of the above method are implemented.

[0022] A computer device, comprising:

[0023] A memory, on which a computer program is stored;

[0024] A processor, configured to execute the computer program in the memory to implement the steps of the above method.

[0025] The beneficial effects of the present invention are:

[0026] The image detail enhancement algorithm of the present application can significantly improve the quality of mine images, help improve the level of automated monitoring and safety management, reduce the dependence on high-performance computing resources at the same time, and improve the real-time performance and economy of the system. This can not only optimize the production process, but also ensure the safety of miners and promote the development of mines towards intelligence and automation.

[0027] By adopting the method of maximizing the fractal dimension, the fractal dimension can better capture many detail information (such as rock structure, mineral distribution, equipment status, etc.) in the mining area image through multi-scale analysis. It can not only enhance the important details in the image, but also ensure the naturalness and authenticity of the details, which is particularly important for the application of intelligent mines. For example, in the environmental monitoring of mines, by enhancing the image details, abnormal changes or dangerous signs in the mining area can be better identified, providing effective support for mine safety management. Description of the Drawings

[0028] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0029] Figure 1 It is a flowchart of the single-mine-image detail enhancement algorithm based on maximizing the fractal dimension in the present invention.

[0030] Figure 2 It is a schematic diagram of the principle of the single-mine-image detail enhancement algorithm based on maximizing the fractal dimension in the present invention.

[0031] Figure 3 It is a schematic diagram of the regression fitting algorithm for local fractal dimension and local fractal length in the present invention.

[0032] Figure 4 It is a subjective comparison chart of the image processing results of the image detail enhancement algorithm in the present invention and other existing algorithms for the coal feeder in the underground air roadway.

[0033] Figure 5 It is a subjective comparison chart of the image processing results of the image detail enhancement algorithm in the present invention and other existing algorithms for the mine tunnel. Specific Embodiments

[0034] The present invention will now be further described in detail in conjunction with the accompanying drawings. These drawings are all simplified schematic diagrams, only illustrating the basic structure of the present invention in a schematic manner, so they only show the components related to the present invention.

[0035] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "transverse", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc. indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings. It 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, so it cannot be understood as a limitation of the present invention. In addition, the features defined as "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present invention, unless otherwise stated, the meaning of "a plurality" is two or more.

[0036] In the description of the present invention, it should be noted that unless otherwise clearly specified and limited, the terms "installation", "connection", and "coupling" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0037] Embodiment 1

[0038] A single-frame mine image detail enhancement algorithm based on maximizing the fractal dimension includes three modules, namely, a gradient calculation module, a fractal analysis module, and an image enhancement module.

[0039] The gradient calculation module is implemented by the per-pixel difference method.

[0040] Starting from the image gradient, the fractal analysis module performs data fitting by changing different scales, and finally solves the slope and intercept of the fitting line, which are the local fractal dimension and local fractal length of the image.

[0041] The image enhancement module first divides the original image into blocks, takes the maximum value of all local fractal dimensions within the block, then inversely calculates the theoretical gradient value of the enhanced image according to the relationship between the gradient values, then calculates the difference to obtain the detail layer, and then multiplies it by an enhancement factor and superimposes it on the original image to obtain the image with enhanced details.

[0042] Based on the above three modules, the specific single-frame mine image detail enhancement algorithm based on maximizing the fractal dimension includes the following steps:

[0043] Step 1: Obtain the original image, and use the per-pixel difference method to calculate the gradients in the vertical and horizontal directions of the original image to prepare for subsequent image fractal analysis.

[0044] Use the difference method to calculate the directional difference values respectively, and then use them to calculate the gradient:

[0045]

[0046] Where and represent the directional difference values in the X and Y directions respectively, and G r (x) represents the gradient of the image pixel point.

[0047] Step 2: Perform fractal analysis on the original image to obtain the local fractal length and local fractal dimension data of the original image to prepare for subsequent image enhancement.

[0048] If we consider each pixel point of an image as a fractal set, and the set of all pixel points as the union of a group of fractal sets, then the corresponding gradient can be regarded as a measure of it. If the dimension of the fractal set is α, then in the measure space defined by using the image gradient as a measure, the density function corresponding to any point x in the fractal set is:

[0049]

[0050] where I(x) is the density function, r represents the scale size of the fractal analysis, K b (x) is a Gaussian kernel with a standard deviation of σ and a radius of b, and f(G r (x)) represents the measure value calculated by using the image gradient as a measure, and z is the integration element introduced for the convolution integral.

[0051] Since the digital images we use now are such that each pixel point is uniformly distributed, the density function I(x) is a constant, and thus it can be deduced that:

[0052]

[0053] where α(x) represents the local fractal dimension at pixel point x, and S(x) is the local fractal length at the pixel point.

[0054] By performing data fitting by varying the scale r, a straight line in the ln-ln coordinate system can be obtained for the scale r and its corresponding measure value f(G r (x)). Then, by solving the slope and intercept of the fitted straight line (4), the local fractal dimension and local fractal length of the image can be approximately obtained:

[0055] ln(f(G r (x))) = α(x)ln 2r + S(x) (4)

[0056] Figure 3 Two examples in the fitting process are listed, by respectively selecting two pixel points in the picture 56028 in the BSDS200 dataset, Figure 3 (a) The small figure is the B channel of (21,37), and the regression fitting straight line analytical formula is ln(f) = 1.9894ln(2r) + 7.2253, obtaining a fractal dimension of 1.9894 and a local fractal length of 7.2253; in figure (b), the small figure is the G channel of (200,266), and the regression fitting straight line analytical formula is ln(f) = 2.0525ln(2r) + 7.4558, obtaining a fractal dimension of 2.0525 and a local fractal length of 7.4558.

[0057] Step 3: Divide the image into blocks, compare the size of the local fractal dimension within the block, and obtain the maximum local fractal dimension. If the original image satisfies formula (4), then the reconstructed image after the original image is divided into blocks also satisfies the relationship of formula (4). The reconstructed image satisfies:

[0058] ln(fenhance d (G r (x)))=α m (x)ln 2r+S(x) (5)

[0059] Among them, f enhanced (G r (x)) represents the gradient value at the reconstructed image pixel x, α m (x) is the maximum value of the fractal dimension in the local area.

[0060] Step 4: Perform pixel-by-pixel gradient reconstruction and combine to obtain the enhanced image gradient. The enhanced image pixel values are calculated from the reconstructed image gradient, the original image and its gradient, and finally the image is reconstructed.

[0061] From equations (4) and (5), we can deduce equation (6). So far, we have obtained the theoretical gradient value f of the enhanced image. enhanced (G r (x)):

[0062]

[0063] Furthermore, reconstructing the image includes:

[0064] An image can generally be decomposed into a smooth layer and a detail layer. Image detail enhancement can be performed by amplifying the detail layer and then superimposing it on the original image:

[0065]

[0066] Among them, H represents the original image, J represents the smoothing layer, K represents the detail layer, and H enhanced represents the image after detail enhancement, and β is the enhancement factor.

[0067] In this invention, the gradient value is regarded as the image detail layer. The difference is that the difference between the enhanced gradient value and the original image gradient value is regarded as the detail layer, which is then multiplied by an enhancement factor and superimposed on the original image to obtain the detail-enhanced image:

[0068] H enhanced =H+β(f enhanced (G r (x))-f(G r (x))) (8)

[0069] Furthermore, the overall process of the algorithm is shown in the following pseudocode:

[0070] Input: Original mine image

[0071] Output: High-quality mine image with enhanced details

[0072] For i = 1 to M:

[0073] For j = 1 to N:

[0074] First, calculate the directional difference values in the X and Y directions through Equation (1) Use them to calculate the image gradient Gr(x):

[0075] For r = 1 to R:

[0076] Perform Gaussian convolution operation on the image at each scale and store the results in a three-dimensional variable

[0077]

[0078] end

[0079] After slicing the three-dimensional variable, perform pixel-by-pixel regression analysis using Equations (2) - (4) to solve the linear equations and obtain the slope and intercept, that is, the local fractal dimension α(x) and local fractal length S(x) of the image

[0080] ln(f(G r (x))) = α(x)ln 2r + S(x)

[0081] Divide the local fractal dimension of the image into blocks, and take αm(x) for the local fractal dimension α(x) within the block;

[0082] According to the image fractal theory, perform gradient reconstruction using Equations (4) - (6)

[0083]

[0084] Perform detail enhancement using Equations (7) - (8)

[0085] H enhanced = H + β(f enhanced (G r (x)) - f(G r (x)))

[0086] end

[0087] end

[0088] In order to intuitively demonstrate the effect of the MFD (Detail Enhancement Method Based on Maximizing Fractal Dimension) algorithm in the present invention, the present invention provides a comparison between it and several other image detail enhancement algorithms. The compared algorithms include GIF, GGIF, WGIF, BFLS, TH and ILS, which represent common methods in the current field of image enhancement. In order to verify the superiority of the algorithm in the present invention, the present invention provides a comparison chart of its subjective visual effects and objective numerical evaluation indicators with other algorithms on classic data sets. In order to objectively evaluate the image quality, we selected the internationally used peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) as evaluation criteria, and formulas (9) and (10) are their calculation methods. PSNR is used to measure the noise and distortion of the image, and SSIM evaluates the structural fidelity of the image. The combination of the two can more comprehensively reflect the image quality.

[0089]

[0090] Where MAX is the maximum value of the pixel in the processed image, and MSE is the mean square error. x and μ y is the brightness mean of the original image and the image to be calculated, δ x and δ y are the variances of the original image and the image to be calculated, δ xy is the covariance between the original image and the image to be calculated, and a1 and a2 are constant terms that have no practical meaning and play a regulatory role.

[0091] Figure 4 and Figure 5 Two examples of mine operation sites are shown. Figure 4 This is a comparison chart of the recognition effect of the coal feeder image in the underground air lane. Figure 4In the small figure (a), it is the original image of the air roadway feeder. The small figure (a1) is the image processed by the GIF algorithm, and the maximum values in the numerical evaluations of PSNR and SSIM are 34.97 dB and 0.9978 respectively. The small figure (a2) is the image processed by the GGIF algorithm, and the values in the numerical evaluations of PSNR and SSIM are 30.88 dB and 0.9853 respectively. The small figure (a3) is the image processed by the WGIF algorithm, and the values in the numerical evaluations of PSNR and SSIM are 33.35 dB and 0.9976 respectively. The small figure (a4) is the image processed by the BFLS algorithm, and the values in the numerical evaluations of PSNR and SSIM are 27.87 dB and 0.9399 respectively. The small figure (a5) is the image processed by the TH algorithm, and the values in the numerical evaluations of PSNR and SSIM are 28.86 dB and 0.9700 respectively. The small figure (a6) is the image processed by the ILS algorithm, and the values in the numerical evaluations of PSNR and SSIM are 28.68 dB and 0.9610 respectively. The small figure (a7) is the image processed by the MFD algorithm in this application, and the values in the numerical evaluations of PSNR and SSIM are 35.68 dB and 0.9980 respectively. The small figure (a8) is the original image GT (ground truth).

[0092] Figure 5 It is a comparison chart of the recognition effects of mine tunnel images. Figure 5 In the small figure (b), it is the original image of the mine tunnel. The small figure (b1) is the image processed by the GIF algorithm, and the values in the numerical evaluations of PSNR and SSIM are 34.15 dB and 0.9842 respectively. The small figure (b2) is the image processed by the GGIF algorithm, and the values in the numerical evaluations of PSNR and SSIM are 31.09 dB and 0.9542 respectively. The small figure (b3) is the image processed by the WGIF algorithm, and the values in the numerical evaluations of PSNR and SSIM are 31.82 dB and 0.9698 respectively. The small figure (b4) is the image processed by the BFLS algorithm, and the values in the numerical evaluations of PSNR and SSIM are 29.22 dB and 0.8979 respectively. The small figure (b5) is the image processed by the TH algorithm, and the values in the numerical evaluations of PSNR and SSIM are 29.22 dB and 0.9083 respectively. The small figure (b6) is the image processed by the ILS algorithm, and the values in the numerical evaluations of PSNR and SSIM are 29.12 dB and 0.9046 respectively. The small figure (b7) is the image processed by the MFD algorithm in this application, and the values in the numerical evaluations of PSNR and SSIM are 36.15 dB and 0.9962 respectively. The small figure (b8) is the original image GT (ground truth).

[0093] Visually, overall, the algorithm in the present invention shows relatively ideal effects in both image detail enhancement and avoiding excessive amplification of noise. Other enhancement algorithms more or less have problems such as over-enhancement leading to picture distortion and serious artifacts. For example Figure 4 , 5 For other algorithms in, it can be clearly seen that the noise in the picture is overly amplified, the picture appears mosaic, and there is serious color distortion in some areas of the picture, with large areas of green and purple patches. This limits their full application in the construction of intelligent mines. In addition, the present invention not only has excellent detail enhancement effects, but also has a relatively small modification range for the original image, and has achieved the maximum values in the numerical evaluations of PSNR and SSIM.

[0094] The present invention also provides the test results of Mean Opinion Score (MOS). MOS test is a subjective standard method widely used in image and video quality assessment, especially in fields such as communication, video processing, and image compression. The MOS test reflects the perception of human observers on the quality of images or videos through the quantification of a group of user ratings.

[0095] Suppose there are N raters rating a certain image or video, then the calculation formula of MOS is formula (9), which is based on the opinions of a group of human observers on the quality of the image or video and calculates the average value.

[0096]

[0097] where score i is the rating of the i-th observer on the image.

[0098] According to the MOS test results (shown in Table 1), the algorithm in the present invention performs optimally or sub-optimally on the three datasets of BSDS200, RealSRSet, and T91, proving its excellent performance in image enhancement. Especially in the processing of low-quality images, this algorithm has significant advantages compared with traditional methods, can effectively improve the image quality, conforms to the visual perception of most people, and demonstrates its stability and superiority in improving visual effects.

[0099] Table 1 MOS test results

[0100]

[0101] To further illustrate the effectiveness of the algorithm, the present invention also provides the quantitative comparison results with other algorithms on different datasets, including BSDS200, RealSRSet, and T91. The algorithms to be compared include GIF, GGIF, WGIF, BFLS, TH, ILS, and ZF. The above algorithms are all common methods in the current field of image detail enhancement. As shown in Table 2. The quantitative results of the MFD algorithm model in this application on the three datasets all perform the best in the table.

[0102] Table 2 Quantitative Comparison Results

[0103]

[0104] Generally speaking, the algorithm in the present invention demonstrates powerful image enhancement capabilities, making the image details clearer and the visual effect more natural. For high-contrast and colorful images, the algorithm effectively retains the color saturation and detail levels, avoiding distortion caused by over-enhancement. The comparison results show that the algorithm not only improves the image quality but also effectively retains details, demonstrating stronger potential. Compared with traditional methods, this algorithm performs better in maintaining image details and structure, avoiding color distortion, excessive noise amplification, and severe gradient artifacts. Through the local fractal adaptive mechanism, the algorithm improves the visual effect while retaining the original structure of the image.

[0105] The detail enhancement method based on maximizing the fractal dimension of the present invention not only has significant academic value but also can play a huge potential in actual mine operations. It can further improve the quality and efficiency of image enhancement, providing more accurate support for intelligent mines. The application of this algorithm in the construction of intelligent mines can effectively solve the problems of traditional image enhancement technologies, improve the image analysis capabilities in aspects such as mine monitoring, equipment management, and environmental protection, and promote the development of mines towards intelligence, efficiency, and safety.

[0106] Embodiment 2

[0107] The embodiment of the present application provides a computer device, which includes a processor and a memory. At least one instruction or at least one program segment is stored in the memory, and the at least one instruction or the at least one program segment is loaded and executed by the processor to implement a single-frame mine image detail enhancement algorithm based on maximizing the fractal dimension as provided in the above method embodiment.

[0108] It should be noted that the above one or more processors and / or other data processing circuits can generally be referred to as "data processing circuits" herein. The data processing circuit can be embodied in software, hardware, firmware or any other combination, either in whole or in part. In addition, the data processing circuit can be a single independent processing module, or be integrated into other components of a computer device (or mobile device) in whole or in part. As involved in the embodiments of this application, the data processing circuit is a kind of processor control (such as the selection of a variable resistance terminal path connected to an interface). The memory can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to an intelligent diagnosis method for rolling bearing faults in the embodiments of this application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, to implement the above method. The memory can include high-speed random access memory, and can also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memories. In some instances, the memory can further include a memory remotely set relative to the processor, and these remote memories can be connected to the computer device through a network. Examples of the above network include but are not limited to the Internet, intranet, local area network, mobile communication network and their combinations. The transmission device is used to receive or send data via a network. Specific examples of the above network can include the wireless network provided by the communication provider of the computer device. In one instance, the transmission device includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station and thus communicate with the Internet. In one instance, the transmission device can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly. The display can be, for example, a touch-screen liquid crystal display (LCD), which enables the user to interact with the user interface of the computer device (or mobile device).

[0109] Embodiment 3

[0110] The embodiment of the present application also provides a computer-readable storage medium, which can be disposed in a server to store at least one instruction or at least one program related to an algorithm for enhancing details of a single mine image based on maximizing the fractal dimension in the method embodiment. The at least one instruction or the at least one program is loaded and executed by the processor to implement the algorithm for enhancing details of a single mine image based on maximizing the fractal dimension provided in the above method embodiment. Optionally, in this embodiment, the above storage medium may be located in at least one of multiple network servers in a computer network. Optionally, in this embodiment, the above storage medium may include, but is not limited to, various media that can store program codes, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk, or an optical disc.

[0111] Inspired by the ideal embodiment of the present invention described above, through the above description, relevant staff can completely make various changes and modifications without departing from the technical idea of the present invention. The technical scope of the present invention is not limited to the content in the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. A single-frame mine image detail enhancement algorithm based on maximizing the fractal dimension, characterized in that, including the following steps, Step 1: Obtain the original image and calculate the gradients of the original image in the vertical and horizontal directions using the difference method; Step 2: Conduct fractal analysis on the original image, using the image gradient as a metric to obtain the local fractal length and local fractal dimension data of the original image; Step 3: Divide the original image into blocks, compare the magnitudes of the local fractal dimensions within the blocks, and obtain the maximum value of the local fractal dimension; Step 4: Perform pixel-by-pixel gradient reconstruction and combine to obtain the enhanced image gradient; Step 5: Calculate the enhanced image pixel values from the reconstructed image gradient, the original image, and its gradient, and finally reconstruct the image.

2. The single-frame mine image detail enhancement algorithm based on maximizing the fractal dimension according to claim 1, characterized in that In step one, the gradients in the vertical and horizontal directions of the original image are calculated using the per-pixel difference method , where and represent the direction difference values in the X and Y directions respectively, represents the gradient of the pixel points in the image.

3. The single-frame mine image detail enhancement algorithm based on maximizing the fractal dimension according to claim 1, characterized in that In step two, a pixel is regarded as a fractal set, and the pixel gradient is regarded as a measure of the pixel. The dimension of the fractal set is , and for any point in the fractal set, the corresponding density function is: , where is the density function. In a digital image, is a constant, represents the scale size of the fractal analysis, is a Gaussian kernel with a standard deviation of and a radius of , represents the measure value calculated using the image gradient as the measure, and z is the integration element introduced for the convolution integral.

4. The single-frame mine image detail enhancement algorithm based on maximizing the fractal dimension according to claim 1, characterized in that In step two, the local fractal dimension at the pixel point , the local fractal length at the pixel point , satisfy .

5. The single-frame mine image detail enhancement algorithm based on maximizing the fractal dimension according to claim 4, characterized in that, In step three, the reconstructed image satisfies: , where represents the gradient value at the pixel point of the reconstructed image, is the maximum value of the fractal dimension within the local region.

6. The single-frame mine image detail enhancement algorithm based on maximizing the fractal dimension according to claim 5, characterized in that, In Step 4, based on the gradient of the reconstructed image and the measure value calculated using the image gradient as a metric, the theoretically gradient value of the enhanced image : .

7. The single-frame mine image detail enhancement algorithm based on maximizing the fractal dimension according to claim 1, characterized in that, In Step 4, the detail layer of the image is magnified and then superimposed on the original image to enhance the image details.

8. The single-frame mine image detail enhancement algorithm based on maximizing the fractal dimension according to claim 1, wherein, In Step 4, the difference between the enhanced gradient value and the original image gradient value is regarded as the detail layer, and then multiplied by an enhancement factor and superimposed on the original image to obtain the image with enhanced details: , where represents the original image, represents the image with enhanced details, is the enhancement factor.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the steps of the method described in any one of claims 1-8.

10. A computer device, characterized in that, Comprising: a memory storing a computer program thereon; a processor for executing the computer program in the memory to implement the steps of the method described in any one of claims 1-8.

Images