Tunnel dust environment image intelligent restoration method and system based on defogging algorithm

Abstract

The present application belongs to the field of image processing, and provides a tunnel dust environment image intelligent recovery method and system based on a defogging algorithm, which obtains a to-be-recovered image from a dim dusty environment in a tunnel; performs frequency domain transmittance optimization based on the to-be-recovered image to obtain a final transmittance, performs atmospheric light clustering optimization on the to-be-recovered image to obtain an atmospheric light value, performs image recovery by using the final transmittance and the atmospheric light value, and obtains a dim recovery image; performs spatial conversion based on the dim recovery image, performs hybrid enhancement on the converted dim recovery image, and converts the dim recovery image back to the original space to obtain a defogging recovery image. The present application solves the problems of tunnel dim dusty environment, influence on image quality, interference on recognition and detection accuracy, etc.

Description

Technical Field

[0001] This invention belongs to the field of image processing technology, specifically relating to a method and system for intelligent image restoration of tunnel dust environment based on dehazing algorithm. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] With the development of intelligent technologies, intelligent detection methods are increasingly being applied to tunnel construction quality inspection, with technologies such as image recognition and machine vision becoming more widely used. However, due to their enclosed and unique environment, tunnels often experience high concentrations of dust and dim lighting. Dust particles significantly scatter and absorb light, leading to a severe decline in the image quality captured by imaging equipment, manifesting as reduced contrast, blurred details, and color distortion. In tunnel inspection and management, image quality directly affects the accuracy and reliability of the results. High-resolution images can more accurately reflect the actual conditions inside the tunnel. The application of intelligent image recovery methods for tunnel dust environments can significantly improve image visibility and quality, providing clearer and more accurate source data for subsequent image analysis and processing, and offering stronger support for tunnel construction quality inspection and management.

[0004] However, existing methods for restoring images of dusty tunnel environments still suffer from insufficient restoration accuracy. Furthermore, during the image restoration process, inaccurate or missing information in the color space leads to a significant decrease in the accuracy and precision of the restored images. Summary of the Invention

[0005] To address the aforementioned issues, this invention proposes an intelligent image restoration method and system for tunnel dusty environments based on a defogging algorithm. This invention solves the problems of image quality being affected by the dim and dusty environment of tunnels, and the interference with recognition and detection accuracy.

[0006] According to some embodiments, the first solution of the present invention provides an intelligent image restoration method for tunnel dust environment based on a dehazing algorithm, which adopts the following technical solution: A method for intelligent image restoration of tunnel dust environments based on dehazing algorithms includes: Acquire the image to be restored from the dim, dusty environment of the tunnel; The final transmittance is obtained by frequency domain transmittance optimization of the image to be restored, and atmospheric light value is obtained by atmospheric light clustering optimization of the image to be restored. The final transmittance and atmospheric light value are used to restore the image and obtain the dimly restored image. Spatial transformation is performed on the dark restored image, and the transformed dark restored image is then blended and enhanced before being converted back to the original space to obtain the dehazed restored image.

[0007] Furthermore, the step of optimizing the frequency domain transmittance based on the image to be restored to obtain the final transmittance includes: Determine the initial transmittance based on the image to be restored; Frequency domain transmittance optimization was performed on the low-frequency and high-frequency components of the initial transmittance to obtain the optimized low-frequency and high-frequency components. The final transmittance is obtained by reconstructing the low-frequency and high-frequency components.

[0008] Further, the atmospheric light value is obtained by performing atmospheric light clustering optimization on the image to be restored, including: The image to be restored is divided into multiple superpixel blocks; Divide the superpixel block with the largest pixel weight into multiple rectangular regions; Each decomposed region in the image to be restored after two decompositions is clustered to obtain atmospheric candidate regions, and the atmospheric light value of the atmospheric candidate regions is used as the final atmospheric light value.

[0009] Furthermore, the step of performing spatial transformation based on the dimly lit restored image, blending and enhancing the transformed dimly lit restored image, and then converting it back to the original space to obtain a dehazed restored image includes: A color space conversion function is established based on the image color space to convert the RGB space of the dimly restored image to the HSV space, thus obtaining the converted dimly restored image. Based on the converted dark restored image, cross-color space joint enhancement is performed to obtain the luminance channel, the corrected hue channel, and the equalized saturation channel. Based on the image binarization function, the brightness channel of the converted dimly lit restored image is binarized to obtain the binarized restored image; Based on the labeling function, the brightness and darkness attributes of each region in the image are restored by labeling and binarization, and the dark areas of the image are extracted. Based on the attention mechanism and the dark areas of the image, the binarized restored image is enhanced to obtain an attention-enhanced brightness channel; The attention-enhanced brightness channel, the corrected hue channel, and the equalized saturation channel are converted back to the original RGB space to obtain the dehazed and restored image.

[0010] Furthermore, the process of using the final transmittance and atmospheric light value to perform image restoration and obtain a dimly lit restored image includes: Based on real-time dust concentration, variance of the grayscale image of the image to be recovered, and infrared thermal imaging temperature gradient, dust concentration is sensed to obtain the time-varying perturbation term of dust concentration. Based on the time-varying perturbation term of dust concentration, final transmittance, and atmospheric light value, image restoration is performed on the image to be restored, resulting in a dimly restored image.

[0011] Furthermore, the image restoration process based on the time-varying perturbation term of dust concentration, final transmittance, and atmospheric light value yields a dimly lit restored image, specifically as follows:

[0012] in, For pixels The image to be recovered at that location. For pixels The dimly lit image was restored. For pixels The final transmittance at that point It is the atmospheric light value. This is the time-varying perturbation term of dust concentration based on real-time dust concentration prediction.

[0013] According to some embodiments, the second aspect of the present invention provides an intelligent image restoration system for tunnel dust environments based on a dehazing algorithm, employing the following technical solution: A smart image restoration system for tunnel dust environments based on dehazing algorithms includes: The image acquisition module is configured to acquire images to be recovered from the dim, dusty environment of the tunnel. The image restoration module is configured to perform frequency domain transmittance optimization on the image to be restored to obtain the final transmittance, perform atmospheric light clustering optimization on the image to be restored to obtain the atmospheric light value, and use the final transmittance and atmospheric light value to perform image restoration to obtain a dimly restored image. The image enhancement module is configured to perform spatial transformation based on the dimly lit restored image, perform blending enhancement on the transformed dimly lit restored image, and then convert it back to the original space to obtain the dehazed restored image.

[0014] According to some embodiments, a third aspect of the present invention provides a computer-readable storage medium.

[0015] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the intelligent restoration method for tunnel dust environment images based on a defogging algorithm as described in the first scheme above.

[0016] According to some embodiments, a fourth aspect of the present invention provides a computer device.

[0017] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the intelligent restoration method for tunnel dust environment images based on defogging algorithms as described in the first embodiment above.

[0018] According to some embodiments, a fifth aspect of the present invention provides a computer program product or computer program.

[0019] A computer program product or computer program includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the steps in the intelligent restoration method for tunnel dust environment images based on defogging algorithms as described in the first embodiment above.

[0020] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention comprehensively solves the degradation problem of dimly lit dusty tunnel images through a three-stage progressive architecture of physical restoration, color enhancement, and dark area brightening. First, frequency domain transmittance optimization effectively suppresses edge halos and block effects, while atmospheric light clustering optimization avoids interference from artificial light sources, resulting in a physically accurate dark recovery base map. Second, the base map is converted to HSV space for cross-color joint enhancement and binarization, accurately marking underexposed dark areas while globally improving contrast and color saturation. Finally, using this marking as attention guide, network enhancement is applied only to the dark areas, achieving high dynamic range compression with bright areas brightened and no overexposure in the bright areas. The image is then converted back to RGB space, ultimately outputting a dehazed and restored image. This method integrates dehazing, color correction, and local illumination, significantly improving target visibility and color fidelity in tunnel environments, and exhibits high robustness and real-time application potential. Attached Figure Description

[0021] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0022] Figure 1 This is a flowchart of the intelligent image restoration method for tunnel dust environment based on dehazing algorithm in an embodiment of the present invention. Figure 2 This is a flowchart of the image processing of the dust environment image restoration model in an embodiment of the present invention; Figure 3 This is a flowchart of the image processing of the dim environment image enhancement model in an embodiment of the present invention; Figure 4 It is the original tunnel image in the dust environment image restoration model input in this embodiment of the invention; Figure 5 This refers to the image transmittance of the image to be restored in this embodiment of the invention. Figure 6 This refers to the image transmittance after frequency domain transmittance optimization in this embodiment of the invention. Figure 7This is a schematic diagram of atmospheric optical spectrum clustering optimization in an embodiment of the present invention; Figure 8 This is the image output after processing by the dust environment image restoration model in this embodiment of the invention; Figure 9 is a schematic diagram of the conversion between RGB and HSV color spaces in an embodiment of the present invention; Figure 10 This is the image obtained after binarization post-processing in the embodiments of the present invention; Figure 11 This is the image output after processing by the dim environment image enhancement model in the embodiments of the present invention. Detailed Implementation

[0023] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0024] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0025] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0026] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0027] Example 1 like Figure 1As shown, this embodiment provides an intelligent image restoration method for tunnel dust environment based on a defogging algorithm. This embodiment uses the application of this method to a server as an example for illustration. It is understood that this method can also be applied to terminals, and can also be applied to systems including terminals, servers, and other components, and can be implemented through interaction between the terminal and the server. The server can be an independent physical server, a server cluster composed of multiple physical servers, or a distributed system. It can also be a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network servers, cloud communication, middleware services, domain name services, CDN security services, and big data and artificial intelligence platforms. The terminal can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, etc., but is not limited to these. The terminal and server can be directly or indirectly connected via wired or wireless communication, which is not limited herein. In this embodiment, the method includes the following steps: Step S1: Obtain the image to be restored from the dimly lit, dusty environment of the tunnel; Step S2: Optimize the frequency domain transmittance based on the image to be restored to obtain the final transmittance, optimize the atmospheric light clustering of the image to be restored to obtain the atmospheric light value, and use the final transmittance and atmospheric light value to restore the image and obtain the dimly restored image. Step S3: Perform spatial transformation based on the dark recovery image, perform blending enhancement on the transformed dark recovery image, and then transform it back to the original space to obtain the dehazed recovery image.

[0028] This embodiment provides an intelligent image restoration method for tunnel dust environment based on a dehazing algorithm. It mainly includes two parts: a dust environment image restoration model and a dim environment image enhancement model. Using this method, an image to be restored, collected in a dim dust environment within a tunnel, is input. After processing by the two models, a clear restored image can be obtained, as detailed below: Step S1: Obtain the image to be restored from the dimly lit, dusty environment of the tunnel, wherein the image to be restored is as follows: Figure 4 As shown; Step S2: Optimize the frequency domain transmittance based on the image to be restored to obtain the final transmittance, optimize the atmospheric light clustering of the image to be restored to obtain the atmospheric light value, and use the final transmittance and atmospheric light value to restore the image and obtain the dimly restored image. like Figure 2 As shown, a data-physical fusion model for dust environment image degradation is established, and the dust environment image degradation model is used to restore the image to be restored. Step S2.1: Determine the initial transmittance based on the image to be restored, optimize the low-frequency and high-frequency components of the initial transmittance in the frequency domain to obtain the optimized low-frequency component and optimized high-frequency component, and reconstruct the final transmittance based on the optimized low-frequency component and optimized high-frequency component. Specifically, the initial transmittance is determined based on the image to be restored, and wavelet decomposition is performed on the initial transmittance to obtain the low-frequency component. and high frequency components ; For low-frequency components An optimized low-frequency component is obtained by using an improved weighted guided filter. As shown in the following formula: (1); in, The image to be recovered. It is the weighting function of the weighted guided filter. In position The original transmittance value at that location, Based on target pixels The central local square window area These are the pixel coordinates in the image; For high frequency components Edge enhancement is performed to obtain optimized high-frequency components. As shown in the following formula: (2); in, It represents an image At pixel position The gradient at that point.

[0029] The final transmittance is obtained by wavelet reconstruction based on optimized low-frequency and high-frequency components. .

[0030] like Figure 5 As shown, the initial transmittance of the image to be restored, after frequency domain transmittance optimization, is as follows: Figure 6 As shown.

[0031] This embodiment effectively separates low-frequency true depth information from high-frequency noise / texture interference in the transmittance map through frequency domain operations, ensuring smooth transmittance in flat areas and maintaining sharp boundaries at abrupt depth changes. Image restoration using this optimized transmittance can fundamentally suppress edge halos, blockiness, and texture residue, resulting in clear object outlines in dimly lit restored images and avoiding white edges or halos produced by traditional dehazing methods at depth discontinuities. Frequency domain processing also offers advantages such as low computational complexity and high parallelism, making it suitable for hardware acceleration and meeting real-time requirements such as tunnel monitoring.

[0032] Step S2.2: Divide the image to be restored into multiple superpixel blocks, divide the superpixel block with the largest pixel weight into multiple rectangular regions, cluster each decomposed region in the image to be restored after two decompositions to obtain atmospheric candidate regions, and use the atmospheric light value of the atmospheric candidate region as the final atmospheric light value. First, the image to be recovered Divided into multiple superpixel blocks of the same size ,Right now: (3); in, It is any superpixel block after the image to be restored is divided into four equal parts. It is the index of the superpixel block. It represents the number of superpixel blocks; Then, the pixel weights are calculated for each superpixel block: (4); in, It is a superpixel block pixel weights, It is the average pixel value of any superpixel block. It is the standard deviation of the pixel values ​​of any superpixel block after the image to be restored is divided into four equal parts; Select the superpixel block with the largest pixel weight It is further divided into multiple rectangular regions of the same size: (5); in, It is any rectangular region after dividing the superpixel block with the largest pixel weight into four equal parts. It is the index of the rectangular region. It is the number of rectangular regions. ; It is a decomposition function.

[0033] Based on the undecomposed remaining pixel blocks and multiple rectangular regions, all decomposition regions of the image to be restored are selected. For each decomposition region, spectral clustering is performed in the RGB color space. The region with the brightest cluster center and the smallest variance is selected as the atmospheric light candidate region. A schematic diagram of atmospheric light value spectral clustering optimization is shown below. Figure 7 As shown, with For example.

[0034] The atmospheric light values ​​of the candidate atmospheric light regions were used as the final atmospheric light values. .

[0035] This embodiment uses clustering optimization to constrain the candidate range of atmospheric light values ​​to the true depth fog / dust region, significantly improving the robustness of atmospheric light estimation. The resulting dimly lit restored image has natural white balance and color fidelity, eliminating color casts and halo effects caused by erroneous atmospheric light, and providing a color-reliable base image for subsequent enhancement.

[0036] Step S2.3: Based on real-time dust concentration, final transmittance, and atmospheric light value, the image to be restored is performed using a dust environment image degradation model to obtain a dimly lit restored image. The calculation is as follows: (6); in, For pixels The image to be recovered at that location, For pixels The dimly lit image was restored. For pixels The final transmittance at a certain point describes the portion of light that is transmitted through the medium to the imaging device without being scattered. It is the atmospheric light value, also known as the ambient light intensity. The dust concentration time-varying perturbation term is based on real-time dust concentration prediction. Based on the real-time dust concentration sensed by the sensor, the dust concentration fluctuation law of the tunnel over time is learned by using, but not limited to, an LSTM network.

[0037] Among them, the time-varying perturbation term of dust concentration based on real-time dust concentration prediction The calculation process is as follows: Based on real-time dust concentration, the variance of the grayscale image of the image to be recovered, and the infrared thermal imaging temperature gradient, dust concentration is sensed to obtain the time-varying perturbation term of dust concentration. The calculation formula is: (7); in, For adaptive weighting coefficients, The real-time dust concentration is obtained directly from the dust sensor. It combines sensor measurements and image estimates to obtain a more robust and accurate concentration estimate, which serves as the time-varying disturbance term for dust concentration. It is a grayscale image of the image to be restored. variance It is the infrared thermal imaging temperature gradient of the image to be recovered.

[0038] In a controlled experimental chamber, coal dust with a particle size ≤10μm was used to simulate a tunnel environment. The concentration gradient was set to 0~500mg / m³ (interval of 50mg / m³). At each concentration level, 24-color standard color cards were photographed, the RGB mean was extracted, the attenuation rate matrix was calculated, and an RGB simulation experiment of dust environment images was carried out. Establish concentration The non-linear mapping function with RGB attenuation (taking the red channel as an example): (8); in, These are the fitting coefficients calibrated through experiments. It is the red channel decay rate function.

[0039] Based on the above three steps, the images to be recovered (such as those acquired in the tunnel environment) will be processed. Figure 4 Image restoration is performed on the image shown, resulting in the output image - a recovered image (as shown). Figure 8 (As shown).

[0040] Step S3: Perform spatial transformation based on the dark recovery image, perform blending enhancement on the transformed dark recovery image, and then transform it back to the original space to obtain the dehazed recovery image.

[0041] like Figure 3 As shown, a low-light environment image enhancement model is used to enhance the environment of a low-light restored image. The implementation process of the low-light environment image enhancement model is as follows: Step S3.1: Establish a color space conversion function based on the image color space to convert the RGB space of the dimly lit restored image to the HSV space, and obtain the converted dimly lit restored image; A color space conversion diagram is shown in Figure 9. The conversion function between RGB space and HSV space is as follows: (9); (10); (11); in, These represent the red, green, and blue color channels of a recovered image from a dimly lit image. These represent the hue, saturation, and brightness channels of a dimly lit image.

[0042] Step S3.2: Based on the converted, dimly lit image, perform cross-color space joint enhancement to obtain the luminance channel, the corrected hue channel, and the equalized saturation channel, including: based on The channel extraction function extracts the luminance channel of the converted, dimly lit restored image; where, The channel extraction function is: (12); in, —Low brightness Channel image, —Extracted after color space conversion Channel image, —Extracted after color space conversion The mean of the channel image, —Empirical threshold for image brightness; Color cast correction is performed on the tone channels of the converted, dimly lit restored image to obtain the corrected tone channels; among them, The channel color cast correction function is: (13); in, This is the color shift compensation coefficient related to dust concentration. These are the original values ​​of the tone channels in the converted, darkened, restored image. It is determined by dust concentration The resulting average hue shift.

[0043] Adaptive histogram equalization is performed on the saturation channels of the converted, dimly lit restored image to obtain equalized saturation channels; among which, The adaptive histogram equalization function for the channel is: (14); in, yes medium gray level Frequency of occurrence It was Dirac function, It is the converted, dimly lit image in coordinates Pixel intensity value at that location, It is an image indexed The marked area It is used to identify the grid positions of image blocks.

[0044] This embodiment utilizes an established color space conversion function to convert dimly lit restored images from RGB to HSV space, achieving complete separation of brightness, hue, and saturation. Based on this, the brightness channel is enhanced, the saturation channel is equalized, and the hue channel is corrected; these three are optimized independently without interference. Therefore, brightness enhancement does not cause hue shift or saturation abnormalities, resulting in output images with natural colors and moderate vibrancy. This completely solves the problems of "grayish fading" in dusty environments and color distortion caused by direct RGB enhancement.

[0045] Step S3.3: Based on the image binarization function, the brightness channel of the converted dimly lit restored image is binarized to obtain the binarized restored image.

[0046] The binarization function is: (15); (16); (17); in, It is the luminance channel of the converted, dark image. It is the image after being processed by the function. It is a binarized image reconstruction. yes exist Location Average window brightness Therefore The adaptive sensitivity parameters are constructed based on this. yes The mean brightness of the image, after binarization, is as follows: Figure 10 As shown.

[0047] Step S3.4: Based on the labeling function, label the brightness and darkness attributes of each region in the binarized image and extract the dark areas of the image; Bright and dark area marking and dark area extraction: Based on the marking function, the brightness and darkness attributes of each region of the binarized image are marked, and the dark areas of the image are extracted; The marking function is: (18); in, It refers to any region defined in the image. yes Minimum area brightness, yes Minimum area brightness, yes The average brightness of the area.

[0048] This embodiment processes the enhanced luminance channel based on an image binarization function, adaptively generating a binarized restored image. Then, a labeling function is used to identify the brightness attributes of each region in the image, accurately extracting the dark areas. This step, based on a more reliable enhanced luminance signal, eliminates interference from bright areas such as lamps and reflections, resulting in a high-confidence dark area mask, providing spatial guidance for subsequent precise local enhancement.

[0049] Step S3.5: Enhance the binarized restored image based on the attention mechanism and the dark areas of the image to obtain the attention-enhanced brightness channel; The attention mechanism utilizes a dark region attention enhancement network for image enhancement. This network can employ a U-Net architecture, adding dark region masks at skip connections and using a PatchGAN structure. A dark region SSIM constraint is added to the loss function, with the function formula as follows: (19); (20); in, It is the brightness value of the dark areas of the image. It is the output image after adaptive gamma correction, i.e., the attention-enhanced restored image; It is a dark area mask for the dark areas of an image, representing the position of each pixel. The possibility of it belonging to a dark area. It is the gamma correction function; The attention mechanism in this embodiment utilizes extracted dark area identifiers to locally enhance the luminance channel, focusing enhancement only on dark areas. This significantly brightens extremely dark areas such as shadows in tunnels, under vehicles, and corners, while preserving the original natural brightness of bright areas. This results in an output image where details in dark areas are clearly visible, while bright areas are neither overexposed nor have halos. The overall dynamic range is effectively compressed, and the balance of light and dark approaches human visual perception.

[0050] Step S3.6: Based on the image color space conversion function, convert the attention-enhanced brightness channel, the corrected hue channel, and the equalized saturation channel back to the original RGB space to obtain the dehazed and restored image.

[0051] Integrating the models and functions mentioned in the above steps yields a low-light environment image enhancement model, the model's flowchart is as follows. Figure 2 As shown, the image output by the dust environment image recovery model (such as...) Figure 8 As shown in the image, inputting the model will yield the enhanced image output by the model (e.g., ...). Figure 11 (As shown).

[0052] This step, based on spatial transformation, with cross-color joint enhancement and binarization marking as the core, and attention-based local enhancement as the means, achieves hierarchical optimization of "global color - brightness balance - dark area brightening", which significantly improves the visibility and visual naturalness of dark dust images in tunnels.

[0053] Example 2 This embodiment provides an intelligent image restoration system for tunnel dust environments based on a dehazing algorithm, including: The image acquisition module is configured to acquire images to be recovered from the dim, dusty environment of the tunnel. The image restoration module is configured to perform frequency domain transmittance optimization on the image to be restored to obtain the final transmittance, perform atmospheric light clustering optimization on the image to be restored to obtain the atmospheric light value, and use the final transmittance and atmospheric light value to perform image restoration to obtain a dimly restored image. The image enhancement module is configured to perform spatial transformation based on the dimly lit restored image, perform blending enhancement on the transformed dimly lit restored image, and then convert it back to the original space to obtain the dehazed restored image.

[0054] The examples and application scenarios implemented by the above modules and corresponding steps are the same, but are not limited to the content disclosed in Embodiment 1 above. It should be noted that the above modules, as part of the system, can be executed in a computer system such as a set of computer-executable instructions.

[0055] The descriptions of each embodiment in the above embodiments have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0056] The proposed system can be implemented in other ways. For example, the system embodiments described above are merely illustrative, and the division of modules described above is only a logical functional division. In actual implementation, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed.

[0057] Example 3 This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in the intelligent restoration method for tunnel dust environment images based on a defogging algorithm as described in Embodiment 1 above.

[0058] Example 4 This embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the intelligent restoration method for tunnel dust environment images based on defogging algorithm described in Embodiment 1 above.

[0059] Example 5 This embodiment provides a computer program product or computer program, including computer instructions stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the steps in the intelligent restoration method for tunnel dust environment images based on defogging algorithm described in Embodiment 1 above.

[0060] Those skilled in the art will understand that embodiments of the present invention can provide methods, systems, or computer program products. Therefore, the present invention can take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0061] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0062] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0063] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0064] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0065] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A method for intelligent image restoration of tunnel dust environment based on dehazing algorithm, characterized in that, include: Acquire the image to be restored from the dim, dusty environment of the tunnel; The final transmittance is obtained by frequency domain transmittance optimization of the image to be restored, and atmospheric light value is obtained by atmospheric light clustering optimization of the image to be restored. The final transmittance and atmospheric light value are used to restore the image and obtain the dimly restored image. Spatial transformation is performed on the dark restored image, and the transformed dark restored image is then blended and enhanced before being converted back to the original space to obtain the dehazed restored image.

2. The intelligent image restoration method for tunnel dust environment based on dehazing algorithm as described in claim 1, characterized in that, The process of optimizing the frequency domain transmittance based on the image to be restored to obtain the final transmittance includes: Determine the initial transmittance based on the image to be restored; Frequency domain transmittance optimization was performed on the low-frequency and high-frequency components of the initial transmittance to obtain the optimized low-frequency and high-frequency components. The final transmittance is obtained by reconstructing the low-frequency and high-frequency components.

3. The intelligent image restoration method for tunnel dust environment based on dehazing algorithm as described in claim 1, characterized in that, The atmospheric light value is obtained by performing atmospheric light clustering optimization on the image to be restored, including: The image to be restored is divided into multiple superpixel blocks; Divide the superpixel block with the largest pixel weight into multiple rectangular regions; Each decomposed region in the image to be restored after two decompositions is clustered to obtain atmospheric candidate regions, and the atmospheric light value of the atmospheric candidate regions is used as the final atmospheric light value.

4. The intelligent image restoration method for tunnel dust environment based on dehazing algorithm as described in claim 1, characterized in that, The process of spatial transformation based on the dimly lit restored image, followed by blending and enhancement of the transformed dimly lit restored image and then converting it back to the original space to obtain a dehazed restored image includes: A color space conversion function is established based on the image color space to convert the RGB space of the dimly restored image to the HSV space, thus obtaining the converted dimly restored image. Based on the converted dark restored image, cross-color space joint enhancement is performed to obtain the luminance channel, the corrected hue channel, and the equalized saturation channel. Based on the image binarization function, the brightness channel of the converted dimly lit restored image is binarized to obtain the binarized restored image; Based on the labeling function, the brightness and darkness attributes of each region in the image are restored by labeling and binarization, and the dark areas of the image are extracted. Based on the attention mechanism and the dark areas of the image, the binarized restored image is enhanced to obtain an attention-enhanced brightness channel; The attention-enhanced brightness channel, the corrected hue channel, and the equalized saturation channel are converted back to the original RGB space to obtain the dehazed and restored image.

5. The intelligent image restoration method for tunnel dust environment based on dehazing algorithm as described in claim 1, characterized in that, The process of image restoration using final transmittance and atmospheric light value to obtain a dimly lit restored image includes: Based on real-time dust concentration, variance of the grayscale image of the image to be recovered, and infrared thermal imaging temperature gradient, dust concentration is sensed to obtain the time-varying perturbation term of dust concentration. Based on the time-varying perturbation term of dust concentration, final transmittance, and atmospheric light value, image restoration is performed on the image to be restored, resulting in a dimly restored image.

6. The intelligent image restoration method for tunnel dust environment based on dehazing algorithm as described in claim 5, characterized in that, The image restoration process, based on the time-varying perturbation term of dust concentration, final transmittance, and atmospheric light value, yields a dimly lit restored image. Specifically: in, For pixels The image to be recovered at that location. For pixels The dimly lit image was restored. For pixels The final transmittance at that point It is the atmospheric light value. This is the time-varying perturbation term of dust concentration based on real-time dust concentration prediction.

7. A smart image restoration system for tunnel dust environment based on dehazing algorithm, characterized in that, include: The image acquisition module is configured to acquire images to be recovered from the dim, dusty environment of the tunnel. The image restoration module is configured to perform frequency domain transmittance optimization on the image to be restored to obtain the final transmittance, perform atmospheric light clustering optimization on the image to be restored to obtain the atmospheric light value, and use the final transmittance and atmospheric light value to perform image restoration to obtain a dimly restored image. The image enhancement module is configured to perform spatial transformation based on the dimly lit restored image, perform blending enhancement on the transformed dimly lit restored image, and then convert it back to the original space to obtain the dehazed restored image.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the intelligent restoration method for tunnel dust environment images based on defogging algorithm as described in any one of claims 1-6.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the intelligent restoration method for tunnel dust environment images based on defogging algorithm as described in any one of claims 1-6.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the steps in the intelligent restoration method for tunnel dust environment images based on a defogging algorithm as described in any one of claims 1-6.

Images