Underground coal mine low-illumination image enhancement method, system and equipment

Abstract

The invention relates to the technical field of image enhancement, and discloses an underground coal mine low-illumination image enhancement method, system and equipment, and the method comprises the steps: firstly, carrying out color correction to better solve a color cast problem under the abnormal illumination of an underground coal mine, and guaranteeing the color naturalness of an image; secondly, on the basis of a noisy Retinex model, three components of noise, illumination and reflection are decomposed, complex noise is removed, and meanwhile, a gamma correction mode is adopted to carry out balanced brightness adjustment; and finally, multi-scale fusion is carried out based on the weight to obtain a final enhanced image, the problem of detail loss of too dark and too bright areas is relieved, image enhancement under abnormal illumination is realized, the safety and rationality of subsequent coal mining operation identification and detection operation development are ensured, and the precision of underground coal mine low-illumination image enhancement is improved.

Description

Technical Field

[0001] The present invention relates to the field of image enhancement technology, and in particular to a method, system and device for low-illuminance image enhancement in a coal mine. Background Art

[0002] With the development of information construction of scientific and technological coal mines, digital video monitoring systems for underground coal mines have begun to be fully deployed. However, the actual video monitoring of underground coal mining faces difficulties such as complex underground environmental background, susceptibility to dust and fog interference, and loss of low-light image details. As a result, the final monitoring image has abnormal low illumination, local light overexposure or too dark, low contrast, missing detail feature information and color distortion. It is impossible to provide high-quality image data for subsequent tasks. Therefore, it is necessary to build an image quality enhancement technology suitable for the underground production environment of coal mines and process the low-illumination images of coal mines. At present, low-illumination image enhancement in coal mines has become one of the key research contents in the field of image processing. For low-illumination image enhancement, histogram equalization, gamma transform, Retinex model decomposition, image fusion and other methods are commonly used. However, since a single method has no obvious effect on local feature detail enhancement and will cause image overexposure, the existing image enhancement algorithm is applied to the actual coal mine underground low-illumination image enhancement. There is a large room for improvement in color contrast, brightness adjustment, detail features, noise estimation and other aspects, which leads to the low accuracy of the existing image enhancement algorithm in the image enhancement of low-illumination images in coal mines. Summary of the invention

[0003] The present invention provides a method, system and device for enhancing low-illuminance images in underground coal mines, so as to solve the problem of low precision of low-illuminance images in underground coal mines in the prior art.

[0004] In order to achieve the above object, the present invention is implemented by the following technical solutions:

[0005] In a first aspect, the present invention provides a method for enhancing low-illumination images in a coal mine, comprising:

[0006] S1: Convert the low-light image to be processed into RGB space to obtain image pixels;

[0007] S2: Calculate the one-dimensional discrete relative entropy of the image pixel points, and then perform gain coefficient constraint processing on the discrete relative entropy to obtain a color-corrected image;

[0008] S3: converting the color-corrected image from the RGB space to the HSI space, establishing a noisy Retinex model, and performing denoising and gamma correction processing on the color-corrected image based on the noisy Retinex model to obtain a processed image;

[0009] S4: performing multi-scale fusion based on a weight map on the processed image to obtain a fused image;

[0010] S5: Convert the fused image from the HSI space to the RGB space to obtain a final enhanced image.

[0011] In a second aspect, the present application provides a low-light image enhancement system for underground coal mines, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method described in the first aspect when executing the computer program.

[0012] In a third aspect, the present application provides a low-light image enhancement device in a coal mine, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method described in the first aspect when executing the computer program.

[0013] Beneficial effects:

[0014] The method for enhancing low-illuminance images in underground coal mines provided by the present invention, firstly, performs color correction to better solve the problem of color deviation under abnormal illumination in underground coal mines, and ensures the naturalness of image colors; secondly, decomposes three major components of noise, illumination, and reflection based on a noisy Retinex model, removes complex noise, and adopts a gamma correction method to perform balanced brightness adjustment; finally, multi-scale fusion is performed based on weights to obtain a final enhanced image, which alleviates the problem of detail loss in too dark and too bright areas, realizes image enhancement under abnormal illumination, ensures the safety and rationality of subsequent coal mining operation identification and detection operations, and improves the accuracy of image enhancement of low-illuminance images in underground coal mines. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 One of the flow charts of a method for enhancing low-light images in underground coal mines according to a preferred embodiment of the present invention;

[0016] Figure 2 This is a second flow chart of a method for enhancing low-light images in underground coal mines according to a preferred embodiment of the present invention;

[0017] Figure 3 is the original image of Example 1 in the preferred embodiment of the present invention;

[0018] Figure 4 The R, G, and B channel histograms of the original image of Example 1 of the preferred embodiment of the present invention;

[0019] Figure 5 Schematic diagram of grayscale histogram distribution of the original image of Example 1 in the preferred embodiment of the present invention;

[0020] Figure 6 An image processed by the method of this application;

[0021] Figure 7 The R, G, and B channel histograms of the image processed by the method of the present application;

[0022] Figure 8 The figure is a schematic diagram of the grayscale histogram distribution of an image processed by the method of the present application. DETAILED DESCRIPTION

[0023] The technical solution of the present invention is described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0024] Unless otherwise defined, the technical terms or scientific terms used in the present invention shall have the usual meanings understood by persons with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. Similarly, words such as "one" or "one" do not indicate quantity restrictions, but indicate the existence of at least one. Words such as "connect" or "connected" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the object being described changes, the relative positional relationship also changes accordingly.

[0025] See also Figure 1 The present application provides a method for enhancing low-light images in underground coal mines, comprising:

[0026] S1: Convert the low-light image to be processed into RGB space to obtain image pixels;

[0027] S2: Calculate the one-dimensional discrete relative entropy of the image pixel points, and then perform gain coefficient constraint processing on the discrete relative entropy to obtain a color-corrected image;

[0028] S3: converting the color-corrected image from the RGB space to the HSI space, establishing a noisy Retinex model, and performing denoising and gamma correction processing on the color-corrected image based on the noisy Retinex model to obtain a processed image;

[0029] S4: performing multi-scale fusion based on a weight map on the processed image to obtain a fused image;

[0030] S5: Convert the fused image from the HSI space to the RGB space to obtain a final enhanced image.

[0031] The above-mentioned method for low-illumination image enhancement in coal mines, firstly, performs color correction to better solve the problem of color deviation under abnormal illumination in coal mines, and ensures the naturalness of image color; secondly, decomposes the three major components of noise, illumination, and reflection based on the noisy Retinex model, removes complex noise, and adopts gamma correction to adjust the brightness evenly; finally, multi-scale fusion is performed based on weights to obtain the final enhanced image, which alleviates the problem of detail loss in too dark and too bright areas, realizes image enhancement under abnormal illumination, ensures the safety and rationality of subsequent coal mining operation identification and detection operations, and improves the accuracy of image enhancement of low-illumination images in coal mines.

[0032] Optionally, the one-dimensional discrete relative entropy of the image pixel points is calculated in S2, satisfying the following relationship:

[0033]

[0034] Among them, k represents 3 color channels, P k,i Indicates the proportion of pixels with gray value i of component k in the image;

[0035] The discrete relative entropy is subjected to gain coefficient constraint processing to satisfy the following relationship:

[0036]

[0037] in, Represents the three corrected gain coefficients, k r , k g , k b Indicates the initial gain coefficients of the three channels, represents the discrete relative entropy of the R channel, represents the discrete relative entropy of the G channel, Represents the discrete relative entropy of the B channel.

[0038] Optionally, in S3, the color correction image is converted from the RGB space to the HSI space, satisfying the following relationship:

[0039] The formula for calculating brightness I is as follows:

[0040]

[0041] In the formula, R represents R channel, G represents G channel, and B represents B channel;

[0042] The formula for calculating color saturation S is as follows:

[0043]

[0044] The formula for calculating hue H is as follows:

[0045]

[0046] Optionally, the noisy Retinex model is established in S3, and denoising and gamma correction are performed on the color-corrected image based on the noisy Retinex model, including:

[0047] A noisy Retinex model is established, and the color-corrected image is denoised based on the noisy Retinex model, satisfying the following relationship:

[0048] I cr =R·L+N;

[0049] Among them, I cr represents the color-corrected image of channel I in the HSI space, R represents the reflectance component, L represents the illumination component, N represents the noise component, and · represents element-by-element multiplication.

[0050] Gamma correction is performed on the denoised image to satisfy the following relationship:

[0051]

[0052] Among them, γ is the parameter that controls the degree of illumination enhancement, I en Represents the I channel image after gamma correction, I end Represents the reconstructed I channel image.

[0053] In this implementation, noise removal is first performed, and then gamma correction is performed on the illumination component in the denoised image, so that brightness adjustment can be achieved.

[0054] Optionally, the S4 includes:

[0055] S41: Calculate the aggregation weight of the synthesized image and normalize the aggregation weight to satisfy the following relationship:

[0056] W k =W L +W C +W S ;

[0057]

[0058] Among them, W S represents the significance weight, W C represents the color contrast weight, W L represents the Laplace contrast weight, Wk Represents the aggregation weight, W~ k represents the normalized aggregation weight, K represents the number of aggregation weight graphs, and δ represents the regularization term. In one example, it can be set to 0.1 to ensure the effect of each input on the output.

[0059] S42: Multi-scale fusion is performed based on the weight map, satisfying the following relationship:

[0060]

[0061] Among them, l represents the number of pyramid layers, and L{·} is the input original image I k Perform Laplace processing, G{·} table normalized weight graph Perform Gaussian filtering.

[0062] The steps of the above method are described in detail below with a complete embodiment.

[0063] Example 1: The data of this embodiment comes from a public dataset of underground coal mines. The image location is selected from a low-light scene in an underground tunnel. A method for enhancing low-light images in underground coal mines based on color correction and noisy multi-scale fusion is shown in the specific flow chart of this method. Figure 2 As shown, the specific implementation steps are as follows:

[0064] Step 1: First, read the low-light image of the underground tunnel of the coal mine and convert the image into RGB space;

[0065] Step 2: Calculate the one-dimensional discrete relative entropy according to the image pixels after spatial transformation, and use the improved gray world algorithm to constrain the gain coefficient of the discrete relative entropy of the R, G, and B three-channel images to achieve color correction of the original image;

[0066] The improved gray world algorithm is applied to the discrete relative entropy of the R, G, and B three-channel images to calculate the gain coefficient constraint. The improved R, G, and B three-channel "constrained" gain coefficient formula is as follows:

[0067]

[0068] in Represents the three corrected gain coefficients, where k r , k g , k b Indicates the initial gain coefficients of the three channels.

[0069] Step 3, convert the image obtained in step 2 from RGB space to HSI space;

[0070] Step 4: Establish a noisy Retinex model for the color-corrected image pixels obtained in step 3 and remove noise, perform gamma correction on the illumination component in the denoised image to achieve brightness adjustment;

[0071] The noisy Retinex model is established, and its specific formula is as follows:

[0072] I cr =R·L+N;

[0073] Among them, I cr represents the color-corrected image of channel I in the HSI space, R represents the reflectance component, L represents the illumination component, N represents the noise component, and · represents element-by-element multiplication.

[0074] Gamma correction is performed on the denoised pixels. The specific formula is as follows:

[0075]

[0076] Among them, γ is the parameter that controls the degree of illumination enhancement, I en Represents the I channel image after gamma correction, I end Represents the reconstructed I channel image.

[0077] Step 5: Perform multi-scale fusion based on the weight map on the image obtained in step 4;

[0078] Calculate the aggregation weight of the image and perform normalization. The specific formula is described as:

[0079] W k =W L +W C +W S ;

[0080]

[0081] Among them, W S represents the significance weight, W C represents the color contrast weight, W L represents the Laplace contrast weight, W k represents the aggregation weight, represents the normalized aggregation weight, K represents the number of aggregation weight graphs, and δ represents the regularization term, which is set to 0.1 here to ensure the effect of each input on the output.

[0082] Secondly, a multi-scale fusion process based on a weight map is performed. The specific formula is described as:

[0083]

[0084] Among them, l represents the number of pyramid layers, and L{·} is the input original image I k Perform Laplace processing, G{·} represents the normalized weight graph W k Perform Gaussian filtering.

[0085] Step 6: Convert the image in step 5 from HSI space to RGB channel to obtain the final enhanced image.

[0086] according to Figure 3 The original image of Example 1 (Example 1 refers to the implementation method of using typical pictures of the public data set as the original experimental data) and Figure 4 The original image R, G, B channel histogram of Example 1 and Figure 6 The image processed by the method of the present invention and Figure 7 Comparison of the R, G, and B channel histograms of the image processed by the method of the present invention shows that the distribution of the processed image on the three color channels is more balanced, and the dynamic range of the values ​​is also expanded. This shows that the color types of the image are richer and the color balance of the image is optimized, making the overall color appear more natural and realistic. In addition, according to Figure 5 The grayscale histogram distribution of the original image of Example 1 is Figure 8 The grayscale histogram distribution of the image processed by the method of the present invention shows that the processed image presents a more uniform distribution in grayscale level, especially in the high grayscale value area, which indicates that the overall brightness of the image has been effectively improved, indicating that the entropy of the processed image is improved compared with the original image, the details and complexity of the image are further improved, and the amount of information contained in the enhanced image is increased. Therefore, the image is visually brighter and has a stronger contrast, and the detailed information such as the miners and slogans in the scene is also clearer and richer.

[0087] The embodiment of the present application also provides a low-light image enhancement system for underground coal mines, including a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the steps of the above method when executing the computer program. The low-light image enhancement system for underground coal mines can implement various embodiments of the above method and achieve the same beneficial effects, which will not be described in detail here.

[0088] The present application also provides a low-light image enhancement device for underground coal mines, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the computer program. The low-light image enhancement device for underground coal mines can implement various embodiments of the above method and achieve the same beneficial effects, which will not be described in detail here.

[0089] The preferred specific embodiments of the present invention are described in detail above. It should be understood that a person skilled in the art can make many modifications and changes based on the concept of the present invention without creative work. Therefore, any technical solution that can be obtained by a person skilled in the art through logical analysis, reasoning or limited experiments based on the concept of the present invention on the basis of the prior art should be within the scope of protection determined by the claims.

Claims

1. A low-illumination image enhancement method for underground coal mines. It is characterized in that include: S1: Convert the low-light image to be processed into RGB space to obtain image pixels; S2: Calculate the one-dimensional discrete relative entropy of the image pixel points, and then perform gain coefficient constraint processing on the discrete relative entropy to obtain a color-corrected image; S3: converting the color-corrected image from the RGB space to the HSI space, establishing a noisy Retinex model, and performing denoising and gamma correction processing on the color-corrected image based on the noisy Retinex model to obtain a processed image; S4: performing multi-scale fusion based on a weight map on the processed image to obtain a fused image; S5: Convert the fused image from the HSI space to the RGB space to obtain a final enhanced image.

2. The method for enhancing low-light images in underground coal mines according to claim 1, It is characterized in that The one-dimensional discrete relative entropy of the image pixel is calculated in S2 to satisfy the following relationship: Among them, k represents 3 color channels, P k,i Indicates the proportion of pixels with gray value i of component k in the image; The discrete relative entropy is subjected to gain coefficient constraint processing to satisfy the following relationship: in, Represents the three corrected gain coefficients, k r , k g , k b Indicates the initial gain coefficients of the three channels, represents the discrete relative entropy of the R channel, represents the discrete relative entropy of the G channel, Represents the discrete relative entropy of the B channel.

3. The method for enhancing low-light images in underground coal mines according to claim 1, It is characterized in that In S3, the color correction image is converted from the RGB space to the HSI space, satisfying the following relationship: The formula for calculating brightness I is as follows: In the formula, R represents R channel, G represents G channel, and B represents B channel; The formula for calculating color saturation S is as follows: The formula for calculating hue H is as follows:

4. The method for enhancing low-light images in underground coal mines according to claim 1, It is characterized in that In S3, a noisy Retinex model is established, and denoising and gamma correction are performed on the color-corrected image based on the noisy Retinex model, including: A noisy Retinex model is established, and the color-corrected image is denoised based on the noisy Retinex model, satisfying the following relationship: I cr =R·L+N; Among them, I cr represents the color-corrected image of channel I in HSI space, R represents the reflectance component, L represents the illumination component, N represents the noise component, and · represents element-by-element multiplication; Gamma correction is performed on the denoised image to satisfy the following relationship: Among them, γ is the parameter that controls the degree of illumination enhancement, I en Represents the I channel image after gamma correction, I end Represents the reconstructed I channel image.

5. The method for enhancing low-light images in underground coal mines according to claim 1, It is characterized in that The S4 includes: S41: Calculate the aggregation weight of the synthesized image and normalize the aggregation weight to satisfy the following relationship: IN k =In L +W C +W S ; Among them, W S represents the significance weight, W C represents the color contrast weight, W L represents the Laplace contrast weight, W k represents the aggregation weight, represents the normalized aggregation weight, K represents the number of aggregation weight graphs, and δ represents the regularization term; S42: Multi-scale fusion is performed based on the weight map, satisfying the following relationship: Among them, l represents the number of pyramid layers, and L{·} is the input original image I k Perform Laplace processing, G{·} represents the normalized weight graph W~ k Perform Gaussian filtering.

6. A low-light image enhancement system for underground coal mines, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, It is characterized in that When the processor executes the computer program, the steps of the method described in any one of claims 1 to 5 are implemented.

7. A low-light image enhancement device for underground coal mines, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, It is characterized in that When the processor executes the computer program, the steps of the method described in any one of claims 1 to 5 are implemented.

Images