Novel brightness and color correction underwater image processing method

Abstract

The invention provides a novel brightness and color correction underwater image processing method, and belongs to the technical field of image processing. The method comprises the following steps: inputting an underwater image, extracting features of the underwater image, and calculating a base layer and a detail layer of the image according to the features; and obtaining a brightness correction image by recovering the brightness of the detail layer. Meanwhile, a maximum attenuation image of the underwater image is estimated, and details and textures of the brightness correction image are enhanced by using the maximum attenuation image, so that a color correction image is obtained. Then, calculating a gradient distribution diagram and a pixel stretching distribution diagram of the color correction diagram in a V channel in the HSV color model; and finally, fusing the two distribution diagrams to enhance the global contrast of the color correction diagram. According to the method, the visual quality of the underwater image is remarkably improved, the image is clearer and more real, and the method has a wide application prospect.

Description

Technical Field

[0001] The invention belongs to the technical field of image processing, and in particular is a novel underwater image processing method for brightness and color correction. Background Art

[0002] Underwater image enhancement is a research direction that has received widespread attention in the field of computer vision and image processing in recent years. With the continuous advancement of science and technology, especially in the context of improved image sensor and processor performance, the acquisition of underwater images has become more and more convenient.

[0003] In the field of marine science research, scientists need to accurately observe and record the behavior, species and distribution of marine life in order to gain a deeper understanding of the complexity of marine ecosystems. Through underwater image enhancement technology, researchers can improve the quality of marine life images, so that species characteristics and behaviors can be observed more clearly. Underwater image enhancement also plays an indispensable role in underwater archaeology and exploration activities. When divers explore shipwrecks or ancient underwater sites, clear images can help them accurately identify and record various objects and structures. In addition, with the increase in global awareness of marine resource management and protection, accurate monitoring of fish and their habitats is particularly critical for sustainable development. In general, underwater image enhancement technology has a wide range of applications and far-reaching significance in many fields such as scientific research, resource management, cultural exploration and artistic expression.

[0004] However, the particularity of the underwater environment often makes the quality of the acquired images lower than expected, which is affected by many factors such as water turbidity, light scattering and absorption, resulting in image blur, color distortion and reduced contrast. In order to address these challenges, a variety of underwater image enhancement methods have been proposed, which can be mainly divided into three types: model-based methods, model-free methods and data-driven methods.

[0005] Model-based methods construct a physical model of the imaging process of degraded images, invert the model using assumptions and prior knowledge, and thus restore clear images. However, these methods are highly sensitive to their underlying assumptions. Model-free methods improve the visual quality of underwater images by adjusting the distribution of pixel values. For example, researchers proposed a color and white balance correction method based on the gray world principle and nonlinear color mapping function to improve the color and white balance distortion problems in underwater images. The downside is that these methods often lead to over-enhancement or the appearance of unnatural colors due to the selective absorption of light. Data-driven methods use neural networks to learn and adaptively enhance images. However, in underwater environments, it is difficult to obtain reference images corresponding to degraded images, which is crucial for training neural networks, so these data-driven methods are constrained by the scarcity and low quality of various underwater image datasets.

[0006] Therefore, how to design a more effective method to improve the quality of underwater images and have outstanding effects on solving the problems of image detail and contrast restoration has become a technical problem that needs to be solved urgently. Summary of the invention

[0007] In view of this, in order to solve the problems of image blur, color distortion and contrast reduction in underwater images, the present invention proposes a novel underwater image processing method for brightness and color correction. In order to achieve the above purpose, the present invention provides the following technical solutions:

[0008] A novel underwater image processing method for brightness and color correction comprises the following steps:

[0009] S1, inputting an underwater image and extracting image features, and calculating a base layer of the underwater image based on the image features;

[0010] S2, calculating a detail layer of the underwater image based on the base layer of the underwater image, and restoring the brightness of the detail layer to obtain a brightness correction map;

[0011] S3, estimating a maximum attenuation map of the input underwater image, and enhancing details and textures of the brightness correction map based on the maximum attenuation map to obtain a color correction map;

[0012] S4, calculating the gradient distribution map and pixel stretching map of the V channel in the HSV color model corresponding to the color correction map;

[0013] S5. Fusing the gradient distribution map and the pixel stretching map to enhance the global contrast of the color correction map.

[0014] Furthermore, the step of extracting the image features of the underwater image comprises:

[0015] ,

[0016] ,

[0017] in, Used to represent image features, represents an input underwater image; , and Represents underwater images , and The maximum pixel value of the channel, represents the number of pixels of the underwater image, and They represent the Laplace filtering result of the underwater image and the pixel mean of the Laplace filtering result respectively.

[0018] The image features are used to separate the base layer and detail layer of the underwater image. The specific calculation formula of the base layer of the underwater image is:

[0019] ,

[0020] in, The base layer representing the underwater image, represents a Gaussian filter, Represents image features, Represents a convolution operation.

[0021] Furthermore, the detail layer of the underwater image is calculated based on the base layer of the underwater image, and the brightness of the detail layer is restored to obtain a brightness correction map, and the specific steps are as follows:

[0022] Calculate the detail layer of the underwater image. The specific formula is:

[0023] ,

[0024] in, represents the detail layer of the underwater image; the brightness of the detail layer is restored to obtain a brightness correction map, and the specific formula is:

[0025] ,

[0026] in, Indicates the brightness correction map.

[0027] Further, the maximum attenuation map of the underwater image is estimated, and the details and texture of the brightness correction map are enhanced based on the maximum attenuation map to obtain a color correction map, and the specific steps are:

[0028] The specific calculation formula of the maximum attenuation diagram is:

[0029] ,

[0030] in, represents the maximum attenuation graph, represents the maximum value function, , and Represents the underwater image , and Channel; the details and texture of the brightness correction image are enhanced to obtain the color correction image. The specific calculation formula is:

[0031] ,

[0032] in, represents the color correction map, is the input underwater image.

[0033] Furthermore, the gradient distribution map and pixel stretching map of the V channel in the HSV color model corresponding to the color correction map are calculated, and the specific steps are as follows:

[0034] Calculate the gradient distribution map of the V channel, the specific formula is:

[0035] ,

[0036] in, Represents the gradient distribution diagram of the V channel, represents the V channel of the color correction map, represents the Sobel operator;

[0037] The specific calculation formula of the pixel stretching map of the V channel is:

[0038] ,

[0039] in, Represents the pixel stretching map of the V channel, Represents the histogram equalization function.

[0040] Furthermore, the pixel stretching distribution map and the gradient distribution map are fused to enhance the global contrast of the color correction map, and the specific steps are as follows:

[0041] The specific formula for fusing the pixel stretching distribution map and the gradient distribution map is:

[0042] ,

[0043] in Represents the V channel of the enhanced color correction map.

[0044] Beneficial effects: The present invention proposes a novel underwater image processing method for brightness and color correction, aiming to solve the problems of underwater image blur, color distortion and reduced contrast caused by factors such as water turbidity, light scattering and absorption.

[0045] By separating the base layer of underwater images, the color veil covering the image surface can be effectively removed, thereby eliminating color deviation and obtaining a more natural color representation. In addition, the fusion strategy combined with the maximum attenuation map can enhance the details and texture of underwater images and improve the overall contrast. The underwater images with improved quality are more visually attractive and conducive to the development of underwater scientific research. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 A schematic flow chart of a novel underwater image processing method for brightness and color correction implemented in the present invention;

[0047] Figure 2 is the input underwater image;

[0048] Figure 3 for Figure 2 The result after enhancement by a novel underwater image processing method for brightness and color correction;

[0049] Figure 4 is the base layer of the input underwater image;

[0050] Figure 5 is the detail layer of the input underwater image;

[0051] Figure 6 is the brightness correction map of the input underwater image;

[0052] Figure 7 is the color correction map of the input underwater image. DETAILED DESCRIPTION

[0053] In order to have a further understanding and recognition of the structural features and the effects achieved by the present invention, a preferred embodiment and accompanying drawings are used for detailed description as follows:

[0054] like Figure 1 As shown, a novel underwater image processing method for brightness and color correction according to the present invention comprises the following steps:

[0055] The first step is to input an underwater image and extract image features, and calculate the base layer of the underwater image based on the image features. The steps are as follows:

[0056] like Figure 2 The underwater image shown has the problems of blur, color distortion and reduced contrast. It is used as the input of this example. First, the image features of the input underwater image are extracted. The specific formula is:

[0057] ,

[0058] ,

[0059] in, Used to represent image features, represents an input underwater image; , and Represents underwater images , and The maximum pixel value of the channel, represents the number of pixels of the underwater image, and They represent the Laplacian filter result of the input image and the pixel mean of the Laplacian filter result respectively.

[0060] Image features are used to separate the base layer and detail layer of underwater images. Further, the specific formula of the base layer is:

[0061] ,

[0062] in, The base layer representing the underwater image, represents a Gaussian filter, Represents image features, Represents a convolution operation.

[0063] Figure 4 The figure shows the base layer of the underwater image. The base layer includes a color veil covering the surface of the underwater image. The present invention achieves the purpose of removing the color deviation of the underwater image by separating the base layer of the underwater image.

[0064] The second step is to calculate the detail layer of the underwater image based on the base layer of the input underwater image, and restore the brightness of the detail layer to obtain a brightness correction map. The specific steps are as follows:

[0065] Calculate the detail layer of the underwater image. The specific formula is:

[0066] ,

[0067] in, Represents the detail layer of the input underwater image.

[0068] Figure 5 The figure shows the detail layer of the underwater image. The detail layer contains the texture information of the underwater image and eliminates the color veil of the underwater image. However, the brightness of the detail layer is low. Further, the brightness of the detail layer is restored to obtain a brightness correction map. The specific formula is:

[0069] ,

[0070] in, Indicates the brightness correction map.

[0071] Figure 6 The figure shows a brightness correction map of an underwater image. Although the brightness correction map restores the brightness, there is an unnatural red tint. It is also necessary to enhance the details and texture of the brightness correction map and eliminate the unnatural red tint in the brightness correction map.

[0072] The third step is to estimate the maximum attenuation map of the input underwater image, and enhance the details and texture of the brightness correction map based on the maximum attenuation map to obtain the color correction map. The specific steps are as follows:

[0073] The specific calculation formula of the maximum attenuation diagram is:

[0074] ,

[0075] in, represents the maximum attenuation graph, represents the maximum value function, , and Represents the underwater image , and Channel; the details and texture of the brightness correction image are enhanced to obtain the color correction image. The specific calculation formula is:

[0076] ,

[0077] in, represents the color correction map, is the input underwater image.

[0078] Figure 7 It is a color correction map, which restores the brightness and color of underwater images.

[0079] The fourth step is to calculate the gradient distribution map and pixel stretching map of the V channel in the HSV color model corresponding to the color correction image. In order to further enhance the overall contrast of the color correction image, the present invention improves the overall contrast of the color correction image by stretching the V channel of the HSV color model.

[0080] Further, the gradient distribution map of the V channel is calculated, and the specific formula is:

[0081] ,

[0082] in, Represents the gradient distribution diagram of the V channel, represents the V channel of the color correction map, Represents the Sobel operator.

[0083] The specific calculation formula of the pixel stretching map of the V channel is:

[0084] ,

[0085] in, Represents the pixel stretching map of the V channel, Represents the histogram equalization function.

[0086] The fifth step is to fuse the pixel stretching distribution map with the gradient distribution map to enhance the global contrast of the color correction map. The specific steps are as follows:

[0087] The specific formula for fusing the pixel stretch distribution map and the gradient distribution map is:

[0088] ,

[0089] in Represents the V channel of the enhanced color correction map.

[0090] Different weights are assigned to different areas of the underwater image through the gradient distribution map, and the stretch distribution map is further fused to increase the dynamic range of the pixels of the color correction image, thereby enhancing the overall contrast of the color correction image.

[0091] like Figure 3 As shown in the figure, the underwater images processed by this method have more natural colors, richer details and better contrast.

[0092] The above is only one of the implementation examples of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A novel underwater image processing method for brightness and color correction, comprising the following steps: S1, inputting an underwater image and extracting image features, and calculating a base layer of the underwater image based on the image features; S2, calculating a detail layer of the underwater image based on the base layer of the underwater image, and restoring the brightness of the detail layer to obtain a brightness correction map; S3, estimating a maximum attenuation map of the input underwater image, and enhancing details and textures of the brightness correction map based on the maximum attenuation map to obtain a color correction map; S4, calculating the gradient distribution map and pixel stretching map of the V channel in the HSV color model corresponding to the color correction map; S5. Fusing the gradient distribution map and the pixel stretching map to enhance the global contrast of the color correction map.

2. A novel underwater image processing method for brightness and color correction according to claim 1, characterized in that: An underwater image is input and image features are extracted, and a base layer of the underwater image is calculated based on the image features. The specific steps are as follows: Extract the image features of underwater images. The specific calculation formula is: ， ， in, Used to represent image features, represents an input underwater image; , and Represents underwater images , and The maximum pixel value of the channel, represents the number of pixels of the underwater image, and They represent the Laplace filtering result of the underwater image and the pixel mean of the Laplace filtering result respectively; Calculate the base layer of the underwater image. The specific calculation formula is: ， in, The base layer representing the underwater image, represents a Gaussian filter, Represents image features, Represents a convolution operation.

3. A novel underwater image processing method for brightness and color correction according to claim 1, characterized in that: The detail layer of the underwater image is calculated based on the base layer of the underwater image, and the brightness of the detail layer is restored to obtain a brightness correction map, and the specific steps are: Calculate the detail layer of the underwater image. The specific formula is: ， in, represents the detail layer of the underwater image; the brightness of the detail layer is restored to obtain a brightness correction map, and the specific formula is: ， in, Indicates the brightness correction map.

4. A novel underwater image processing method for brightness and color correction according to claim 1, characterized in that: The maximum attenuation map of the underwater image is estimated, and the details and texture of the brightness correction map are enhanced based on the maximum attenuation map to obtain a color correction map. The specific steps are: The specific calculation formula of the maximum attenuation map is: ， in, represents the maximum attenuation graph, represents the maximum value function, , and Represents the underwater image , and Channel; the details and texture of the brightness correction image are enhanced to obtain the color correction image. The specific calculation formula is: ， in, represents the color correction map, is the input underwater image.

5. A novel underwater image processing method for brightness and color correction according to claim 1, characterized in that: The specific steps of calculating the gradient distribution map and pixel stretching map of the V channel in the HSV color model corresponding to the color correction map are as follows: Calculate the gradient distribution map of the V channel, the specific formula is: ， in, Represents the gradient distribution diagram of the V channel, represents the V channel of the color correction map, represents the Sobel operator; the specific calculation formula of the pixel stretching map of the V channel is: ， in, Represents the pixel stretching map of the V channel, Represents the histogram equalization function.

6. A novel underwater image processing method for brightness and color correction according to claim 1, characterized in that: The gradient distribution map and the pixel stretching map are integrated to enhance the global contrast of the color correction map, and the specific steps are as follows: The specific formula for fusing the gradient distribution map and the pixel stretching map is: ， in Represents the V channel of the enhanced color correction map.

Images