A New Underwater Image Processing Method for Brightness and Color Correction

Abstract

The present invention proposes a novel underwater image processing method for brightness and color correction, belonging to the technical field of image processing. The method includes: inputting an underwater image and extracting its features, calculating the base layer and the detail layer of the image according to its features. By restoring the brightness of the detail layer, a brightness correction map is obtained. At the same time, estimating the maximum attenuation map of the underwater image, and using the maximum attenuation map to enhance the details and textures of the brightness correction map, thereby obtaining a color correction map. Then, calculating the gradient distribution map and the pixel stretching distribution map of the V channel in the HSV color model of the color correction map. Finally, fusing these two distribution maps to enhance the global contrast of the color correction map. This method significantly improves the visual quality of underwater images, making the images clearer and more realistic, and has broad application prospects.

Description

Technical Field

[0001] The present invention belongs to the technical field of image processing, and specifically relates to a new underwater image processing method for brightness and color correction. Background Art

[0002] Underwater image enhancement has been a research direction that has received extensive attention in the fields of computer vision and image processing in recent years. With the continuous progress of technology, especially against the background of the improvement of the performance of image sensors and processors, the acquisition of underwater images has become increasingly convenient.

[0003] In the field of marine scientific research, scientists need to accurately observe and record the behaviors, species, and distributions of marine organisms to deeply understand the complexity of the marine ecosystem. Through underwater image enhancement technology, researchers can improve the quality of marine organism images, thereby observing species characteristics and behaviors more clearly. In underwater archaeology and exploration activities, underwater image enhancement also plays an indispensable role. When divers explore sunken ship remains or ancient underwater sites, clear images can help them accurately identify and record various items and structures. In addition, with the increasing global awareness of marine resource management and protection, accurately monitoring fish and their habitats is particularly crucial for sustainable development. Generally speaking, underwater image enhancement technology has wide 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 makes the quality of the acquired images often lower than expected. Affected by many factors such as water turbidity, light scattering, and absorption, the images are blurred, the colors are distorted, and the contrast is reduced. To address these challenges, various underwater image enhancement methods have been proposed, 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 and use assumptions and prior knowledge to invert the model to restore clear images. However, these methods are highly sensitive to their basic assumptions. Model-free methods improve the visual quality of underwater images by adjusting the distribution of pixel values. For example, for instance, researchers have proposed a color and white balance correction method based on the gray world principle and a nonlinear color mapping function, which has improved the color and white balance distortion problems in underwater images. The drawback is that due to the selective absorption of light, these methods often result in over-enhancement or unnatural colors. Data-driven methods use neural networks to learn and adaptively enhance images. However, in the underwater environment, it is difficult to obtain a reference image corresponding to the degraded image, which is crucial for training neural networks. Therefore, these data-driven methods are restricted 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 has outstanding effects on solving the problems of image details and contrast restoration has become a technical problem that needs to be solved urgently. Summary of the Invention

[0007] In view of this, aiming at the problems of image blurring, color distortion and contrast reduction in underwater images, the present invention proposes a new underwater image processing method for brightness and color correction. To achieve the above object, the present invention provides the following technical solutions:

[0008] A new underwater image processing method for brightness and color correction, comprising the following steps:

[0009] S1. Input an underwater image and extract image features, and calculate the base layer of the underwater image based on the image features;

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

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

[0012] S4. Calculate 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. Fuse the gradient distribution map and the pixel stretching map to enhance the global contrast of the color correction map.

[0014] Further, the specific steps of extracting the image features of the underwater image are:

[0015] ,

[0016] ,

[0017] wherein, is used to represent the image features, represents the input underwater image; , and respectively represent the maximum pixel values of the underwater image , and channels, represents the number of pixels of the underwater image, and respectively represent the Laplacian filtering result and the pixel mean value of the Laplacian filtering result of the underwater image.

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

[0019] ,

[0020] where, represents the base layer of the underwater image, represents the Gaussian filter, represents the image features, represents the convolution operation.

[0021] Furthermore, based on the base layer of the underwater image, calculate the detail layer of the underwater image, and restore the brightness of the detail layer to obtain a brightness correction map. The specific steps are as follows:

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

[0023] ,

[0024] where, represents the detail layer of the underwater image; the formula for restoring the brightness of the detail layer to obtain a brightness correction map is:

[0025] ,

[0026] where, represents the brightness correction map.

[0027] Furthermore, estimate the maximum attenuation map of the underwater image, and enhance the details and textures of the brightness correction map based on the maximum attenuation map to obtain a color correction map. The specific steps are as follows:

[0028] The specific calculation formula for the maximum attenuation map is:

[0029] ,

[0030] where, represents the maximum attenuation map, represents the maximum value function, , and respectively represent the , and channels of the underwater image; the formula for enhancing the details and textures of the brightness correction map to obtain a color correction map is:

[0031] ,

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

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

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

[0035] ,

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

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

[0038] ,

[0039] where, represents the pixel stretching map of the V channel, represents the histogram equalization function.

[0040] Furthermore, fuse the pixel stretching distribution map and the gradient distribution map to enhance the global contrast of the color correction map. 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] where represents the V channel of the enhanced color correction map.

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

[0045] By separating the base layer of the underwater image, it can effectively remove the color veil covering the image surface, thereby eliminating color deviation and obtaining a more natural color performance. In addition, combining the fusion strategy of the maximum attenuation map can enhance the details and textures of the underwater image and improve the overall contrast. The underwater image with improved quality is more visually appealing and helps to carry out underwater scientific research. BRIEF DESCRIPTION OF THE DRAWINGS

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

[0047] Figure 2 is the input underwater image;

[0048] Figure 3 is Figure 2 the result enhanced by the 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 implementation manner

[0053] To have a further understanding and recognition of the structural features and achieved effects of the present invention, the following is a detailed description in conjunction with preferred embodiments and accompanying drawings:

[0054] As Figure 1 shown, a novel underwater image processing method for brightness and color correction according to the present invention includes 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] As Figure 2 shown, the underwater image has problems of blurring, color distortion, and reduced contrast, which is used as the input of this example. First, extract the image features of the input underwater image. The specific formula is:

[0057] ,

[0058] ,

[0059] wherein, is used to represent the image features, represents the input underwater image; , and respectively represent the maximum pixel values of the underwater image , and channels, represents the number of pixels of the underwater image, and respectively represent the Laplacian filtering result of the input image and the pixel mean of the Laplacian filtering result.

[0060] The image feature is used to separate the base layer and the detail layer of the underwater image. Further, the specific formula for the base layer is:

[0061] ,

[0062] where, represents the base layer of the underwater image, represents the Gaussian filter, represents the image feature, represents the convolution operation.

[0063] Figure 4 As shown, it is the base layer of the underwater image, and the base layer contains the color veil covering the surface of the underwater image. The present invention aims to remove 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] where, represents the detail layer of the input underwater image.

[0068] Figure 5 As shown, it is 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, but the brightness of the detail layer is low; further, restore the brightness of the detail layer to obtain a brightness correction map. The specific formula is:

[0069] ,

[0070] where, represents the brightness correction map.

[0071] Figure 6 As shown, it is the brightness correction map of the underwater image. Although the brightness correction map restores the brightness, there is an unnatural red tone, and it is still necessary to enhance the details and texture of the brightness correction map and eliminate the unnatural red tone 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 a color correction map. The specific steps are as follows:

[0073] The specific calculation formula for the maximum attenuation map is as follows:

[0074] ,

[0075] wherein, represents the maximum attenuation map, represents the maximum value function, , and respectively represent the , and channels of the underwater image; Details and textures of the enhanced brightness correction map are obtained to get the color correction map. The specific calculation formula is as follows:

[0076] ,

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

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

[0079] Step 4: Calculate the gradient distribution map and pixel stretching map of the V channel in the HSV color model corresponding to the color correction map. To further enhance the overall contrast of the color correction map, the present invention enhances the overall contrast of the color correction map by stretching the V channel of the HSV color model.

[0080] Furthermore, calculate the gradient distribution map of the V channel. The specific formula is as follows:

[0081] ,

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

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

[0084] ,

[0085] wherein, represents the pixel stretching map of the V channel, represents the histogram equalization function.

[0086] Step 5: Fuse the pixel stretching distribution map and 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 stretching distribution map and the gradient distribution map is as follows:

[0088] ,

[0089] where represents the V channel of the enhanced color correction map.

[0090] By assigning different weights to different regions of the underwater image through the gradient distribution map, further fusing the stretching distribution map, and increasing the dynamic range of the pixels of the color correction map, the overall contrast of the color correction map is enhanced.

[0091] As Figure 3 shown, the underwater image processed by this method has 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 replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within 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. Input the underwater image and extract the image features, and calculate the base layer of the underwater image based on the image features ; S2. Calculate the detail layer of the underwater image based on the base layer of the underwater image, and restore the brightness of the detail layer to obtain a brightness correction map. The specific steps are as follows: Calculate the detail layer of the underwater image. The specific formula is: ， Among them, represents the detail layer of the underwater image; to restore the brightness of the detail layer and obtain a brightness correction map, the specific formula is: ， Among them, represents the brightness correction diagram; S3. Estimate the maximum attenuation map of the input underwater image, and enhance the details and textures of the brightness correction map based on the maximum attenuation map to obtain a color correction map. The specific steps are as follows: The specific calculation formula of the maximum attenuation map is: ， Among them, represents the maximum attenuation graph, represents the maximum value function, , and respectively represent the , and channels of the underwater image; The details and textures of the enhanced brightness correction graph are obtained to obtain the color correction graph, and the specific calculation formula is: ， Among them, represents the color correction diagram, is the input underwater image; S4. Calculate the gradient distribution map and pixel stretching map of the V channel in the HSV color model corresponding to the color correction map; S5. Fuse the gradient distribution map and the pixel stretching map to enhance the global contrast of the color correction map.

2. The novel underwater image processing method for brightness and color correction according to claim 1, characterized in that, Input the underwater image and extract image features, and calculate the base layer of the underwater image based on the image features. The specific steps are as follows: Extract the image features of the underwater image. The specific calculation formula is: ， ， Among them, is used to represent image features, represents the input underwater image; , and respectively represent the maximum pixel values of the underwater image 、 and channels, represents the number of pixels of the underwater image, and respectively represent the Laplacian filtering result of the underwater image and the pixel mean of the Laplacian filtering result; Calculate the base layer of the underwater image. The specific calculation formula is: ， Among them, represents the base layer of the underwater image, represents the Gaussian filter, represents the image feature, represents the convolution operation.

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

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

Images