Ocean detection image processing method and device

Abstract

The invention relates to the technical field of image processing, and discloses an ocean exploration image processing method and device, and the method comprises the following steps: S101, generating a red channel histogram; step S102, determining a gap center gray scale according to the red channel histogram, and converting the gap center gray scale into a gap center wavelength; s103, calculating the effective attenuation coefficient of the red light according to the central wavelength of the notch; s104, a red channel compensation gain value is obtained through calculation according to the red light effective attenuation coefficient; and S105, obtaining a corrected ocean detection image by using the red channel compensation gain value. According to the invention, by constructing a spectrum drift sensing mechanism based on double-peak notch characteristics in an image red channel, self-calibration of response change of the optical filter in a deep-sea high-pressure environment is realized, and image color cast and target distortion caused by blue shift of the optical filter are effectively avoided, so that high-quality image data are provided for subsequent ocean tasks.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image processing, more particularly, it relates to a marine exploration image processing method and device. BACKGROUND

[0002] In the marine image exploration technology, an image sensor carrying a red, green and blue filter is widely used for image data acquisition. However, in the deep-sea high-pressure environment, the optical structure of the filter will be significantly affected, especially the multi-layer interference filter, the film layer thickness will be micron-level compressed under the pressure of hundreds to thousands of bars, resulting in a blue shift of the central cut-off wavelength. Although the spectrum drift is subtle, due to the extremely steep cut-off characteristics of the filter, it can cause the red channel image brightness to decrease, the color to shift, and the key target spectral features to be lost.

[0003] In addition, the red band is easily absorbed by water, and the image relies more on the red channel, and the filter blue shift further inhibits the red light transmittance, further reducing the image quality. At the same time, this blue shift cannot be perceived and corrected by automatic exposure, white balance or gamma correction, because its essence is the change of the physical properties of the filter, not the change of the environmental brightness. And because the filter blue shift is nonlinear at different depths and camera poses, traditional methods based on sensor parameters or depth sensor static compensation are difficult to adapt to complex operating scenarios.

[0004] Therefore, there is an urgent need for a method that can perceive the filter spectrum drift based on the image itself and perform adaptive correction in real time. SUMMARY

[0005] The present application provides a marine exploration image processing method and device to solve the technical problems in the background art.

[0006] The present application provides a marine exploration image processing method, comprising the following steps: Step S101, extracting red channel pixels from the collected original marine exploration image to generate a red channel histogram; Step S102, determining the notch center gray level by detecting the valley position between the main peak and the secondary peak in the red channel histogram, and converting the notch center gray level to the notch center wavelength using the first calibration coefficient and the second calibration coefficient; Step S103, calculating the difference between the preset illumination light source center wavelength and the notch center wavelength to obtain the wavelength drift, and mapping the wavelength drift to the red light effective attenuation coefficient using the third calibration coefficient and the fourth calibration coefficient; Step S104, calculating the red channel compensation gain value by combining the wavelength drift and the red light effective attenuation coefficient through the filter static transmittance curve; Step S105, multiply the red channel pixel by the red channel compensation gain value to obtain the corrected red channel pixel, and combine the original green channel pixel and the blue channel pixel, and output the corrected ocean detection image.

[0007] Further, the pixels of the original ocean detection image are arranged in a red, green, green, and blue manner in a 2x2 unit, the red channel pixels are extracted by traversing the original ocean detection image according to the rule that the row and column are both 0, and an array with a length corresponding to the pixel gray scale value and with all element values being 0 is initialized, and then the element values at the index positions corresponding to the pixel gray scale values of each pixel are added by one by traversing the red channel pixels, and the occurrence frequency of the red channel pixels of different gray scale values is recorded as the red channel histogram.

[0008] Further, the gray scale value corresponding to the minimum frequency valley between the maximum frequency peak and the second maximum frequency peak in the red channel histogram is determined as the notch center gray scale, the first calibration coefficient is multiplied by the notch center gray scale, and then the second calibration coefficient is added to obtain the notch center wavelength.

[0009] Further, the first calibration coefficient and the second calibration coefficient are obtained by establishing a linear fitting relationship between the gray scale value and the actual wavelength by shooting monochromatic light of a known wavelength in a laboratory environment.

[0010] Further, step S103 specifically includes the following steps: Step S201, statistically analyzing the wavelength drift of consecutive frames in a preset time window with a length of M to obtain a steady-state drift mean and a steady-state drift standard deviation; Wherein the wavelength drift of each frame is equal to the difference between the preset illumination light source center wavelength and the notch center wavelength, wherein the length M of the preset time window and the preset illumination light source center wavelength are both self-defined parameters; Step S202, calculating the spectral attenuation coefficient by using the third calibration coefficient and the fourth calibration coefficient in combination with the steady-state drift mean and the steady-state drift standard deviation; Step S203, converting the camera real-time depth reading into a depth attenuation coefficient by a polynomial fitting function using the water body optical property parameter; Step S204, calculating the difference between the spectral attenuation coefficient and the depth attenuation coefficient, and calculating the dynamic weight according to the difference; Step S205, fusing the spectral attenuation coefficient and the depth attenuation coefficient by using the dynamic weight to calculate the red light effective attenuation coefficient.

[0011] Further, by placing the camera in a hyperbaric chamber, taking images of a standard reflector plate at different known pressures, obtaining the steady-state drift mean and steady-state drift standard deviation according to step S201, simultaneously measuring the red light effective attenuation coefficient at the corresponding pressure using a water optical instrument, and finally fitting the third calibration coefficient and the fourth calibration coefficient by least squares method.

[0012] Further, the camera filter is placed in the light path of a spectrometer, and the ratio of the light intensity transmitted through the filter to the incident light intensity is measured at each wavelength in a predetermined wavelength range to obtain a filter static transmittance curve.

[0013] Further, the first determination result is obtained by substituting the center wavelength of the preset illumination light source into the filter static transmittance curve, and the second determination result is obtained by substituting the sum of the center wavelength of the preset illumination light source and the wavelength drift into the filter static transmittance curve, and the transmittance ratio is calculated by the ratio of the first determination result to the second determination result.

[0014] Further, the red channel compensation gain value is obtained by multiplying the transmittance ratio by the water column absorption compensation factor, wherein the water column absorption compensation factor is obtained by taking the natural constant as the base number and performing exponential operation on the product of the red light effective attenuation coefficient and the spectral depth, wherein the spectral depth is the ratio of the real-time depth reading of the camera to the cosine value of the angle between the camera and the vertical direction.

[0015] The present application provides a kind of marine exploration image processing device, comprising: The red channel histogram generation module is used to extract red channel pixels from the acquired original marine exploration image and generate a red channel histogram. The notch center wavelength conversion module is used to determine the notch center gray level by detecting the valley position between the main peak and the secondary peak in the red channel histogram, and convert the notch center gray level to the notch center wavelength using the first calibration coefficient and the second calibration coefficient. The red light effective attenuation coefficient mapping module is used to calculate the wavelength drift by the difference between the center wavelength of the preset illumination light source and the notch center wavelength, and map the wavelength drift to the red light effective attenuation coefficient using the third calibration coefficient and the fourth calibration coefficient. The red channel compensation gain value generation module is used to calculate the red channel compensation gain value by the filter static transmittance curve, combined with the wavelength drift and the red light effective attenuation coefficient. The marine exploration image correction module is used to multiply the red channel compensation gain value by the red channel pixel to obtain the corrected red channel pixel, and combine it with the original green channel pixel and blue channel pixel to output the corrected marine exploration image.

[0016] The beneficial effects of the present application are as follows: the present application realizes self-calibration and self-compensation of the response change of the filter under the deep-sea high-pressure environment by constructing a spectral drift perception mechanism based on the bimodal notch feature in the red channel of the image, compared with the traditional method which depends on external pressure sensors or static parameter setting, the present application does not need to increase any hardware, but can dynamically estimate the red light effective attenuation coefficient through the observable features of the image itself, and then accurately adjust the red channel gain, significantly improve the image color stability and spectral consistency, and under the condition of complex water body and extreme depth, the present application can effectively avoid the image color cast and target distortion caused by the filter blue shift, ensure that the image basis for subsequent mineral identification, biological discrimination and other key tasks is real and reliable, and has good practicability and engineering adaptability. BRIEF DESCRIPTION OF DRAWINGS

[0017] Fig. 1 is a flow chart of a marine exploration image processing method of the present application; Fig. 2 is a flow chart of calculating the red light effective attenuation coefficient of the present application; Fig. 3 is a schematic diagram of a marine exploration image processing device of the present application.

[0018] In the figure: the red channel histogram generation module 301, the notch center wavelength conversion module 302, the red light effective attenuation coefficient mapping module 303, the red channel compensation gain value generation module 304, and the marine exploration image correction module 305. DETAILED DESCRIPTION

[0019] The subject matter described herein will now be discussed with reference to example implementations. It should be understood that the discussion of these implementations is merely meant to provide a better understanding of the subject matter described herein and can be changed in function and arrangement without departing from the scope of the present disclosure. Various processes or components can be omitted, substituted, or added according to desired implementations. Additionally, features described in relation to some examples can be combined in other examples.

[0020] It should be noted that the technical terms or scientific terms used in one or more embodiments of the present application should be understood as the usual meaning understood by those skilled in the art unless otherwise defined. The terms "first", "second" and the like used in one or more embodiments of the present application do not represent any order, number or importance, but are only used to distinguish different components. The terms "include" or "contain" and the like mean that the elements or objects appearing before the terms cover the elements or objects listed after the terms and their equivalents, and do not exclude other elements or objects. The terms "connected" or "connected" and the like are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to represent relative positional relationships, and when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0021] As shown in Figs. 1-3 A method for processing marine exploration images, comprising the following steps: Step S101, extracting red channel pixels from the acquired original marine exploration images to generate a red channel histogram; Step S102, determining the notch center gray level by detecting the valley value position between the main peak and the secondary peak in the red channel histogram, and converting the notch center gray level into a notch center wavelength using a first calibration coefficient and a second calibration coefficient; Step S103, calculating the difference between the preset illumination light source center wavelength and the notch center wavelength to obtain the wavelength drift, and mapping the wavelength drift to a red light effective attenuation coefficient using a third calibration coefficient and a fourth calibration coefficient; Step S104, calculating the red channel compensation gain value by combining the wavelength drift and the red light effective attenuation coefficient through the filter static transmittance curve; Step S105, multiplying the red channel pixels by the red channel compensation gain value to obtain corrected red channel pixels, and combining them with the original green channel pixels and blue channel pixels to output the corrected marine exploration image.

[0022] It should be noted that the format of the original marine exploration image is Bayer-RAW. The pixels of the image in this format are arranged in a red, green, green, blue mosaic pattern in a 2x2 unit, and the light information is captured in a Bayer filter arrangement. The Bayer filter is a color filter array overlaid on the image sensor, used to capture color images. This format completely retains the dynamic range and original photoelectric response accuracy captured by the sensor, avoids interference with the spectral characteristics of the red channel by the camera's built-in processing flow, and thus provides undistorted basic data for accurate positioning of the notch center gray level.

[0023] In one embodiment of the present application, during the operation of the camera-equipped deep-sea exploration platform, the camera lens axis is kept at an angle of 5-15 degrees with the surface of the exploration target, while ensuring that the light source is coaxial with the lens and the angle between the illumination direction and the lens axis is no more than 3 degrees, which can minimize the interference of the blue shift of the filter angle incident on the red channel spectrum caused by excessively large incident angles, i.e., the wavelength shift reflects the degree of blue shift, ensuring that the spectral shift caused by deep-sea pressure becomes the dominant factor, thereby improving the stability of the mapping relationship between the wavelength shift and the deep-sea pressure and the accuracy of the subsequent calculation of the red channel compensation gain value.

[0024] In one embodiment of the present application, the pixels of the original marine exploration image are arranged in the order of red, green, green, and blue in a 2x2 unit, the red channel pixels are extracted according to the rule that the row and column are both 0, and an array with a length corresponding to the pixel gray value and all elements being 0 is initialized, then the elements at the index positions corresponding to the pixel gray values of each red channel pixel are incremented by 1, and the occurrence frequency of the red channel pixels of different gray values is recorded as the red channel histogram.

[0025] It should be noted that the pixel gray value is a self-defined parameter, for example, the pixel gray value range of the original marine exploration image is generally [0, 4095], i.e., the corresponding array length is 4096, or [0, 65535], i.e., the corresponding array length is 65536; for example, the red channel pixel value is 30 and the occurrence frequency is 50, then the element value at the array index of 30 is 50.

[0026] In one embodiment of the present application, the gray values of the maximum frequency peak and the second maximum frequency peak are determined in the red channel histogram, then the gray value corresponding to the minimum frequency valley between the two is determined as the notch center gray value, the first calibration coefficient is multiplied by the notch center gray value, then added to the second calibration coefficient to obtain the notch center wavelength.

[0027] It should be noted that the notch center wavelength is calculated as follows: wherein represents the notch center gray value, and represent the first calibration coefficient and the second calibration coefficient, respectively.

[0028] In one embodiment of the present application, the first calibration coefficient and the second calibration coefficient are obtained by establishing a linear fitting relationship between the gray value and the actual wavelength through shooting monochromatic light of known wavelength in a laboratory environment, for example, through calibration of a 620nm standard light source, the obtained first calibration coefficient is 0.015nm and the second calibration coefficient is 562nm, the specific values vary with the sensor model and are not described here.

[0029] For example, the main peak with the highest frequency in the interval of 3000 to 3500 of the gray scale value corresponds to the normal red response greater than 630 nm, the secondary peak with the second highest frequency in the interval of 2500 to 2800 of the gray scale value corresponds to the scattering signal of 590 to 610 nm, and the frequency in the interval of 2800 to 3000 between the two peaks has the lowest valley value at the gray scale value of 2900, so that the notch center gray scale is determined as 2900, and then the notch center wavelength is calculated as 605.5 nm according to the above calibration coefficient.

[0030] In an embodiment of the present application, as shown in Fig. 2 Step S103 specifically includes the following steps: Step S201, performing statistical analysis on the wavelength drift of the continuous frames in a preset time window with a length of M to obtain a steady-state drift mean value and a steady-state drift standard deviation; Wherein the wavelength drift of each frame is equal to the difference between the preset illumination light source center wavelength and the notch center wavelength, wherein the length M of the preset time window and the preset illumination light source center wavelength are self-defined parameters, and preferably, the length M of the preset time window is set to 10, and the preset illumination light source center wavelength is set according to the hardware parameters, for example, set to 620 nm; Step S202, using the third calibration coefficient and the fourth calibration coefficient to calculate the spectral attenuation coefficient in combination with the steady-state drift mean value and the steady-state drift standard deviation; The calculation formula of the spectral attenuation coefficient is as follows: Wherein represents the steady-state drift mean value, represents the steady-state drift standard deviation, and respectively represent the third calibration coefficient and the fourth calibration coefficient; Step S203, using the water body optical property parameters to convert the camera real-time depth reading into a depth attenuation coefficient through a polynomial fitting function; The calculation formula of the depth attenuation coefficient is as follows: Wherein represents the camera real-time depth reading, and a, b and c respectively represent the first water body optical property parameter, the second water body optical property parameter and the third water body optical property parameter; Step S204, calculating the difference between the spectral attenuation coefficient and the depth attenuation coefficient, and calculating the dynamic weight according to the difference; The calculation formula of the dynamic weight is as follows: Wherein represents a difference between the spectral attenuation coefficient and the depth attenuation coefficient, represents a sensitivity, and exp represents an exponential function with a natural constant as a base number; In step S205, the spectral attenuation coefficient and the depth attenuation coefficient are fused by using a dynamic weight to calculate a red effective attenuation coefficient. red effective attenuation coefficient The calculation formula is as follows: .

[0031] In an embodiment of the present application, the camera is placed in a hyperbaric chamber, and images of a standard reflector plate are taken at different known pressures (for example, 100 bar, 200 bar, and 300 bar, etc.). The steady-state drift mean value and the steady-state drift standard deviation are obtained according to step S201, and the red effective attenuation coefficient under the corresponding pressure is measured by using a water optical instrument at the same time. Finally, the third calibration coefficient and the fourth calibration coefficient are obtained by fitting and backstepping by using the least square method.

[0032] It should be noted that the red effective attenuation coefficient under the corresponding pressure can also be measured by using the water optical instrument at different depths (for example, 100 m, 200 m, and 500 m, etc.) of the target sea area, and the water optical characteristic parameters are obtained by fitting and backstepping by using the least square method.

[0033] It should be noted that the default value of the sensitivity is 0.5. Other parameters can be fixed first, and then the size of the sensitivity can be manually adjusted. The subjective evaluation of the human eye or the objective image quality index (for example, SSIM) is used as a standard to realize the effect of self-adaptive adjustment of the sensitivity according to the real-time depth reading of the camera, so as to ensure that the quality of the corrected marine exploration image is optimal.

[0034] In an embodiment of the present application, the camera filter is placed in the light path of a spectrometer, and the ratio of the light intensity transmitted through the filter to the incident light intensity is measured at each wavelength in a preset wavelength range (usually covering the red channel working wavelength band of the camera filter, for example, 600 nm to 700 nm), to obtain a filter static transmittance curve.

[0035] In an embodiment of the present application, the preset illumination light source center wavelength is substituted into the filter static transmittance curve to obtain a first measurement result, and the sum of the preset illumination light source center wavelength and the wavelength drift is substituted into the filter static transmittance curve to obtain a second measurement result. The ratio of the first measurement result to the second measurement result is calculated to obtain a transmittance ratio, which quantifies the influence of the spectral response change of the camera filter caused by the deep-sea pressure on the red light signal.

[0036] In one embodiment of the present application, the red channel compensation gain value is obtained by multiplying the transmittance ratio value and a water column absorption compensation factor, wherein the water column absorption compensation factor is obtained by taking the product of the red light effective attenuation coefficient and the spectral depth as the base of the natural constant and performing exponential operation, and wherein the spectral depth is the ratio of the camera real-time depth reading and the cosine value of the angle between the camera and the vertical direction.

[0037] It should be noted that the red channel compensation gain value is used to compensate the intensity attenuation of red light in the water body due to absorption and scattering, that is, to quantify and correct the influence of the water body environment at different depths on the red light signal in an exponential form, and in addition, to convert the camera oblique optical path into a vertical optical path to ensure that the attenuation calculation is consistent with the actual water column length through which the light passes, thereby improving the accuracy of the red channel compensation gain value calculation.

[0038] It should be noted that the red channel compensation gain value is calculated according to the following formula: , wherein represents a first measurement result obtained by substituting the center wavelength of the preset illumination light source into the filter static transmittance curve, represents a second measurement result obtained by substituting the sum of the center wavelength of the preset illumination light source and the wavelength shift into the filter static transmittance curve, represents the red light effective attenuation coefficient, represents the camera real-time depth reading, represents the angle between the camera and the vertical direction.

[0039] It should be noted that the original red channel pixels are subjected to pixel-by-pixel gain adjustment to obtain corrected red channel pixels, so as to accurately compensate the red channel spectral drift and water body attenuation caused by the deep sea environment, and then the corrected red channel pixels are fused with the original green channel pixels and blue channel pixels without change to reconstruct the image data with color balance, thereby completing the red channel self-calibration of the marine exploration image.

[0040] In one embodiment of the present application, as shown in Fig. 3 , the present application provides a marine exploration image processing device, comprising: a red channel histogram generation module 301 configured to extract red channel pixels from the collected original marine exploration image and generate a red channel histogram; a notch center wavelength conversion module 302 configured to determine a notch center gray scale by detecting the valley position between the main peak and the secondary peak in the red channel histogram, and convert the notch center gray scale into a notch center wavelength by using a first calibration coefficient and a second calibration coefficient; a red light effective attenuation coefficient mapping module 303, configured to calculate a wavelength drift by calculating a difference between the center wavelength of the preset illumination light source and the center wavelength of the notch, and map the wavelength drift to a red light effective attenuation coefficient by using a third calibration coefficient and a fourth calibration coefficient; a red channel compensation gain value generation module 304, configured to calculate a red channel compensation gain value by using the filter static transmittance curve, the wavelength drift and the red light effective attenuation coefficient; a marine exploration image correction module 305, configured to multiply the red channel pixel by the red channel compensation gain value to obtain a corrected red channel pixel, and combine the corrected red channel pixel with the original green channel pixel and the blue channel pixel to output a corrected marine exploration image.

[0041] It should be noted that the intervals and threshold values are set for ease of comparison, and the size of the threshold value depends on the amount of sample data and the base number set by those skilled in the art for each group of sample data, as long as the proportional relationship between the parameters and the quantized values is not affected. And the above formula is a dimensionless calculation of the value, and the formula is obtained by software simulation of a large amount of data to obtain a formula closest to the actual situation, and the preset parameters in the formula are set by those skilled in the art according to the actual situation.

[0042] The above describes the embodiments of the present embodiment, but the present embodiment is not limited to the specific embodiments described above, and the specific embodiments described above are only illustrative and not limiting, and those skilled in the art can make many forms under the inspiration of the present embodiment, which all belong to the protection of the present embodiment.

Claims

1. A method of processing a marine survey image, characterized by, The method comprises the following steps: Step S101, extracting red channel pixels from the collected original marine detection image to generate a red channel histogram; Step S102, determining the notch center gray level by detecting the valley value position between the main peak and the secondary peak in the red channel histogram, and converting the notch center gray level into a notch center wavelength using a first calibration coefficient and a second calibration coefficient; Step S103, calculating the wavelength drift by calculating the difference between the preset illumination light source center wavelength and the notch center wavelength, and mapping the wavelength drift to a red light effective attenuation coefficient using a third calibration coefficient and a fourth calibration coefficient; Step S104, calculating the red channel compensation gain value by combining the wavelength drift and the red light effective attenuation coefficient through the filter static transmittance curve; Step S105, multiplying the red channel pixels by the red channel compensation gain value to obtain corrected red channel pixels, and combining the corrected red channel pixels with the original green channel pixels and blue channel pixels to output a corrected marine detection image.

2. The method of claim 1, wherein, The pixels of the original marine detection image are arranged in a red, green, green, and blue manner in a 2x2 unit. The red channel pixels are extracted according to the rule that the row pixel position and the column pixel position are both 0, and an array with a length corresponding to the pixel gray level value and element values of 0 is initialized. The element values corresponding to the array index positions of each pixel gray level value are added by 1, and the occurrence frequency of the red channel pixels of different gray level values is recorded as the red channel histogram.

3. The method of claim 1, wherein, First, the gray level values of the maximum frequency peak and the second maximum frequency peak are determined in the red channel histogram, and then the gray level value corresponding to the minimum frequency valley between the two is determined as the notch center gray level. The first calibration coefficient is multiplied by the notch center gray level, and then the second calibration coefficient is added to obtain the notch center wavelength.

4. The method of claim 1, wherein, The first calibration coefficient and the second calibration coefficient are obtained by establishing a linear fitting relationship between the gray level value and the actual wavelength by shooting monochromatic light with a known wavelength in a laboratory environment.

5. The method of claim 1, wherein, Step S103 specifically comprises the following steps: Step S201, statistically analyzing the wavelength drift of consecutive frames within a preset time window with a length of M to obtain a steady-state drift mean and a steady-state drift standard deviation; wherein the wavelength drift of each frame is equal to the difference between the preset illumination light source center wavelength and the notch center wavelength, and wherein the length M of the preset time window and the preset illumination light source center wavelength are both self-defined parameters; Step S202, calculating the spectral attenuation coefficient using the third calibration coefficient and the fourth calibration coefficient in combination with the steady-state drift mean and the steady-state drift standard deviation; Step S203, converting the camera real-time depth reading to a depth attenuation coefficient using a polynomial fitting function based on the water body optical property parameters; Step S204, calculating the difference between the spectral attenuation coefficient and the depth attenuation coefficient, and calculating the dynamic weight based on the difference; Step S205, fusing the spectral attenuation coefficient and the depth attenuation coefficient using the dynamic weight to calculate the red light effective attenuation coefficient.

6. The method of claim 5, wherein, The camera is placed in a pressure tank, and images of a standard reflector plate are taken at different known pressures. The mean and standard deviation of the steady-state drift are obtained according to step S201, and the water optical instrument is used to measure the red light effective attenuation coefficient at the corresponding pressure. Finally, the third and fourth calibration coefficients are obtained by fitting and back-calculating using the least squares method.

7. The method of claim 1, wherein, The camera filter is placed in the light path of a spectrometer, and the ratio of the light intensity transmitted through the filter to the incident light intensity is measured at each wavelength in a predetermined wavelength range to obtain a filter static transmittance curve.

8. The method of claim 1, wherein, The first measurement result is obtained by substituting the center wavelength of the preset illumination light source into the filter static transmittance curve, and the second measurement result is obtained by substituting the sum of the center wavelength of the preset illumination light source and the wavelength drift into the filter static transmittance curve. The transmittance ratio is obtained by calculating the ratio of the first measurement result to the second measurement result.

9. The method of claim 8, wherein, The red channel compensation gain value is obtained by multiplying the transmittance ratio by the water column absorption compensation factor, wherein the water column absorption compensation factor is obtained by taking the natural constant as the base number and performing exponential operation on the product of the red light effective attenuation coefficient and the spectral depth, and the spectral depth is the ratio of the real-time depth reading of the camera to the cosine value of the angle between the camera and the vertical direction.

10. A processing device of a marine exploration image, characterized by, The method for processing an ocean exploration image according to any one of claims 1 to 9 is executed, comprising: a red channel histogram generation module for extracting red channel pixels from the collected original ocean exploration image and generating a red channel histogram; a notch center wavelength conversion module for determining the notch center gray scale by detecting the valley position between the main peak and the secondary peak in the red channel histogram, and converting the notch center gray scale to the notch center wavelength using the first and second calibration coefficients; a red light effective attenuation coefficient mapping module for calculating the difference between the preset illumination light source center wavelength and the notch center wavelength to obtain the wavelength drift, and mapping the wavelength drift to the red light effective attenuation coefficient using the third and fourth calibration coefficients; a red channel compensation gain value generation module for calculating the red channel compensation gain value by combining the wavelength drift and the red light effective attenuation coefficient through the filter static transmittance curve; an ocean exploration image correction module for multiplying the red channel pixels by the red channel compensation gain value to obtain corrected red channel pixels, and combining the corrected red channel pixels with the original green channel pixels and blue channel pixels to output the corrected ocean exploration image.

Images