A skylight polarization navigation system and method suitable for marine environment

Abstract

The present invention discloses a skylight polarization navigation system and method suitable for marine environment, belonging to the field of navigation technology. In the system, a gyro-stabilized platform is fixed on the ground of a ship, a DOFP polarization imager is detachably mounted on the gyro-stabilized platform, a fisheye lens is mounted on the front end of the DOFP polarization imager, and a bandpass filter is placed in the C-mount interface of the DOFP polarization imager, and is located behind the light outlet of the fisheye lens. The system is used to capture real-time skylight polarization pattern images, calculate the degree of polarization DOP value, the angle of polarization AOP value and the DOP gradient value, identify the output neutral point position based on the calculation results, and perform solar meridian detection and extraction on the AOP image converted to the incident light coordinate system; determine the heading angle of the carrier based on the extracted neutral point position and the solar meridian. The present invention realizes real-time, high-precision polarization pattern measurement and heading angle calculation.

Description

Technical Field

[0001] The present invention belongs to the field of navigation technology, and in particular relates to a skylight polarization navigation system and method suitable for marine environments. Background Art

[0002] At present, maritime navigation mainly relies on GPS and inertial navigation. The measurement and research of the polarization pattern of skylight on the sea surface is not sufficient, especially the research on the polarization distribution characteristics of skylight at sea and its feasibility in navigation is not deep enough. There are two main difficulties in polarization navigation at sea. First, the sea surface is not calm. The turbulence and tilt caused by the waves on the hull make it impossible for the polarization sensor to align with the zenith. When the sensor is tilted, the central area of ​​the image will no longer correspond to the zenith area. This causes the selected sky area to change constantly under dynamic conditions, and it is impossible to establish a unified heading reference for the polarized light compass, which greatly affects the accuracy of polarization navigation. Second, the marine environment and climate are changeable. Clouds and sea fog will affect the propagation and polarization characteristics of skylight. In particular, cloud cover or ship building coverage will weaken or change the symmetrical distribution of polarization signals, making navigation unstable.

[0003] Skylight polarization patterns are the polarization distribution of sunlight scattered by the Earth's atmosphere, providing valuable directional information and have been used in a variety of navigation techniques. These patterns are dominated by Rayleigh scattering, forming a stable and measurable polarization distribution, including the solar meridian, neutral point, and symmetrical arrangements of the degree of polarization (DOP) and angle of polarization (AOP). The spatial distribution of these polarization signals varies with the position of the sun, atmospheric conditions, and the latitude of the observer. Skylight polarization patterns have been used by humans and animals for hundreds of years, and studies have shown that many insects, birds, and marine life rely on unique polarization angles and intensities to maintain navigation accuracy.

[0004] Heading information is a basic requirement for long-term autonomous navigation, especially in environments where satellite signals are unavailable or electromagnetic interference is severe. Under these conditions, current inertial navigation systems (INS) accumulate errors, while magnetic compasses are susceptible to interference and cannot meet the accuracy requirements of unmanned platforms. To address these issues, autonomous navigation systems are often combined with auxiliary navigation methods such as celestial navigation, visual navigation, acoustic navigation, and geophysical navigation. Although these systems are able to correct the drift of INS, they are usually costly and susceptible to external interference. Inspired by the natural navigation mechanism of animals, skylight polarization information can be used as a stable, autonomous heading reference that is undisturbed and has no time-dependent error accumulation. Bionic polarization navigation exploits the polarization patterns in the sky to obtain heading information by detecting and calculating polarization signals.

[0005] The design of polarized skylight navigation sensors is inspired by the biological polarization detection mechanism and has developed along two main technical paths: non-imaging sensors and imaging sensors. Non-imaging sensors are relatively simple in structure and collect polarized light intensity in a specific direction through discrete photodetectors. For example, the early prototype developed by Labhart and Meyer combined polarizers, photodetectors, and circuit processing modules to simulate biological compound eyes. These sensors use blue light band filters to detect the direction of polarized skylight. Compared with the ultraviolet spectrum, detection in the blue light band is easier to implement and more effective.

[0006] Imaging sensors use advances in semiconductor technology (such as CMOS and CCD) to capture polarization data for the entire sky. Based on the data acquisition method, these sensors are generally divided into time series sensors, multi-channel sensors, and focal plane division (DOFP) sensors. The calculation of the polarization direction is achieved by extracting the symmetry axis (solar meridian) from the polarization information image. Although non-imaging sensors have the advantages of fast response and high instantaneous accuracy in good weather conditions, their performance is greatly limited in bad weather. Imaging sensors show stronger anti-interference capabilities and can detect polarized skylight from multiple directions, although their accuracy is slightly lower than that of non-imaging sensors. Summary of the invention

[0007] In order to solve the above technical problems, the present invention provides a skylight polarization navigation system and method suitable for marine environment, which reduces the sensor tilt caused by sea surface instability, resulting in polarization mode measurement errors and the influence of local obstruction (such as clouds, hull structure) on polarization signal symmetry and navigation stability.

[0008] To achieve the above object, the present invention adopts the following technical solution:

[0009] A skylight polarization navigation system suitable for marine environment, the system comprises a gyro-stabilized platform, a DOFP polarization imager, a fisheye lens and a bandpass filter, wherein:

[0010] The gyro-stabilized platform is fixed on the hull ground, the DOFP polarization imager is detachably mounted on the gyro-stabilized platform, the fisheye lens is mounted on the front end of the DOFP polarization imager, and the bandpass filter is placed in the C-mount interface of the DOFP polarization imager and is located behind the light outlet of the fisheye lens;

[0011] The DOFP polarization imager includes a monochrome CMOS sensor, which includes a top layer of a microlens array, a middle layer of a polarizing filter array, and a bottom layer of a super pixel array; the microlens array is used to focus light; the middle layer is a polarizing filter array with directions of 0°, 45°, 90° and 135°; the bottom layer of the super pixel array is used to convert light signals into digital signals; wherein each super pixel includes four sub-pixels, and the photodiode corresponding to each sub-pixel is aligned with a polarizing filter in the middle layer of the polarizing filter array.

[0012] On the other hand, the present invention provides a skylight polarization navigation method suitable for marine environment, which is applied to the aforementioned skylight polarization navigation system suitable for marine environment, and comprises the following steps:

[0013] Step 1, using the skylight polarization navigation system suitable for the marine environment to capture real-time skylight polarization mode images, calculate the degree of polarization DOP value, the polarization angle AOP value and the DOP gradient value, identify and isolate the effective area that is not blocked by clouds and hull buildings based on the calculation results, and delineate the area of ​​interest based on the effective area;

[0014] Step 2: convert the AOP image in the region of interest from the observation coordinate system to the navigation coordinate system, and then to the incident light coordinate system, and reconstruct the AOP pattern through denoising;

[0015] Step 3, quantify the AOP image in the observation coordinate system, identify the output neutral point position, and detect and extract the solar meridian of the AOP image converted to the incident light coordinate system;

[0016] Step 4: Determine the heading angle of the carrier based on the extracted neutral point position and the solar meridian.

[0017] The beneficial effects of the present invention are:

[0018] The present invention proposes a method to achieve real-time, high-precision polarization mode measurement and heading angle calculation. The algorithm combines adaptive sky area shielding, coordinate system conversion, and dual feature recognition technology in polarization mode processing. It compensates for the tilt of the hull through coordinate conversion, and combines the neutral point detection and solar meridian extraction methods in the region of interest (ROI) to ensure that high measurement accuracy can be maintained under partial occlusion conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a schematic diagram of the structure of a skylight polarization navigation system suitable for marine environments according to the present invention;

[0020] Figure 2 This is a flow chart of a skylight polarization navigation method suitable for marine environment according to the present invention;

[0021] Figure 3 This is a schematic diagram of the celestial coordinate system;

[0022] Figure 4 This is a schematic diagram for calculating the heading angle;

[0023] Figure 5 The figure is a schematic diagram of the principle of a skylight polarization navigation method suitable for a marine environment according to the present invention. DETAILED DESCRIPTION

[0024] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0025] The present invention aims to solve the two main challenges of polarization navigation in marine environments: one is the sensor tilt caused by unstable sea surface, which leads to polarization mode measurement errors; the other is the impact of local occlusion (such as clouds and hull structure) on the symmetry of polarization signals and navigation stability. To this end, the present invention proposes a skylight polarization navigation system and method suitable for marine environments, which are introduced below.

[0026] On the one hand, the present invention provides a skylight polarization navigation system suitable for marine environment, which is a gyro-stabilized all-sky DOFP polarization imaging system, aiming to provide stable real-time polarization measurement for dynamic marine environment. The system includes a gyro-stabilized platform 1, a DOFP polarization imager 2, a fisheye lens 3 and a bandpass filter 4, and its main structure is as follows: Figure 1 shown.

[0027] The gyro-stabilized platform 1 is fixed on the hull ground, and the DOFP polarization imager 2 is supported on the gyro-stabilized platform 1 by tripods, and is bound by gluing and rolling tape, and is vertically facing the sky. The fisheye lens 3 is installed at the front end of the DOFP polarization imager 2, and the bandpass filter 4 is placed in the C-mount interface of the DOFP polarization imager 2, behind the light outlet of the fisheye lens 3.

[0028] The gyro-stabilized platform 1 integrates an electronic balancing mechanism and a digital control system to ensure image stability in the case of hull motion. The platform has good environmental adaptability, is IP waterproof grade certified, can withstand harsh conditions, and is very suitable for marine deployment. Its installation compensation mechanism effectively reduces the pitch and roll motion caused by waves, and controls the angle deviation within a range of less than 0.5°, which is crucial for accurate polarization measurement.

[0029] At the heart of the imaging setup is an Edmund fisheye lens 3 that captures full-sky images with a 185° field of view. Designed with a 2 / 3" image circle and C-mount design, the lens has a focal length of 1.8 mm and supports applications up to megapixel resolution. The large depth of field design allows stable focus at distances from 100 mm to infinity without the need for adjustment. In addition, a manual aperture lock prevents overexposure and ensures that the aperture setting remains stable under external shocks.

[0030] To selectively capture polarized light from the visible to near-infrared, the system is equipped with a set of manually switchable bandpass filters4, with center wavelengths of 410 nm, 450 nm, 500 nm, 675 nm, and 870 nm. Each filter has a half-width (FWHM) of 10 nm and a diameter of 24 mm, allowing for easy integration into C-mount systems.

[0031] The DOFP polarization imager 2 is equipped with a 5-megapixel monochrome CMOS sensor with microlenses and polarization filters integrated on its surface. Specifically, the sensor adopts a three-layer precision structure design: the top layer is a microlens array for focusing light; the middle layer is a polarization filter array with 0°, 45°, 90° and 135° directions; the bottom layer is a super pixel array for converting light signals into digital signals. Each super pixel 5 consists of four sub-pixels 6, and the photodiode 8 corresponding to each sub-pixel 6 is equipped with a polarization filter 7, so that four polarization states can be captured simultaneously. This unique arrangement allows the system to effectively modulate polarized light and minimize aliasing artifacts in traditional frequency domain modulation methods. The strict alignment of the polarization filter array with the pixel matrix ensures that parallax errors are minimized, thereby generating high-quality polarization images. Through built-in software, the system is able to process data in real time and output key polarization parameters, including the degree of linear polarization (DOLP), the angle of linear polarization (AOLP) and the intensity (I).

[0032] (1)

[0033] (2)

[0034] in,

[0035] (3)

[0036] The formula uses Stokes parameters to describe the characteristics of polarized light. Equivalent to represents the total intensity of light, Equivalent to represents the difference between the horizontal polarization intensity and the vertical polarization intensity, Equivalent to It represents the difference between the polarization components at +45° and -45°. The degree of linear polarization (DOLP) is used to quantify the degree of polarization of light, ranging from 0 to 1, while the angle of polarization (AOLP) represents the angle of the polarization direction of light, ranging from -90° to +90° relative to the horizontal axis.

[0037] The gyro-stabilized all-sky DOFP polarization imaging system supports real-time spectral polarization data acquisition across the entire sky. Compared with the time-segmented measurement method, the system eliminates the asynchronous error introduced by the rotating polarizer and provides polarization data with consistent spatial and temporal characteristics, making it particularly suitable for dynamic scenes. In addition, compared with other polarimeter, the system has a more compact and convenient design, achieving a balance between ease of operation and data accuracy.

[0038] On the other hand, in order to achieve accurate heading estimation in complex environments such as frequent weather changes, partial cloud cover, and dynamic ocean conditions, based on the aforementioned system, the present invention also provides a skylight polarization navigation method suitable for marine environments. This method solves the unique challenges caused by partial cloud cover, hull cover, and wave tilt. Although the gyro-stabilized platform 1 can compensate for most of the hull's tilt, there are still some minor inaccuracies. In addition, since the polarization image covers a larger field of view (FOV), partial hull cover and partial cloud cover in the sky are still sources of interference. If the heading information is extracted directly from the unprocessed polarization pattern, the accuracy and reliability cannot be guaranteed. Therefore, the present invention proposes this method, and the flow chart is as follows: Figure 2 As shown, the proposed method can mitigate these effects and improve the robustness of heading calculation under complex conditions. The proposed method workflow integrates four key steps to enhance the consistency and accuracy of ocean polarization heading estimation.

[0039] Step 1: Divide the region of interest (ROI); ROI refers to the largest circumscribed circle area excluding clouds and ship buildings. The effective area is the area with good sky polarization pattern, which can refer to the full sky polarization area that is not blocked by clouds and ship buildings. It is included in the ROI, and ROI must be a valid area.

[0040] like Figure 5 As shown, this is a schematic diagram of the algorithm principle, combined with Figure 2 First, the all-sky DOFP polarization imaging system captures real-time skylight polarization pattern images, including the distribution of DOP and AOP in the sky. Based on these images, the system calculates the polarization degree DOP value, polarization angle AOP value, and DOP gradient (▽DOP) value, and then uses these data to establish an effective mask MASK to identify and isolate the effective area that is not blocked by clouds and hull buildings, thereby defining the ROI area.

[0041] It is generated by pixel-by-pixel calculation and is defined as follows:

[0042] (4)

[0043] in, and Indicates the DOP range threshold of the effective area, Set to the upper limit of the DOP gradient to exclude areas where the DOP pattern is severely disturbed by occlusion or clouds. (x, y) represents pixel coordinates.

[0044] Since the DOP gradient is usually low in clear sky areas, areas with dramatic DOP changes caused by cloud cover or other interference can be identified by high gradient values. Setting it to 0.017 can effectively exclude the occluded area without affecting the accuracy of ROI selection. After masking the occluded area, the largest continuous ROI is identified from the remaining valid DOP area and superimposed on the AOP image for subsequent processing.

[0045] Step 2: Rebuild the AOP mode;

[0046] Due to the tilt of the polarization compass, the sky area in the field of view no longer matches the corresponding area in the horizontal state. Therefore, it is necessary to establish a correspondence between the observed image pixels and the incident light. This includes converting the AOP image in the region of interest ROI from the observation (tilted) coordinate system to the navigation (horizontal) coordinate system, and then to the incident light coordinate system. Through denoising, the AOP pattern is finally reconstructed, that is, from Convert the graph to picture

[0047] like Figure 3 As shown in FIG. 1 , the celestial coordinate system includes the observation coordinate system, the navigation coordinate system and the incident light coordinate system. Wherein, O represents the position of the observer and P represents a point on the celestial sphere.

[0048] is the observation coordinate system. If the observation plane is horizontal, then Pointing to the zenith, if the observation plane is tilted, and denote the roll angle and pitch angle respectively, and They represent the azimuth and zenith angle in the observation coordinate system respectively.

[0049] is the navigation coordinate system, Pointing to the zenith, , is the azimuth and zenith angle of the observation point P in the navigation coordinate system.

[0050] is the incident light coordinate system, Pointing in the direction of observation, Here, i stands for incident light, which has changed the characters in other places. The AOP of incident light is the angle between the electric field vector E of the light and the horizontal direction on the polarization plane. This angle represents the direction of vibration of the electric field vector. express.

[0051] From the observation coordinate system To the navigation coordinate system The transformation matrix It can be expressed as:

[0052] (5)

[0053] Next, from the navigation coordinate system To the incident light coordinate system The transformation matrix It can be expressed as:

[0054] (6)

[0055] The AOP measured by the polarization system is:

[0056] (7)

[0057] therefore, It can be expressed as:

[0058] (8)

[0059] Step 3: Feature recognition and extraction;

[0060] The neutral point and the solar meridian are key features in the sky polarization pattern. The neutral point is the region of the sky where the degree of polarization (DOP) is zero and the angle of polarization (AOP) is rotated. The solar meridian is important for symmetry, especially in clear, cloudless conditions, where the sky polarization pattern is symmetrical along the solar meridian. The neutral point is usually located close to or along the solar meridian, but its exact location can vary depending on the solar altitude, observation time, and atmospheric conditions.

[0061] The AOP distribution of the direct imaging is in the observation coordinate system, which is conducive to identifying the neutral point feature. The neutral point is the intersection of the four regions of AOP -90°, -45°, 45° and 90°. This is different from the AOP distribution in the incident light coordinate system, which is more conducive to identifying the solar meridian feature for polarization navigation. Therefore, the present invention is more conducive to identifying the neutral point feature in the observation coordinate system. The graph is quantified to identify the output neutral point position and the incident light coordinate system after coordinate transformation. The solar meridian is detected and extracted from the map, and the neutral point position is used for judgment to ensure the correct calculation of the heading angle.

[0062] To identify and extract the neutral point, the algorithm first reads the observed AOP image of the skylight polarization pattern. in Represents the pixel coordinates of the image. After smoothing filtering, the noise interference is reduced. Here, Gaussian filtering is used, and its mathematical expression is:

[0063] (9)

[0064] in, represents the Gaussian kernel function, is the filter window size. Represents the relative offset in a small local neighborhood used during convolution.

[0065] Next, the smoothed observed AOP image The image is quantized and divided into four regions, representing -90°, -45°, 45°, and 90°. The quantization process is defined as:

[0066] (10)

[0067] After quantization, the image Edge detection is performed on the image to identify the boundaries of different AOP areas. Canny edge detection is used, and its expression is as follows:

[0068] (11)

[0069] in, is the detected edge image, indicating the boundaries between quantized regions.

[0070] The edge image Perform morphological operations (dilation and erosion) to connect the boundaries of adjacent regions. Get the processed edge image Find the point where the four quantization areas intersect, that is, the intersection of the four areas with AOP values ​​at -90°, -45°, 45° and 90°: ,in, Represents the set of detected neutral points.

[0071] Then, the detected neutral point set Use the DBSCAN clustering method to group and identify the true neutral point clusters. For each cluster, calculate its centroid coordinates to obtain the final location of the neutral point:

[0072] (12)

[0073] in, Indicates Neutral point cluster groups, is the centroid coordinate of the cluster. Finally, the centroid coordinates of all clusters It is output as the basis for determining the solar meridian.

[0074] For solar meridian identification and extraction, the AOP image of the incident skylight polarization mode is first read and smoothed as in formula (9) to reduce noise interference. The Canny operator is used to perform edge detection on the preprocessed image as in formula (11). The difference is that the Hough transform is then used to detect straight lines, and each edge point is Converted to polar coordinate space, the Hough transform equation is:

[0075] (13)

[0076] in, is the vertical distance from the point to the origin, is the angle between the vertical line and the horizontal axis. By accumulating votes, find the one with the largest cumulative value. The corresponding straight line.

[0077] All detected straight lines form a set In order to screen out the solar meridians, these lines are filtered according to their average AOP values, and only the lines close to ±90° are retained as candidate solar meridians. (14)

[0078] Using the neutral point coordinates The straight line that best meets the solar meridian condition among the candidate solar meridians is further determined. Distance from neutral point Do the calculations and choose the line closest to the neutral point:

[0079] (15)

[0080] in, The parametric equation is , Represents the neutral point and the straight line Choose the smallest distance. Corresponding As the solar meridian, where , , The coefficients that represent the parametric equation of a line, usually a constant.

[0081] The output is identified as the line equation of the solar meridian and its location in the image.

[0082] Step 4: Determine the heading angle;

[0083] The azimuth information of the solar meridian can be used to determine the heading angle of the carrier. The relationship between the solar meridian and the carrier heading is as follows: Figure 4 The angle between the solar meridian and the carrier heading is shown as When the DOFP imaging system is aligned with the carrier heading, the angle between the solar meridian and the true north direction in the "east-north" coordinate system is expressed as Using the position and time information provided by the positioning and time equipment, combined with the astronomical calendar method, the following calculations can be performed:

[0084] (16)

[0085] (17)

[0086] In the formula, is the sun's altitude, is the geographical latitude, is the solar right ascension angle, is the solar hour angle.

[0087] Therefore, the heading angle It can be calculated by the following formula:

[0088] (18)

[0089] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A skylight polarization navigation system suitable for marine environment, characterized in that: The system includes a gyro-stabilized platform, a DOFP polarization imager, a fisheye lens and a bandpass filter, wherein: The gyro-stabilized platform is fixed on the hull ground, the DOFP polarization imager is detachably mounted on the gyro-stabilized platform, the fisheye lens is mounted on the front end of the DOFP polarization imager, and the bandpass filter is placed in the C-mount interface of the DOFP polarization imager and is located behind the light outlet of the fisheye lens; The DOFP polarization imager includes a monochrome CMOS sensor, which includes a top layer of a microlens array, a middle layer of a polarization filter array, and a bottom layer of a super pixel array; the microlens array is used to focus light; the middle layer is a polarization filter array with directions of 0°, 45°, 90° and 135°; the bottom layer of the super pixel array is used to convert the optical signal into a digital signal; wherein each super pixel includes four sub-pixels, and the photodiode corresponding to each sub-pixel is aligned with a polarization filter of the middle layer of the polarization filter array; The system captures real-time skylight polarization pattern images, calculates polarization degree DOP values, polarization angle AOP values ​​and DOP gradient values, identifies and isolates effective areas that are not blocked by clouds or ship buildings based on the calculation results, and defines interest areas based on the effective areas, including performing pixel-by-pixel calculations based on polarization degree DOP values, polarization angle AOP values ​​and DOP gradient values ​​to establish an effective mask: (4) in, and Indicates the range threshold of the degree of polarization DOP value in the effective area, Set to the upper limit of the DOP gradient to exclude areas where the DOP pattern is severely disturbed by occlusion or clouds. (x, y) represents pixel coordinates; After excluding the occluded area, identifying the largest continuous region of interest from the remaining valid area, and superimposing the region of interest on the AOP image; Based on the AOP image recognition output in the observation coordinate system, the neutral point position and the solar meridian are output, and the heading angle of the carrier is determined according to the neutral point position and the solar meridian.

2. The skylight polarization navigation system suitable for marine environment according to claim 1, characterized in that: The fisheye lens has a 2 / 3 inch image circle, uses a C-mount interface, and has a focal length of 1.8 mm.

3. A skylight polarization navigation method suitable for marine environment, applied to a skylight polarization navigation system suitable for marine environment as claimed in claim 1 or 2, characterized in that: The steps include: Step 1: After superimposing the region of interest on the AOP image, the AOP image is converted from the observation coordinate system to the navigation coordinate system, and then converted to the incident light coordinate system, and the AOP pattern is reconstructed through denoising; Step 2, quantify the AOP image in the observation coordinate system, identify the output neutral point position, and detect and extract the solar meridian of the AOP image converted to the incident light coordinate system; Step 3: Determine the heading angle of the carrier based on the extracted neutral point position and the solar meridian.

4. The skylight polarization navigation method suitable for marine environment according to claim 3, characterized in that: In step 1, O represents the position of the observer, P represents a point on the celestial sphere, is the observation coordinate system, and Denote the roll angle and pitch angle respectively; is the navigation coordinate system, , is the azimuth and zenith angle of the observation point P in the navigation coordinate system; is the incident light coordinate system, represents the direction of vibration of the electric field vector; then the vector From the observation coordinate system To the navigation coordinate system The transformation matrix It is expressed as: (5) From the navigation coordinate system To the incident light coordinate system The transformation matrix It is expressed as: (6) The polarization angle AOP measured by the polarization system is: (7) Therefore, the direction of the electric field vector vibration is It is expressed as: (8)。 5. The skylight polarization navigation method suitable for marine environment according to claim 3, characterized in that: The step 2 includes smoothing the observed AOP image The image is quantized and divided into four regions, representing -90°, -45°, 45°, and 90° respectively. The quantization process is defined as: (10) After quantization, the image Perform edge detection on the AOP to identify the boundaries of different AOP regions: (11) in, is the detected edge image, indicating the boundary between quantized regions, and Canny represents the edge detection operation; The detected edge image Perform morphological operations to connect the boundaries of adjacent regions and obtain the processed edge image , find the intersection point of the four quantization areas, that is, the intersection point between the four areas of polarization angle AOP values ​​at -90°, -45°, 45° and 90° ,in, represents the set of detected neutral points; For the detected neutral point set Use the DBSCAN clustering method to group and identify the true neutral point clusters. For each cluster, calculate its centroid coordinates to obtain the final location of the neutral point: (12) in, Indicates is the centroid coordinate of the cluster; For solar meridian identification and extraction, the AOP image of the incident skylight polarization pattern is read, and after smoothing, the edge detection is performed using formula (11). The Hough transform is used to transform each edge point Converted to polar coordinate space, the Hough transform equation is: (13) in, is the vertical distance from the point to the origin, is the angle between the vertical line and the horizontal axis. By voting cumulatively, find the one with the largest cumulative value The corresponding straight line; All detected straight lines form a set , for the set The average AOP value is filtered and the straight lines close to ±90° are retained as candidate solar meridians: （14） Using the neutral point coordinates Determine the line among the candidate solar meridians that best satisfies the solar meridian criteria: (15) in, The parametric equation is , Represents the neutral point and the straight line distance, choose the smallest Corresponding As the solar meridian, represents the constant coefficient of the parametric equation of the line; The output is identified as the line equation of the solar meridian and its location in the image.

6. The skylight polarization navigation method suitable for marine environment according to claim 3, characterized in that: The step 3 comprises using the position and time information provided by the positioning and timing equipment and combining it with the astronomical calendar method to calculate: (16) (17) In the formula, is the sun's altitude, is the geographical latitude, is the solar right ascension angle, is the solar hour angle, is the angle between the solar meridian and the carrier heading; Heading Angle Calculated by the following formula: (18) In the formula, It represents the angle between the solar meridian and the true north direction in the "east-north" coordinate system when the skylight polarization navigation suitable for the ocean environment is aligned with the carrier heading.

Images