Visual celestial navigation method based on sun centroid extraction and sea-sky-line fitting

Abstract

The invention provides a visual celestial navigation method based on sun centroid extraction and sea-sky-line fitting. Belongs to the application of image processing technology in the field of celestial navigation. The method mainly comprises the steps of extracting a high-precision sea-sky-line by using a textural feature segmentation and Canny-Hough transform combined algorithm, and extracting a sky region based on a sea-sky-line segmentation result. A solar facula image candidate region and a sun candidate contour are constructed through a sky region, a dual optimization strategy of region similarity screening and facula feature scoring is proposed, and sub-pixel-level sun centroid coordinate calculation is realized by combining an intensity weighted center method and nonlinear least square two-dimensional Gaussian model fitting. And further constructing a camera coordinate system space vector through the orthogonal projection point of the sea-sky-line and the sun centroid, calculating a sun elevation angle based on a vector included angle, and performing angle compensation by adopting an atmospheric refraction correction model. And finally, fusing multi-moment observation data through a sun shift line positioning method, completing autonomous calculation of the ship position, and improving the positioning precision.

Description

Technical Field

[0001] The invention belongs to the technical field of image processing, and in particular relates to a visual astronomical navigation method based on sun centroid extraction and sea-sky line fitting. Background Art

[0002] During ocean voyages, satellite navigation systems like GPS and BeiDou can fail due to factors such as ionospheric disturbances, equipment failures, or military interference. This exposes traditional navigation methods to critical issues such as positioning signal interruption, accumulated coordinate drift, and a loss of heading reference. This poses a serious threat to the operational reliability of maritime safety systems, including ship route planning, attitude control, and emergency avoidance. Therefore, to establish an autonomous navigation system independent of satellite signals and enhance positioning robustness in extreme environments, it is necessary to focus on developing visual astronomical navigation technology based on celestial elevation angle measurement and astronomical triangulation.

[0003] Currently, researchers at home and abroad have proposed numerous theories for visual astronomical navigation, which can be broadly categorized into three types. The first category involves traditional navigation methods based on geometric optics. For example, these methods use a sextant to measure the solar altitude and combine accurate time with ephemeris to calculate geographic coordinates. This method relies on clear horizon visibility and a stable atmospheric refraction correction model. However, in practice, this method is susceptible to turbulent sea conditions, cloud cover, and rapid changes in the solar altitude at low latitudes, leading to cumulative positioning errors. The second category involves astronomical navigation methods based on multi-star positioning vector technology. Through large-field-of-view imaging and star map filtering, multiple celestial targets are simultaneously extracted, matched, and identified for navigation settlement. Common star map recognition and matching methods include those based on star vector features and those based on star statistical features. However, because these methods require imaging of three or more stars and place high demands on equipment, their navigation performance in daylight scenarios is suboptimal. The third category involves celestial navigation methods based on sun sensors and attitude meters. These methods use sun sensors to obtain the heliocentric direction vector. Combined with the three-axis attitude angle data output by the attitude meter, a geometric positioning model based on celestial coordinate system transformation is constructed to obtain the observed horizontal and elevation angles of the sun, providing direct observational data for navigation. These methods have a clear physical observation model and a verifiable navigation mechanism, and can be implemented with relatively inexpensive equipment. A typical approach is astronomical positioning using an array camera based on parallax measurement, which generally produces ideal navigation results. This method uses an attitude and heading reference system (AHRS) to measure the camera's Euler angles, transform the pixel coordinates of the sun's center of mass in the original image into the sun's center of mass direction vector in the horizontal coordinate system, and calculate the sun's altitude angle as data for astronomical positioning. However, since the camera itself does not have the ability to output its own attitude angles, equipment assembly errors are inevitable, and the calculated sun's center of mass direction vector is prone to deviation and lacks stability. Summary of the Invention

[0004] The present invention provides a visual celestial navigation method based on solar centroid extraction and sea-sky-line fitting, so as to solve the problem that the existing visual celestial navigation algorithm calculates the solar altitude angle inaccurately, resulting in unsatisfactory positioning results, and improve the positioning accuracy.

[0005] The technical means adopted in the present invention are as follows: A visual astronomical navigation method based on solar centroid extraction and sea-sky-line fitting includes the following steps: S1. Acquire a target image, where the target image includes the sun and the sea-sky line; S2. performing a dedistortion process on the target image to obtain a dedistorted image, and then performing a cropping process on the dedistorted image to obtain a valid area image; S3, extracting the sea-sky line from the effective area image; S4. Extracting a sky area from the dedistorted image based on a sea-sky line, then obtaining a sun spot image candidate area and a sun candidate outline from the sky area, and determining sub-pixel accurate sun center of mass coordinates by combining the sun spot image candidate area and the sun candidate outline; S5. Obtaining a spatial angle formed by the center of mass of the sun and the orthogonal projection point of the sea-sky line, and correcting the spatial angle based on an atmospheric refraction correction formula to obtain the solar altitude angle at the current observation time and location; S6. Perform dead reckoning based on the initial ship position, course and speed to obtain the estimated ship position one hour later; calculate the theoretical altitude / azimuth of the sun at the initial moment and one hour later respectively, calculate the intercept based on the measured altitude, and draw two astronomical ship position lines on the Mercator projection chart; after translating the initial position line along the course for the actual distance voyaged, the intersection of the translated initial position line and the position line one hour later is the updated actual observed ship position.

[0006] Furthermore, the target image is one of an infrared image, a visible light image and a polarized light image.

[0007] Furthermore, obtaining a candidate area of ​​the sun spot image from the sky area includes: An ideal Gaussian template is used to simulate the sun spot image. The Pearson correlation coefficient method is used to calculate the normalized correlation coefficient between the simulated sun spot image and a small area of ​​the same size in the sky area. The area with the required confidence level is taken as the candidate area of ​​the sun spot image.

[0008] Furthermore, a candidate sun outline is obtained from the sky area, including: Morphological operations are performed on the sky area to obtain several candidate sun outlines, which are then scored and screened based on the sun spot features. The scoring formula based on the sun spot features is:

[0009] in, Indicates the i The brightness weight of the candidate sun contour area, Indicates the i Region shape weights, Indicates the i The normalized average brightness of the region, Indicates the i The circularity and ellipticity of each region are weighted, and the expression is as follows:

[0010]

[0011]

[0012]

[0013] in, represents all the pixels in the i-th sun candidate contour area, Indicates the brightness variance of the current candidate sun contour area. Represents the shape variance of the current sun candidate contour area, Indicates the circularity of the current sun candidate contour area. Indicates the ellipticity of the current sun candidate contour region.

[0014] Furthermore, the calculation formula of the space angle is:

[0015] in, represents the pixel coordinates of the sun's center of mass, K represents the camera intrinsic parameter matrix, Represents the pixel coordinates of the orthogonal projection point on the sea-sky line.

[0016] This invention provides a visual celestial navigation method based on sun centroid extraction and sea-sky line fitting. First, an image containing the sun and sea-sky line is input, including but not limited to infrared, visible light, and polarized light images. Distortion correction is performed using the Zhang Zhengyou calibration method, and the effective observation area is cropped to provide a distortion-free image base for subsequent processing. Next, the sea-sky area is extracted through texture feature segmentation. Canny edge detection and Hough transform linear fitting are combined to accurately locate the sea-sky line. For solar targets, candidate regions are screened based on regional similarity and solar spot feature scoring. Sub-pixel centroid positioning is achieved using the intensity-weighted center method and two-dimensional Gaussian model fitting. Based on this, an orthogonal projection camera model is constructed. The sun centroid and the sea-sky line tangent point are mapped into three-dimensional vectors, and the angle between the vectors is calculated. An empirical formula for atmospheric refraction is introduced to correct for environmental interference, resulting in an accurate true solar altitude angle output. Finally, the method integrates dead reckoning with celestial navigation principles. A theoretical altitude angle is generated through dead position prediction. The astronomical position line is constructed by combining the measured altitude angle difference. Multi-time observation data fusion positioning is achieved through voyage translation.

[0017] Compared with the prior art, the present invention has the following advantages: 1. The system input images used in the present invention cover images containing sea-skyline and sun elements. It not only supports the input of visible light images, but is also compatible with multiple types such as infrared images and polarized light images.

[0018] 2. The present invention adopts a scoring mechanism based on regional similarity and sun spot characteristics to extract the region of interest containing the sun spot, calculates the initial centroid coordinates based on the intensity weighted center algorithm, reduces the number of iterations of the least squares method to achieve two-dimensional Gaussian fitting, and obtains high-precision centroid coordinates.

[0019] 3. The present invention combines the camera orthogonal projection model to directly measure the space vector angle between the center of mass of the sun and the tangent point of the sea-sky line, introduces the empirical formula of atmospheric refraction to correct environmental interference, and outputs the accurate true solar altitude angle. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0021] Figure 1 The flowchart of a visual celestial navigation method based on solar centroid extraction and sea-sky line fitting in an embodiment of the present invention is shown.

[0022] Figure 2 is the original image input in the embodiment of the present invention.

[0023] Figure 3 This is a dedistorted image corrected by the camera distortion model in an embodiment of the present invention.

[0024] Figure 4 This is the segmented sea-sky line area in the embodiment of the present invention.

[0025] Figure 5 This is a sea-sky-line detection diagram in an embodiment of the present invention.

[0026] Figure 6 This is the sun region obtained by screening the dedistorted image in the embodiment of the present invention.

[0027] Figure 7 This is the solar center coordinate diagram obtained by sub-pixel center extraction of the sea-sky line detection diagram in an embodiment of the present invention. DETAILED DESCRIPTION

[0028] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0029] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0030] like Figure 1 As shown, the present invention provides a visual astronomical navigation method based on solar centroid extraction and sea-sky line fitting, which specifically includes the following steps.

[0031] S1. Acquire a target image, where the target image includes the sun and the sea-sky line.

[0032] This step is used to input a target image containing the sun and sea-sky line. I .

[0033] S2. Dedistort the target image to obtain a dedistorted image, and then crop the dedistorted image to obtain a valid region image. If the rectified image cannot completely fill the rectangular canvas, the minimum bounding rectangle of the contour is calculated through image binarization and contour detection, serving as the cropping region.

[0034] This step is used to I Perform dedistortion processing and use Zhang Zhengyou calibration method to obtain the camera intrinsic parameter matrix K , radial distortion parameters and tangential distortion parameters The radial distortion parameter describes the degree of deviation of the image point along the radial direction (from the center of the image to the outside), and the tangential distortion parameter describes the degree of deviation of the image point along the tangential direction (perpendicular to the radial direction). The dedistorted image is obtained through the camera distortion model. I_d . Set the cropping range and crop the dedistorted image I_d The invalid area is obtained to obtain the valid area image.

[0035] S3. Extracting the sea-sky line from the effective area image.

[0036] In this application, the sea-sky region is obtained by extracting the sea-sky region based on texture features. I_c , using the A median filter suppresses salt-and-pepper noise, and Canny edge detection and Hough transform line detection are used to extract the sea-sky line in the image. Alternatively, texture-based sea-sky region extraction methods can be found in the following literature.

[0037] Zhang, Y., Hu, Q., Li, D.et al.Texture feature-based local adaptiveOtsu segmentation and Hough transform for sea-sky line detection.MultimedTools Appl83, 34477–34498 (2024). S4. Extracting a sky area from the dedistorted image based on a sea-sky line, then obtaining a sun spot image candidate area and a sun candidate outline from the sky area, and determining sub-pixel accurate sun center of mass coordinates by combining the sun spot image candidate area and the sun candidate outline.

[0038] This step is based on the sea-sky-line detection results of S3. First, the dedistorted image is I_d Perform sky region segmentation. Use an ideal Gaussian template to simulate a sun spot image, calculate its normalized cross-correlation coefficient with a small area of ​​the same size in the sky area, and select areas with a confidence level above 0.7 as candidate regions. Next, perform morphological operations on the sky area and score and screen based on the sun spot characteristics to obtain candidate sun contours. Combine the candidate regions and candidate contours, and select the ROI with the greatest overlap to determine the ROI. In the ROI slice, the intensity-weighted center method is used, combined with the nonlinear least squares method to fit a two-dimensional Gaussian model, and output the sub-pixel accurate sun center of mass coordinates.

[0039] The scoring formula based on the sun spot feature is as follows:

[0040] in, represents the brightness weight, represents the shape weight, represents the normalized average brightness of each contour, The circularity and ellipticity weights of each contour are expressed as follows:

[0041]

[0042]

[0043]

[0044] in, represents all the pixels in the i-th sun candidate contour area, Indicates the brightness variance of the current candidate sun contour area. Represents the shape variance of the current sun candidate contour area, Indicates the circularity of the current sun candidate contour area. Indicates the ellipticity of the current candidate sun contour area. Default weight is 0.5.

[0045] The circularity and ellipticity formulas are as follows:

[0046]

[0047] in, and denote the area and perimeter of the contour, respectively. and They represent the lengths of the major and minor axes of the fitted ellipse, respectively.

[0048] S5. Obtain a spatial angle formed by the center of mass of the sun and the orthogonal projection point of the sea-sky line, and correct the spatial angle based on an atmospheric refraction correction formula to obtain the solar altitude angle at the current observation time and location.

[0049] According to the sextant measurement principle, when the edge of the sun is tangent to the sea-sky line, the solar altitude angle is the spatial angle formed by the orthogonal projection point of the sun's center of mass and the sea-sky line. First, the sun's center of mass in the image pixel coordinate system is Point of tangency with the sea-sky line The coordinates are converted to the corresponding coordinates in the image's physical coordinate system. Next, the three-dimensional spatial vectors of the two are solved in the camera coordinate system using an orthogonal projection of the camera model. The geometric angle between the two vectors is then calculated. This angle is corrected using an atmospheric refraction correction algorithm based on empirical formulas to obtain the solar altitude angle at the current observation time and location.

[0050] The spatial geometric angle is calculated using the following formula:

[0051] in, represents the pixel coordinates of the sun's center of mass, K represents the camera intrinsic parameter matrix, It represents the pixel coordinates of the orthogonal projection point on the sea-sky line. Its expression is as follows:

[0052]

[0053]

[0054] in, D represents the algebraic offset from the sun's center of mass to the sea-horizon line, represents the sea-sky-line inclination angle, Indicates the polar diameter of the sea-sky line.

[0055] The atmospheric refraction correction formula is:

[0056] in, P Indicates the atmospheric pressure value, T Indicates the temperature value.

[0057] The final measured true solar altitude angle is:

[0058] S6. Perform dead reckoning based on the initial ship position, heading and speed to obtain the estimated ship position one hour later; calculate the theoretical sun altitude / azimuth at two times respectively, calculate the intercept based on the measured altitude, and draw two astronomical position lines on the Mercator projection chart; after translating the initial position line along the heading for the actual voyage, the intersection of the two lines is the updated actual observed ship position.

[0059] The scheme and effects of the present invention are further illustrated below through specific application examples.

[0060] like Figure 2-7 As shown, this embodiment uses an image containing the sea-sky line and the sun as the system input, such as Figure 2 As shown. First, the input image is distorted and corrected to obtain Figure 3 The corrected image shown in the figure is then used to extract the sea-sky region using a sea-sky region extraction method based on texture features, as shown in Figure 4 The median filter is used to suppress the salt and pepper noise, and the sea-sky line in the image is extracted through Canny edge detection and Hough transform line detection, as shown in Figure 5 The real sun spot area is determined by combining the sun spot image candidate area and the sun candidate outline, as shown in FIG. Figure 6 And calculate the sub-pixel accurate solar center coordinates, as shown in Figure 7 Then, the spatial angle formed by the orthogonal projection point of the sun's mass center and the sea-sky line is obtained, and the atmospheric refraction correction formula is used to correct it, and the obtained Observe the solar altitude angle at all times. Finally, repeat the above steps after 1 hour and get Finally, the astronomical position line is constructed by combining the measured altitude angle difference, and the astronomical position line is determined by the voyage translation. Time position.

[0061] This method utilizes a celestial navigation method based on visual imaging, sea-sky line fitting, and solar centroid extraction. First, input images containing the sun and sea-sky line, including but not limited to infrared, visible, and polarized images, are used. Zhang Zhengyou's calibration method is used to correct distortion and crop the effective observation area, providing a distortion-free image base for subsequent processing. Next, the sea-sky area is extracted through texture feature segmentation. Canny edge detection and Hough transform linear fitting are combined to accurately locate the sea-sky line. For the solar target, candidate regions are screened based on regional similarity and solar spot feature scores. Sub-pixel centroid localization is achieved using the intensity-weighted center method and two-dimensional Gaussian model fitting. Based on this, an orthogonal projection camera model is constructed. The solar centroid and the sea-sky line tangent point are mapped into three-dimensional vectors, and the angle between the vectors is calculated. An empirical formula for atmospheric refraction is introduced to correct for environmental interference, resulting in an accurate true solar elevation angle. Finally, the method integrates dead reckoning with celestial navigation principles. A theoretical elevation angle is generated through dead position prediction. The astronomical position line is constructed by combining the measured elevation angle difference. Positioning is achieved by fusion of multi-time observation data through voyage translation.

[0062] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A visual astronomical navigation method based on solar centroid extraction and sea-sky-line fitting, characterized in that: The following steps are involved: S1. Acquire a target image, where the target image includes the sun and the sea-sky line; S2. performing a dedistortion process on the target image to obtain a dedistorted image, and then performing a cropping process on the dedistorted image to obtain a valid area image; S3, extracting the sea-sky line from the effective area image; S4. Extracting a sky area from the dedistorted image based on a sea-sky line, then obtaining a sun spot image candidate area and a sun candidate outline from the sky area, and determining sub-pixel accurate sun center of mass coordinates by combining the sun spot image candidate area and the sun candidate outline; S5. Obtaining a spatial angle formed by the center of mass of the sun and the orthogonal projection point of the sea-sky line, and correcting the spatial angle based on an atmospheric refraction correction formula to obtain the solar altitude angle at the current observation time and location; S6. Perform dead reckoning based on the initial ship position, course and speed to obtain the estimated ship position one hour later; calculate the theoretical altitude / azimuth of the sun at the initial moment and one hour later respectively, calculate the intercept based on the measured altitude, and draw two astronomical ship position lines on the Mercator projection chart; after translating the initial position line along the course for the actual distance voyaged, the intersection of the translated initial position line and the position line one hour later is the updated actual observed ship position.

2. The visual celestial navigation method based on sun centroid extraction and sea-sky line fitting according to claim 1, characterized in that: The target image is one of an infrared image, a visible light image and a polarized light image.

3. The visual celestial navigation method based on sun centroid extraction and sea-sky line fitting according to claim 1, characterized in that: Obtain the candidate area of ​​the sun spot image from the sky area, including: An ideal Gaussian template is used to simulate the sun spot image. The Pearson correlation coefficient method is used to calculate the normalized correlation coefficient between the simulated sun spot image and a small area of ​​the same size in the sky area. The area with the required confidence level is taken as the candidate area of ​​the sun spot image.

4. The visual celestial navigation method based on sun centroid extraction and sea-sky line fitting according to claim 1, characterized in that: Get the sun candidate outline from the sky area, including: Morphological operations are performed on the sky area to obtain several candidate sun outlines, which are then scored and screened based on the sun spot features. The scoring formula based on the sun spot features is: Among them, w b (i) represents the brightness weight of the i-th sun candidate contour area, w s (i) represents the shape weight of the i-th region, represents the normalized average brightness of the i-th region, S i represents the weighted circularity and ellipticity of the i-th region, and the expression is as follows: S i =αC i +(1-α)∈ i Among them, Ω i represents all the pixels in the i-th sun candidate contour area, Indicates the brightness variance of the current sun candidate contour area, Represents the shape variance of the current sun candidate contour area, C i Indicates the circularity of the current sun candidate contour area, ∈ i Indicates the ellipticity of the current sun candidate outline region.

5. The visual celestial navigation method based on sun centroid extraction and sea-sky line fitting according to claim 1, characterized in that: The calculation formula of the space angle is: Among them, (u sun ,v sun ) represents the pixel coordinates of the sun's center of mass, K represents the camera's intrinsic parameter matrix, (u p ,v p ) represents the pixel coordinates of the orthogonal projection point on the sea-sky line.

Images