A method for directional orientation of atmospheric polarized light using solar azimuth region

Abstract

This invention discloses a method for orientation using atmospheric polarized light in the solar azimuth region, provides a visual modulation model of polarization angle in pixel coordinates, and presents an orientation strategy utilizing the solar azimuth region. This invention incorporates the horizontal attitude angle into the polarization angle in the pixel coordinates and achieves orientation using the separation point and the image principal point. Specifically, a seed line extraction algorithm is provided to update the solar azimuth in real time, fusing the morphological representation and polarization angle representation of the solar azimuth for optimal estimation. Finally, the horizontal attitude angle is input from an external sensor, such as inertial navigation, to complete the orientation. This invention can directly calculate the polarization angle in the pixel coordinates under tilted conditions, effectively improving the orientation accuracy and speed of atmospheric polarized light navigation.

Description

Technical Field

[0001] This invention belongs to the field of navigation technology, specifically a method for orientation using atmospheric polarized light in the solar azimuth region. Background Technology

[0002] By detecting the regular patterns formed by AOP / DOP on the celestial sphere, passive, error-free, and precise navigation can be achieved. It is generally believed that the AOP pattern observed from the ground or near the ground is identical in nature under any weather conditions. Due to this robustness of the AOP pattern, navigation methods utilizing AOP are considered to have great potential. Generally, based on the detection device, biomimetic polarization navigation methods can be divided into two types: point-source polarization navigation methods based on photodiodes and imaging polarization navigation methods based on image polarimeters. Point-source polarization navigation methods based on photodiodes require additional sensors to address solar ambiguity. Achieving a large field of view in this type of method necessitates complex mechanical structures and long scan times. The polarization angle obtained using this method results in each pixel having its own coordinate system, i.e., the incident light coordinate system. Furthermore, this navigation method, due to its small field of view, is highly susceptible to interference from weather, obstructions, and other factors.

[0003] To address the aforementioned issues, an imaging polarization navigation method utilizing an image polarimeter has been proposed. However, it is regrettable that a theoretical model for this method is currently lacking. Furthermore, imaging polarization navigation methods primarily focus on the polarization angle in the incident light coordinate system, with limited research on the polarization angle in the pixel system. More critically, a pixel-based polarization navigation method for tilted orientations is missing. The inability to resolve orientation issues under tilted orientations severely limits the application of imaging polarization navigation methods; in clear skies, this navigation error is considered the largest source of error in polarization navigation methods. Summary of the Invention

[0004] Purpose of the invention: In order to solve the problem that existing navigation and orientation methods cannot use the polarization angle in the pixel coordinate system for navigation and orientation under tilt conditions, this invention provides a method for orientation using atmospheric polarized light in the solar azimuth region.

[0005] Technical solution: A method for directional orientation of atmospheric polarized light using solar azimuth regions, comprising the following steps:

[0006] Step 1: Use an image polarimeter to detect four-channel polarization images, and calculate the polarization angle based on the four-channel polarization images.

[0007] Step 2: Obtain the solar altitude angle and solar azimuth angle, and obtain the observation point altitude angle and observation point azimuth angle. Using the pixel coordinate system theoretical polarization angle model, convert the polarization angle in the incident light coordinate system to the polarization angle in the pixel coordinate system, and establish a theoretical model of the polarization angle. Ignoring measurement noise, the obtained polarization angle is the polarization angle calculated from the four-channel polarization image. The polarization angle is obtained through the theoretical model of the polarization angle. With solar azimuth Relationships;

[0008] Step 3: Use the region growing model to adjust the polarization angle. The solar azimuth region is extracted, and the initial growth point is set as the image principal point. Region growth is performed on the current polarization angle image at the initial growth point, and the grown binary image is calculated. The grown binary image is filtered using image morphology operations, the centroid of the filtered binary image is calculated, the seed line is updated according to the centroid, and the updated seed line is used as the starting growth point for the solar azimuth region extraction of the next frame of polarization angle image.

[0009] Step 4: Calculate the first approximate solar azimuth angle based on the updated seed line. The binary image of the sun is divided into equal-field-of-view image blocks along the sun's direction. The second approximate solar azimuth angle is then calculated using these equal-field-of-view image blocks. All measurements of solar azimuth angle Represented as: according to The optimal estimate of the solar azimuth angle was calculated. C is to make The quantity to be determined when the minimum value is reached;

[0010] The horizontal attitude angle is calculated using the inertial navigation system. Let the absolute heading angle be ψ. The absolute heading angle ψ is calculated based on all measured solar azimuth angles and their optimal estimates.

[0011]

[0012] in, Represents the solar azimuth angle in the carrier coordinate system.

[0013] Furthermore, in step one, let the polarization angle be... The formula for calculating the polarization angle is:

[0014]

[0015] Where I(*) represents the polarization intensity in the * direction.

[0016] Furthermore, in step two, the polarization angle in the incident light coordinate system is expressed as:

[0017]

[0018] Among them, h s A s h represents the solar altitude angle and solar azimuth angle. p With A p The elevation angle and azimuth angle of the observation point;

[0019] Transform the polarization angle from the incident light coordinate system to the pixel coordinate system, then the polarization angle in the pixel coordinate system is... It can be represented as:

[0020]

[0021] In the formula:

[0022]

[0023] in:

[0024]

[0025] In the formula, sc is the field of view angle, rl is the resolution, (x p y p () represents the pixel position in the pixel coordinate system. These are the coordinates of the principal point of the image in the pixel coordinate system.

[0026] right Simplifying, we get:

[0027]

[0028] in:

[0029]

[0030] Taking the observation direction as the direction of the sun, the above formula simplifies to:

[0031]

[0032] Introducing the three-axis attitude angle, similarly, we obtain the polarization angle in the pixel coordinate system:

[0033]

[0034] in:

[0035]

[0036] In the formula, the superscript b represents the carrier coordinate system;

[0037] Similarly, simplifying the above equation along the direction of the sun:

[0038]

[0039] in, This represents the solar azimuth angle in the carrier coordinate system.

[0040] Furthermore, in step three, the grown binary image B AOP The calculation method for (i, j) is as follows:

[0041]

[0042] Where i and j are the width and height of a pixel, and RG represents the region growing model.

[0043] Furthermore, in step three, the grown binary image is filtered using image morphology shrinkage and clearing operations. The filtered binary image B SR (i, j) is represented as:

[0044] B SR (i, j) = C(S(B) AOP (i,j))) (18).

[0045] Furthermore, in step three, the method for calculating the centroid of the filtered binary image is as follows:

[0046]

[0047] in:

[0048]

[0049] x s y s Let m represent the x and y coordinates of the centroid, respectively. pq Represents the relationship between the values ​​of p and q and B. SR The calculated values ​​of (i, j), p, q ∈ 0, 1, m 10 This represents the calculated value when p is 1 and q is 0; m 01 The calculated value represents p = 0 and q = 1; m 00 This represents the calculated value when p is 0 and q is 0.

[0050] Furthermore, in step three, the updated seed line is represented as follows:

[0051]

[0052] Where X and Y are the independent and dependent variables in the linear equation, These are the coordinates of the principal point of the image in the pixel coordinate system.

[0053] Furthermore, in step four, the first approximate solar azimuth angle The calculation method is as follows:

[0054]

[0055] Furthermore, in step four, the binary image of the sun direction is divided into n equal-field-of-view image blocks along the sun direction and represented as B. AOP，1 (i, j), B AOP，2 (i, j), ..., B AOP，n (i, j), then the set of equal field-of-view image patches B FOV (i, j) is represented as:

[0056] B FOV (i, j) = [B AOP，1 (i, j), B AOP，2 (i, j), ..., B AOP，n (i, j)] (23)

[0057] Calculating the second approximate solar azimuth using equal field-of-view image patches

[0058]

[0059] Where M and N are the width and height of the pixel position with a value of 1 in the equal field of view image block.

[0060] Beneficial Effects: Compared with existing polarized light orientation methods, this invention directly utilizes the polarization angle in the pixel coordinate system for navigation, simplifying the orientation process for atmospheric polarized light navigation. This invention establishes a polarization angle visual modulation model in the pixel coordinate system and performs morphological representation of the solar azimuth region of the polarization angle image before orientation. It introduces a seed line and a horizontal attitude angle to perform navigation calculations on the obtained solar azimuth region, achieving accurate orientation of the polarized light compass under tilt conditions. This invention can directly utilize the polarization angle in the pixel coordinate system to achieve atmospheric polarized light orientation under tilt conditions, solving the problem of significantly reduced orientation accuracy due to tilt during atmospheric polarized light orientation, thereby effectively improving the orientation accuracy and speed in atmospheric polarized light navigation. Attached Figure Description

[0061] Figure 1 A flowchart illustrating the atmospheric polarization orientation method using the solar azimuth region;

[0062] Figure 2 A comparison chart of orientation results using various orientation methods at a 360° heading angle;

[0063] Figure 3 A comparison chart of polarization angle errors for various orientation methods under tilted conditions. Detailed Implementation

[0064] The present invention will be further explained and described below with reference to the accompanying drawings and specific embodiments.

[0065] A method for orientation of atmospheric polarized light using solar azimuth regions (SAO), such as... Figure 1 As shown, it includes the following steps:

[0066] Step 1: Use an image polarimeter to detect a four-channel polarization image. The four channels are the following four directions, and the polarization angle is set as follows: The polarization angle is calculated from the four-channel polarization image using the following formula:

[0067]

[0068] Where I(*) represents the polarization intensity in the * direction.

[0069] Step 2: Establish a theoretical model for the polarization angle to obtain the polarization angle. With solar azimuth The relationship between them lays the theoretical foundation for subsequent calculations.

[0070] To meet real-time navigation requirements, the polarization angle is simplified based on the theoretical polarization angle model in the pixel coordinate system.

[0071] The solar elevation angle and solar azimuth angle are obtained, as are the elevation angle and azimuth angle of the observation point. Using the pixel coordinate system theoretical polarization angle model, the polarization angle in the incident light coordinate system is converted into the polarization angle in the pixel coordinate system, thus establishing a theoretical model of the polarization angle.

[0072] Specifically, in the pixel coordinate system theoretical polarization angle model calculation, it is assumed that the electric vector direction is always perpendicular to the plane formed by the observation point, the sun's projection point, and the observer. Therefore, when observing the sun and dust points in the sky in the horizon frame, the 3D model of the polarization angle in the incident light coordinate system can be expressed as:

[0073]

[0074] Among them, h s A s h represents the solar altitude angle and solar azimuth angle. p With A p The elevation angle and azimuth angle of the observation point;

[0075] Transform the polarization angle from the incident light coordinate system to the pixel coordinate system, then the polarization angle in the pixel coordinate system is... It can be represented as:

[0076]

[0077] In the formula:

[0078]

[0079] in:

[0080]

[0081] In the formula, sc is the field of view angle, rl is the resolution, (x p y p () represents the pixel position in the pixel coordinate system. These are the coordinates of the principal point of the image in the pixel coordinate system.

[0082] right Simplifying, we get:

[0083]

[0084] in:

[0085]

[0086] Taking the observation direction as the direction of the sun, the above formula simplifies to:

[0087]

[0088] Introducing the three-axis attitude angle, similarly, we obtain the polarization angle in the pixel coordinate system:

[0089]

[0090] in:

[0091]

[0092] In the formula, the superscript b represents the carrier coordinate system;

[0093] Similarly, simplifying the above equation along the direction of the sun:

[0094]

[0095] in, The solar azimuth angle is in the carrier coordinate system. Ignoring measurement noise, the obtained polarization angle is the polarization angle calculated from the four-channel polarization image. The relationship between the polarization angle and the solar azimuth angle is obtained through the theoretical model of the polarization angle. Therefore, navigation and orientation can be completed by extracting the polarization angle in the solar direction.

[0096] Step 3: Perform seed line extraction algorithm. Use a region growing model to extract the solar azimuth region based on the polarization angle. Set the initial growth point of the region growing model as the principal point of the image. Perform region growing on the current polarization angle image at the initial growth point and calculate the grown binary image B. AOP The calculation method for (i, j) is as follows:

[0097]

[0098] Where i and j are the width and height of a pixel. RG represents the region growing model.

[0099] The grown binary image is filtered using image morphology shrinkage and clearing operations. The filtered binary image B SR (i, j) is represented as:

[0100] B SR (i, j) = C(S(B) AOP (i, j))) (18)

[0101] Calculate the centroid of the filtered binary image:

[0102]

[0103] in:

[0104]

[0105] x s y s Let m represent the x and y coordinates of the centroid, respectively. pq Represents the relationship between the values ​​of p and q and B. SR The calculated values ​​of (i, j), p, q ∈ 0, 1, m 10 This represents the calculated value when p is 1 and q is 0; m 01 The calculated value represents p = 0 and q = 1; m 00 This represents the calculated value when p is 0 and q is 0.

[0106] The seed line is updated based on the centroid, and the updated seed line is used as the starting growth point for extracting the solar azimuth region in the next frame of the polarization angle image. The updated seed line is represented as follows:

[0107]

[0108] Where X and Y are the independent and dependent variables in the linear equation, These are the coordinates of the principal point of the image in the pixel coordinate system.

[0109] Step 4: Calculate the first approximate solar azimuth angle based on the updated seed line. The calculation method is as follows:

[0110]

[0111] Divide the binary image along the sun's direction into n equal-field-of-view image blocks and represent them as B. AOP，1 (i, j), B AOP，2 (i, j), ..., BAOP，n (i, j), image block B with equal field of view FOV (i, j) is represented as:

[0112] B FOV (i, j) = [B AOP，1 (i, j), B AOP，2 (i, j), ..., B AOP，n (i, j)] (23)

[0113] Calculating the second approximate solar azimuth using equal field-of-view image patches

[0114]

[0115] Where M and N are the width and height of the pixel position with a value of 1 in the equal field of view image block.

[0116] Thus, all measured solar azimuth angles are expressed as follows:

[0117]

[0118] according to The optimal estimate of the solar azimuth angle was calculated. C is to make The quantity to be determined when the minimum value is reached;

[0119] The horizontal attitude angle is calculated using the inertial navigation system. Let the absolute heading angle be ψ. The absolute heading angle ψ is calculated based on all measured solar azimuth angles and their optimal estimates.

[0120]

[0121] in:

[0122] in, Represents the solar azimuth angle in the carrier coordinate system.

[0123] To verify the effectiveness of this method, the following comparative experiment was conducted. The polarization compass was constructed based on a LUCID PHX050S-P / Q polarization camera and a TX2 sensor. Figure 2 This is a diagram showing the orientation results of this method at a 360° heading angle. (See diagram below.) Figure 3This image shows the orientation results of this method under tilt conditions. The methods used for comparison include: fitted meridian (fit), Zen-ROI region, image principal point region (SAO-PRP), separation point region (SAO-SEP), and the solar azimuth region (SAO) proposed in this embodiment. Reference values ​​are provided by the inertial navigation system (IMU), and tilt angles are illustrated using pitch and roll angles. Figure 2 , Figure 3 It can be seen that the method has the highest accuracy under 360° heading angle and tilt conditions, and exhibits good robustness under different conditions.

[0124] The above description is only a preferred 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 within the protection scope of the present invention.

Claims

1. A method for directional orientation of atmospheric polarized light using a solar azimuth region, characterized in that, Includes the following steps: Step 1: Use an image polarimeter to detect four-channel polarization images, and calculate the polarization angle based on the four-channel polarization images. Step 2: Obtain the solar altitude angle and solar azimuth angle, and obtain the observation point altitude angle and observation point azimuth angle. Using the pixel coordinate system theoretical polarization angle model, convert the polarization angle in the incident light coordinate system to the polarization angle in the pixel coordinate system, and establish a theoretical model of the polarization angle. Ignoring measurement noise, the obtained polarization angle is the polarization angle calculated from the four-channel polarization image. The polarization angle is obtained through the theoretical model of the polarization angle. With solar azimuth Relationships; Step 3: Use the region growing model to adjust the polarization angle. The solar azimuth region is extracted, and the initial growth point is set as the image principal point. Region growth is performed on the current polarization angle image at the initial growth point, and the grown binary image is calculated. The grown binary image is filtered using image morphology operations, the centroid of the filtered binary image is calculated, the seed line is updated according to the centroid, and the updated seed line is used as the starting growth point for the solar azimuth region extraction of the next frame of polarization angle image. Step 4: Calculate the first approximate solar azimuth angle based on the updated seed line. The binary image of the sun is divided into equal-field-of-view image blocks along the sun's direction. The second approximate solar azimuth angle is then calculated using these equal-field-of-view image blocks. All measurements of solar azimuth angle Represented as: according to The optimal estimate of the solar azimuth angle was calculated. C is to make The quantity to be determined when the minimum value is reached; The horizontal attitude angle is calculated using the inertial navigation system. Let the absolute heading angle be ψ. The absolute heading angle ψ is calculated based on all measured solar azimuth angles and their optimal estimates. in, Represents the solar azimuth angle in the carrier coordinate system.

2. The atmospheric polarization light orientation method using the solar azimuth region according to claim 1, characterized in that, In step one, let the polarization angle be... The formula for calculating the polarization angle is: Where I(*) represents the polarization intensity in the * direction.

3. The atmospheric polarization light orientation method using the solar azimuth region according to claim 1 or 2, characterized in that, In step two, the polarization angle in the incident light coordinate system is expressed as: Among them, h s A s h represents the solar altitude angle and solar azimuth angle. p With A p The elevation angle and azimuth angle of the observation point; Transform the polarization angle from the incident light coordinate system to the pixel coordinate system, then the polarization angle in the pixel coordinate system is... It can be represented as: In the formula: in: In the formula, sc is the field of view angle, rl is the resolution, (x p y p () represents the pixel position in the pixel coordinate system. These are the coordinates of the principal point of the image in the pixel coordinate system. right Simplifying, we get: in: Taking the observation direction as the direction of the sun, the above formula simplifies to: Introducing the three-axis attitude angle, similarly, we obtain the polarization angle in the pixel coordinate system: in: In the formula, the superscript b represents the carrier coordinate system; Similarly, simplifying the above equation along the direction of the sun: in, This represents the solar azimuth angle in the carrier coordinate system.

4. The atmospheric polarization light orientation method using the solar azimuth region according to claim 1 or 2, characterized in that, In step three, the grown binary image B AOP The calculation method for (i, j) is as follows: Where i and j are the width and height of a pixel, and RG represents the region growing model.

5. The atmospheric polarization light orientation method using the solar azimuth region according to claim 4, characterized in that, In step three, the grown binary image is filtered using image morphology shrinkage and clearing operations. The filtered binary image B SR (i, j) is represented as: B SR (i,j)=C(S(B AOP (i,j))) (18)。 6. The atmospheric polarization light orientation method using the solar azimuth region according to claim 5, characterized in that, In step three, the centroid calculation method for the filtered binary image is as follows: in: x s y s Let m represent the x and y coordinates of the centroid, respectively. pq Represents the relationship between the values ​​of p and q and B. SR The calculated values ​​of (i, j), p, q ∈ 0, 1, m 10 This represents the calculated value when p is 1 and q is 0; m 01 The calculated value represents p = 0 and q = 1; m 00 This represents the calculated value when p is 0 and q is 0.

7. The atmospheric polarization light orientation method using the solar azimuth region according to claim 6, characterized in that, In step three, the updated seed line is represented as follows: Where X and Y are the independent and dependent variables in the linear equation, These are the coordinates of the principal point of the image in the pixel coordinate system.

8. The atmospheric polarization light orientation method using the solar azimuth region according to claim 7, characterized in that, In step four, the first approximate solar azimuth angle The calculation method is as follows:

9. The atmospheric polarization light orientation method using the solar azimuth region according to claim 4, characterized in that, In step four, the binary image of the sun direction is divided into n equal field-of-view image blocks along the sun direction and represented as B. AOP,1 (i, j), B AOP，2 (i, j), ..., B AOP,n (i, j), then the set of equal field-of-view image patches B FOV (i, j) is represented as: B FOV (i,j)=[B AOP,1 (i,j),B AOP,2 (i,j),…,B AOP,n (i,j)] (23) Calculating the second approximate solar azimuth using equal field-of-view image patches Where M and N are the width and height of the pixel position with a value of 1 in the equal field of view image block.

Images