Visible light positioning method based on geometrical characteristics of lamp

Abstract

The invention discloses a visible light positioning method based on geometrical characteristics of a lamp, and belongs to the field of visible light communication in the technical field of communication. According to the method, firstly, a camera is used for capturing geometric features of a lamp, then, a normal vector of the lamp in an auxiliary coordinate system is calculated, in the process, the calculated normal vector has the problem of multiple solutions, the camera virtual offset technology is adopted, the uncertainty of the multiple solutions is effectively eliminated, and finally, through a single-view geometric method, the normal vector of the lamp in an auxiliary coordinate system is calculated. And the posture and the position of the camera are accurately determined, so that an accurate and efficient positioning technology is realized. Simulation results show that the positioning errors of the sample points are within 10cm.

Description

Technical Field

[0001] The present invention relates to the field of visible light communication, and in particular to a visible light positioning method based on geometric features of lamps. Background Art

[0002] With the rapid development of mobile Internet and smart devices, the demand for location tracking and navigation has increased significantly. As the mainstream positioning system, the Global Positioning System (GPS) is widely used in aviation, vehicles and portable devices, but it performs poorly in indoor environments and is difficult to meet the needs of precise positioning. Therefore, indoor positioning technology has become an important research direction. Existing indoor positioning technologies such as WiFi, Bluetooth, radio frequency identification (RFID), ultra-wideband (UWB) and ultrasonic positioning (Ultrasonic) have limitations in positioning accuracy, coverage, cost and environmental adaptability, especially in high-precision demand scenarios. It is difficult to provide sufficient accuracy and reliability. Indoor positioning technology based on visible light communication has attracted widespread attention due to its unique advantages.

[0003] In the current research field, camera-based visible light positioning (VLP) methods often treat lamps as point sources to achieve precise positioning, thereby ignoring their geometric features. Although camera-based visible light positioning methods perform well under specific conditions, their limitations in field of view (FOV) may affect their application performance in a wider range of scenarios. In addition, in order to implement camera-based visible light positioning methods, expensive modifications to existing lighting facilities may be required, which becomes a major economic burden when deployed on a large scale.

[0004] On the other hand, some single-lamp positioning methods that consider the geometric characteristics of lamps often require the introduction of additional mobile models or auxiliary sensors to assist in positioning. This approach undoubtedly increases the complexity of the positioning system, increases the cost, and limits its applicability and flexibility in specific application scenarios. Summary of the invention

[0005] In view of the shortcomings in the prior art, the present invention provides a visible light positioning method based on the geometric characteristics of lamps, which achieves high-precision indoor positioning without the need for additional sensors or mobile models, and enhances the accuracy and robustness of the positioning system.

[0006] The present invention achieves the above technical objectives through the following technical means.

[0007] A visible light positioning method based on the geometric characteristics of lamps:

[0008] The camera captures the high-speed flashing light source signal of the lamp into a stripe pattern, demodulates the stripe pattern to extract the lamp ID, and searches the corresponding lamp position coordinates through the system database; the lamp is a round lamp;

[0009] The geometric feature information of the lamp is obtained. Based on the geometric features of the lamp, the general equation of the camera imaging surface ellipse is fitted. The relationship between the camera coordinate system and the image coordinate system is used to obtain the side surface equation of the elliptical cone formed by the camera and the imaging surface ellipse. The auxiliary coordinate system O is introduced. a -x a y a z a , simplifying the side surface equation of the elliptical cone, thereby representing the normal vector of the lamp in the auxiliary coordinate system;

[0010] The virtual offset method of the camera is used to discard multiple solutions of the normal vector of the lamp in the auxiliary coordinate system, and the unique normal vector of the lamp in the auxiliary coordinate system is determined;

[0011] Determine the normal vector of the lamp in the camera coordinate system based on the conversion relationship between the camera coordinate system and the auxiliary coordinate system and the unique normal vector of the lamp in the auxiliary coordinate system. Obtain the rotation angle α of the world coordinate system around the x-axis and the rotation angle β around the y-axis based on the conversion relationship between the normal vector of the lamp in the world coordinate system and the normal vector of the lamp in the camera coordinate system. Obtain the rotation angle γ of the world coordinate system around the z-axis based on the conversion relationship between the direction vector of the center of the lamp G and the marked point M and the unit direction vector of the point G and the point M in the world coordinate system in the camera coordinate system.

[0012] Based on the rotation angles α, β, and γ, obtaining the posture of the camera;

[0013] The position coordinates of the camera are obtained from the posture of the camera, the coordinates of the center G of the lamp and the marking point M in the world coordinate system, and the coordinates of the center G of the lamp and the marking point M in the camera coordinate system.

[0014] Furthermore, the side surface equation of the elliptical cone is:

[0015] A 2 (x c ) 2 +Bf 2 x c y c +Cf 2 (y c ) 2 +Dfx c z c +Efy c z c +(z c ) 2 =0

[0016] Among them, A, B, C, D, E are constant coefficients, and f is the image coordinate system O i -x i y i The coordinate origin and camera coordinate system O c -x c y c z c The distance between the coordinate origins.

[0017] Furthermore, the side surface equation of the elliptical cone is simplified as follows:

[0018] λ1(x a ) 2 +λ2(y a ) 2 +λ3(z a ) 2 =0

[0019] Among them, λ1, λ2, λ3 are characteristic matrices The characteristic value of .

[0020] Furthermore, the normal vector of the lamp in the auxiliary coordinate system for:

[0021]

[0022] Where T stands for transpose.

[0023] Furthermore, the normal vector of the lamp in the auxiliary coordinate system is in, Align the camera along the normal vector Virtually offset by m units, the updated camera coordinates are expressed as:

[0024]

[0025] The normal vector corresponding to the minimum distance between the updated camera coordinates and the center G of the lamp is the only normal vector of the lamp in the auxiliary coordinate system, that is:

[0026]

[0027] in, Indicates the coordinates of the center G of the lamp in the auxiliary coordinate system.

[0028] Furthermore, the unique normal vector of the lamp in the auxiliary coordinate system is According to the expression of auxiliary plane z a =k * x a +b, the plane expression of the lamp is:

[0029] z a =k * x a +b LED

[0030] in, k * k represents the minimum distance between the updated camera coordinates and the center G of the lamp. i , b represents a constant, They respectively represent the coordinates of the lamp diameter end points E1 and E2 in the auxiliary coordinate system.

[0031] Furthermore, the rotation angle α of the world coordinate system around the x-axis and the rotation angle β around the y-axis are:

[0032]

[0033] Among them, a1, a2, and a3 are the normal vectors of the lamp in the camera coordinate system. Elements in

[0034] The conversion relationship between the normal vector of the lamp in the world coordinate system and the normal vector of the lamp in the camera coordinate system is:

[0035]

[0036] in, is the normal vector of the lamp in the world coordinate system, is the camera's posture, matrix matrix matrix

[0037] Furthermore, the rotation angle γ of the world coordinate system around the z-axis is:

[0038]

[0039] Among them, b1, b2, and b3 are the elements in the direction vector of the center G of the lamp and the marking point M;

[0040] The conversion relationship between the direction vector of the center G and the marking point M of the lamp and the unit direction vector of the point G and the point M in the world coordinate system is:

[0041]

[0042] Furthermore, the position coordinates of the camera are obtained from the posture of the camera, the coordinates of the center G of the lamp and the mark point M in the world coordinate system, and the coordinates of the center G of the lamp and the mark point M in the camera coordinate system, specifically:

[0043] The coordinates of the center G of the lamp and the mark point M in the world coordinate system And the coordinates of the center G of the lamp and the marker point M in the camera coordinate system Substitute into the conversion formula Get the position coordinates of a camera respectively Among them, k is an arbitrary point, including the center G of the lamp and the marking point M.

[0044] Furthermore, Take the average value as the final position coordinate of the camera, that is:

[0045] The beneficial effects of the present invention are:

[0046] (1) The present invention obtains the position coordinates of the camera from the posture of the camera, the coordinates of the center G of the lamp and the mark point M in the world coordinate system, and the coordinates of the center G of the lamp and the mark point M in the camera coordinate system, thereby realizing visible light indoor positioning without introducing additional sensors and mobile models.

[0047] (2) The present invention utilizes the geometric features of the lamp to fit the general equation of the camera imaging surface ellipse, further obtains the side surface equation of the elliptical cone formed by the camera and the imaging surface ellipse, introduces an auxiliary coordinate system to simplify the side surface equation of the elliptical cone, thereby characterizing the normal vector of the lamp in the auxiliary coordinate system, and then utilizes the virtual offset method of the camera to discard the multiple solutions of the normal vector of the lamp in the auxiliary coordinate system, and determine the unique normal vector of the lamp in the auxiliary coordinate system. The normal vector of the lamp in the camera coordinate system is determined by the unique normal vector of the lamp in the auxiliary coordinate system, and then the rotation angles around the x-axis and the y-axis are obtained. The posture of the camera is obtained in combination with the rotation angle around the z-axis, thereby improving the robustness of the positioning method.

[0048] (3) The general equation of the camera imaging surface ellipse and the side surface equation of the elliptical cone of the present invention both adopt mathematical models, and high-precision positioning is achieved through accurate mathematical models and calculation methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 This is a model diagram of the visible light positioning system of the present invention;

[0050] Figure 2 A schematic diagram of virtual movement of the camera in the auxiliary coordinate system of the present invention;

[0051] Figure 3 The elliptical cone of the present invention is a Oz a Projection drawing on a plane;

[0052] Figure 4A comparison diagram of the cumulative distribution functions of positioning errors of the three positioning methods described in the present invention;

[0053] Figure 5 This is a comparison diagram of the impact of image noise on three positioning methods described in the present invention;

[0054] Figure 6 This is a comparison diagram of the effects of different heights on three positioning methods described in the present invention;

[0055] Figure 7 This is a flow chart of the visible light positioning method based on the geometric features of lamps described in the present invention. DETAILED DESCRIPTION

[0056] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments, but the protection scope of the present invention is not limited thereto.

[0057] Assign a unique ID to each lamp, and store the position coordinates of each lamp in the system database. At the same time, embed the ID information through modulation to facilitate camera recognition; use the rolling shutter effect of the camera to capture the high-speed flashing lamp light source signal into a stripe pattern, demodulate the stripe pattern to extract the lamp ID, and search the corresponding lamp position coordinates through the system database. The camera captures the geometric feature information of the lamp and calculates the normal vector of the lamp; in order to solve the multi-solution problem generated in the process of calculating the normal vector of the lamp, the present invention effectively solves the multi-solution problem through the virtual offset of the camera; finally, use the single view geometry method to accurately determine the attitude and position of the camera, see Figure 7 The method of the present invention is not limited to a single circular lamp, but is also applicable to multiple lamps and lamps of other geometric shapes. The present invention takes a single circular lamp as an example, and the specific implementation includes the following steps and is performed in the order described:

[0058] Step (1), using the geometric features of the lamp captured by the camera, calculate the normal vector of the lamp in the auxiliary coordinate system. In this process, the calculated normal vector has the problem of multiple solutions;

[0059] Figure 1 The VLP system shown in the figure consists of a circular lamp and a camera. When the camera is tilted, the camera imaging surface captures an ellipse, which is denoted as Π e In the elliptical cone formed by the camera and the imaging surface ellipse, there is only one ellipse parallel to the circular lamp plane, and this ellipse is denoted as Π * , Figure 1 Where M is the mark point on the lamp. The system operates in five different coordinate systems, namely the world coordinate system ( w -x w y w z w ), camera coordinate system (O c -xc y c z c ), image coordinate system (O i -x i y i ), pixel coordinate system (O p -u p v p ) and the auxiliary coordinate system (O a -x a y a z a ), the x of the image coordinate system i ,y i Axis and camera coordinate system x c ,y c The axes are parallel, and the distance between the origins of the two coordinate systems is the focal length f. Given the radius r0 of the lamp, the world coordinates of the center G of the lamp and the mark point M are

[0060] The camera takes a picture of the lamp and uses image processing technology (such as edge detection) to extract the geometric feature information of the lamp.

[0061] Based on the geometric characteristics of the lamp, the least squares method is used to fit the camera imaging surface ellipse Π e The general equation of A(x i ) 2 +Bx i y i +C(y i ) 2 +Dx i +Ey i +1=0, where A, B, C, D, and E are constant coefficients.

[0062] Using the relationship between the camera coordinate system and the image coordinate system, the side surface equation of the elliptical cone is obtained: Af 2 (x c ) 2 +Bf 2 x c y c +Cf 2 (y c ) 2 +Dfx c z c +Efy c z c +(z c ) 2 =0.

[0063] Introducing the auxiliary coordinate system (O a -x a y a z a), simplify the side surface equation of the elliptical cone:

[0064] λ1(x a ) 2 +λ2(y a ) 2 +λ3(z a ) 2 =0

[0065] Among them, λ1, λ2, λ3 are characteristic matrices The transformation relationship between the camera coordinate system and the auxiliary coordinate system is is the matrix consisting of eigenvalues ​​and eigenvectors.

[0066] Plane Π * Normal vector in the auxiliary coordinate system, that is, the normal vector of the lamp in the auxiliary coordinate system It is expressed as:

[0067]

[0068] Where T stands for transpose.

[0069] The normal vector of the lamp under ACS From the expression of , we can see that there are two solutions.

[0070] Step (2), for the existing multiple solution problem, the virtual offset method of the camera is used to discard the solution.

[0071] l1′ is the intersection line between the ceiling and the imaging surface. l1 is translated q units and passes through the center of the circle. The equation of the straight line is:

[0072]

[0073] in, b1=0;

[0074] The intersection points of the straight line equation l1′ and the ellipse of the imaging plane are E1′ and E2′. There must be a diameter in the lamp that is parallel to E1′E2′. The two endpoints of this diameter are E1 and E2. Using the optical center O of the camera c , and construct ΔE1O with points E1, E2, E1′, E′2 c E2 and ΔE1′O c E2′. Using ΔE1O c E2 and ΔE1′O c E2′, and obtain the coordinates of the center G of the lamp in the camera coordinate system.

[0075] The normal vector of the lamp in the auxiliary coordinate system is in, The present invention proposes a camera virtual offset method to make the camera move along the normal vector Virtually offset by m units, the updated camera coordinates are expressed as:

[0076]

[0077] The distance between the updated camera coordinates and the center G of the lamp is used to determine the unique normal vector of the lamp in the auxiliary coordinate system. Specifically, the normal vector corresponding to the minimum distance between the updated camera coordinates and the center G of the lamp is the unique normal vector of the lamp in the auxiliary coordinate system, that is:

[0078]

[0079] in, Indicates the coordinates of the center G of the lamp in the auxiliary coordinate system.

[0080] Step (3), based on the single view geometry method, estimate the camera's pose and position.

[0081] Figure 3 Describes the elliptical cone at x a Oz a The projection above. The unique normal vector of the lamp in the auxiliary coordinate system is According to the expression of auxiliary plane z a =k * x a +b, write the plane expression of the lamp, that is:

[0082] z a =k * x a +b LED ;

[0083] in, k * k represents the minimum distance between the updated camera coordinates and the center G of the lamp. i , b represents a constant, They respectively represent the coordinates of the intersection points E1 and E2 in the auxiliary coordinate system.

[0084] The projection point M′ of the marker point M and the optical center O of the camera a The connecting line segment is M′O a , using line segment M′O a and the plane expression of the lamp to obtain the coordinates of the marker point M in the auxiliary coordinate system Right now:

[0085]

[0086] in, They respectively represent the coordinates of the projection point M′ of the marked point M in the auxiliary coordinate system.

[0087] Using the relational Obtain the coordinates of the marker point M in the camera coordinate system

[0088] Based on the conversion relationship between the camera coordinate system and the auxiliary coordinate system, the unique normal vector of the lamp in the auxiliary coordinate system Determine the normal vector of the lamp in the camera coordinate system: According to the normal vector of the lamp in the world coordinate system and The conversion relationship between Get the rotation angle α of the world coordinate system around the x-axis and the rotation angle β around the y-axis, that is:

[0089]

[0090] Among them, the matrix matrix matrix

[0091] In the camera coordinate system, record the center G of the lamp and the direction vector of the marker point M use and The conversion relationship between Get the rotation angle γ of the world coordinate system around the z-axis, that is:

[0092]

[0093] in, It is the unit direction vector of the center G of the lamp and the marker point M in the world coordinate system.

[0094] Based on the rotation angles α, β, and γ, substitute into the corresponding matrix to obtain the camera posture, that is,

[0095] The coordinates of the center G of the lamp and the mark point M in the world coordinate system The coordinates of the center G and the marker M of the lamp in the camera coordinate system Substitute the conversion formulas for both Get the position coordinates of a camera respectively Among them, k is an arbitrary point, including the center G of the lamp and the marking point M.

[0096] In order to further reduce the error, Take the average value as the final position coordinate of the camera, that is:

[0097]

[0098] In this embodiment, a circular lamp with a radius of 10 cm is selected, and the performance comparison between the visible light positioning method (i.e., V-PCF) proposed in the present invention and the traditional IMU-assisted positioning method (i.e., IMU-VLP) and the PnP method is shown in FIG. Figure 4 , Figure 5 and Figure 6 shown.

[0099] The cumulative distribution functions of the positioning errors of the three positioning methods are as follows: Figure 4 In a room of 8m×8m×3m, the positioning accuracy of the method proposed in the present invention is better than that of the other two algorithms, and more than 97% of the positioning errors are less than 10cm, which is significantly higher than 68% of the PnP method and 33% of the IMU-VLP method.

[0100] Figure 5 The effect of image noise on the three positioning methods. Image noise is modeled in the image coordinate system with a mean of 0 and a variance of 10. -3 -50 -3 In a room of 8m×8m×3m, the positioning error of the present invention is only increased by 3cm, which is significantly increased compared with the PnP method and the IMU-VLP method. The method proposed in the present invention has better robustness to noise.

[0101] Figure 6 The influence of different heights on the three positioning methods. The height of the simulation environment increases from 3m to 11cm. In a 3m×3m room, the positioning error of the method proposed in the present invention is always kept below 8cm, and its performance is better than the other two methods, which further proves that the method proposed in the present invention has good adaptability and robustness to height changes.

[0102] In summary, the present invention uses the geometric feature information of lamps to achieve high-precision indoor positioning without the need for additional sensors or mobile models. Compared with other positioning methods, the method proposed in the present invention has higher positioning accuracy and stronger adaptability to the environment. The simulation results show that the positioning error of the sample points is within 10 cm.

[0103] The embodiments are preferred implementations of the present invention, but the present invention is not limited to the above-mentioned implementations. Any obvious improvements, substitutions or modifications that can be made by those skilled in the art without departing from the essential content of the present invention belong to the protection scope of the present invention.

Claims

1. A visible light positioning method based on the geometric characteristics of lamps, characterized in that: The camera captures the high-speed flashing light source signal of the lamp into a stripe pattern, demodulates the stripe pattern to extract the lamp ID, and searches the corresponding lamp position coordinates through the system database; the lamp is a round lamp; The geometric feature information of the lamp is obtained. Based on the geometric features of the lamp, the general equation of the camera imaging surface ellipse is fitted. The relationship between the camera coordinate system and the image coordinate system is used to obtain the side surface equation of the elliptical cone formed by the camera and the imaging surface ellipse. The auxiliary coordinate system O is introduced. a -x a y a z a , simplifying the side surface equation of the elliptical cone, thereby representing the normal vector of the lamp in the auxiliary coordinate system; The virtual offset method of the camera is used to discard multiple solutions of the normal vector of the lamp in the auxiliary coordinate system, and the unique normal vector of the lamp in the auxiliary coordinate system is determined; Determine the normal vector of the lamp in the camera coordinate system based on the conversion relationship between the camera coordinate system and the auxiliary coordinate system and the unique normal vector of the lamp in the auxiliary coordinate system. Obtain the rotation angle α of the world coordinate system around the x-axis and the rotation angle β around the y-axis based on the conversion relationship between the normal vector of the lamp in the world coordinate system and the normal vector of the lamp in the camera coordinate system. Obtain the rotation angle γ of the world coordinate system around the z-axis based on the conversion relationship between the direction vector of the center of the lamp G and the marked point M and the unit direction vector of the point G and the point M in the world coordinate system in the camera coordinate system. Based on the rotation angles α, β, γ, obtaining the posture of the camera; The position coordinates of the camera are obtained from the posture of the camera, the coordinates of the center G of the lamp and the marking point M in the world coordinate system, and the coordinates of the center G of the lamp and the marking point M in the camera coordinate system.

2. The visible light positioning method based on the geometric characteristics of lamps according to claim 1, characterized in that: The side surface equation of the elliptical cone is: Af 2 (x c ) 2 +Bf 2 x c y c +Cf 2 (y c ) 2 +Dfx c z c +Efy c z c +(z c ) 2 =0 Among them, A, B, C, D, E are constant coefficients, and f is the image coordinate system O i -x i y i The coordinate origin and camera coordinate system O c -x c y c z c The distance between the coordinate origins.

3. The visible light positioning method based on the geometric characteristics of lamps according to claim 2, characterized in that: Simplify the side surface equation of the elliptical cone: λ1(x a ) 2 +λ2(y a ) 2 +λ3(z a ) 2 =0 Among them, λ1, λ2, λ3 are characteristic matrices The characteristic value of .

4. The visible light positioning method based on the geometric characteristics of lamps according to claim 3, characterized in that: Normal vector of the lamp in the auxiliary coordinate system for: Where T stands for transpose.

5. The visible light positioning method based on the geometric characteristics of lamps according to claim 4, characterized in that: The normal vector of the lamp in the auxiliary coordinate system is i=1,2, where Align the camera along the normal vector Virtually offset by m units, the updated camera coordinates are expressed as: The normal vector corresponding to the minimum distance between the updated camera coordinates and the center G of the lamp is the only normal vector of the lamp in the auxiliary coordinate system, that is: in, Indicates the coordinates of the center G of the lamp in the auxiliary coordinate system.

6. The visible light positioning method based on the geometric characteristics of lamps according to claim 5, characterized in that: The unique normal vector of the lamp in the auxiliary coordinate system is According to the expression of auxiliary plane z a =k * x a +b, the plane expression of the lamp is: z a =k * x a +b LED in, k * k represents the minimum distance between the updated camera coordinates and the center G of the lamp. i , b represents a constant, They respectively represent the coordinates of the lamp diameter end points E1 and E2 in the auxiliary coordinate system.

7. The visible light positioning method based on the geometric characteristics of lamps according to claim 6, characterized in that: The rotation angle α of the world coordinate system around the x-axis and the rotation angle β around the y-axis are: Among them, a1, a2, and a3 are the normal vectors of the lamp in the camera coordinate system. Elements in The conversion relationship between the normal vector of the lamp in the world coordinate system and the normal vector of the lamp in the camera coordinate system is: in, is the normal vector of the lamp in the world coordinate system, is the camera's posture, matrix matrix matrix 8. The visible light positioning method based on the geometric characteristics of lamps according to claim 7, characterized in that: The rotation angle γ of the world coordinate system around the z-axis is: Among them, b1, b2, and b3 are the elements in the direction vector of the center G of the lamp and the marking point M; The conversion relationship between the direction vector of the center G and the marking point M of the lamp and the unit direction vector of the point G and the point M in the world coordinate system is:

9. The visible light positioning method based on the geometric characteristics of lamps according to claim 1, characterized in that: The position coordinates of the camera are obtained from the posture of the camera, the coordinates of the center G of the lamp and the mark point M in the world coordinate system, and the coordinates of the center G of the lamp and the mark point M in the camera coordinate system, specifically: The coordinates of the center G of the lamp and the mark point M in the world coordinate system And the coordinates of the center G of the lamp and the marker point M in the camera coordinate system Substitute into the conversion formula Get the position coordinates of a camera respectively Among them, k is an arbitrary point, including the center G of the lamp and the marking point M.

10. The visible light positioning method based on the geometric characteristics of lamps according to claim 9, characterized in that: right Take the average value as the final position coordinate of the camera, that is:

Images