A Pixel Coordinate Correction Method for Electric Vehicle Charging Holes Based on Affine Transformation

Abstract

The present invention discloses a method for correcting the pixel coordinates of an electric vehicle charging hole based on affine transformation. When the camera is directly facing the charging port, the projective transformation is approximated as an affine transformation. The original coordinates of the center point of the charging hole are obtained through a target detection method. According to the characteristics of maintaining collinearity and distance ratio before and after affine transformation, the pixel coordinates are corrected with reference to the geometric dimensions between the charging holes. Successively, parallel line fitting, correction in the direction perpendicular to the parallel lines, and correction in the parallel line direction by introducing a correction amount are carried out. Finally, an expression of the corrected pixel coordinates of the charging hole after affine transformation with respect to the correction amount is obtained. Finally, by constructing an error function to minimize the error between the corrected pixel coordinates and the original pixel coordinates, the correction amount is solved, and thus the center pixel coordinate value of the charging hole after affine transformation is obtained. The present invention can significantly improve the positioning accuracy of the charging port.

Description

Technical Field

[0001] The present invention specifically relates to a method for correcting pixel coordinates of an electric vehicle charging hole based on affine transformation. Background Art

[0002] With the continuous popularization of electric vehicles, people's demand for vehicle charging is increasing day by day. The automatic charging technology of electric vehicles can optimize the user charging experience and avoid the danger of personnel using electricity, which is one of the current research hotspots. At present, the automatic charging solutions are mainly divided into wireless automatic charging technology and cable conduction connection charging based on an automatic charging robot. The high voltage, high power required for electric vehicle charging, as well as equipment modification and maintenance, bring additional challenges to the cost and safety of wireless automatic charging; at the same time, the low energy transfer efficiency makes it a huge problem to achieve fast charging. The cable conduction connection charging method based on an automatic charging robot avoids the above disadvantages, has low cost, high charging efficiency, and almost no ionization radiation, and is safer and more reliable. The identification and positioning of the electric vehicle charging interface are the key to realizing cable conduction connection automatic charging. The identification and positioning accuracy of the charging interface determine whether the charging gun can successfully dock with the charging interface, and determine the reliability of the automatic charging system. However, it is still a challenging research problem to improve the positioning accuracy of the charging interface in a complex background and lighting environment.

[0003] In the current charging interface positioning solutions based on monocular vision, most solutions obtain the charging interface features through traditional object detection algorithms or object detection algorithms based on deep learning, obtain the center pixel coordinates of the charging holes, and then use the pnp (perspective-n-point) algorithm to locate the pose information of the charging interface. The accuracy of the center pixel coordinates of the charging holes directly affects the positioning accuracy of the charging interface. And it is a great challenge to ensure the positioning accuracy of the charging interface in a complex light environment and complex scene. The positioning accuracy of the charging interface needs to be improved to ensure the plugging and unplugging success rate of the automatic charging system. Summary of the Invention

[0004] Aiming at the deficiencies in the prior art, the present invention provides a method for correcting pixel coordinates of an electric vehicle charging hole based on affine transformation, which is used to improve the positioning accuracy of the charging interface.

[0005] The present invention realizes the above technical purpose through the following technical means.

[0006] A method for correcting pixel coordinates of an electric vehicle charging hole based on affine transformation:

[0007] Assume that there are n charging holes in the charging interface. When the camera is facing the charging interface directly, the image of the charging interface collected by the camera is transmitted to the computer, and the original coordinates (u i , vi ), determine a set of parallel lines according to the coordinates (u i , v i ), where i = 1, ..., n;

[0008] Determine the projection points (u i , v i ) on the corresponding parallel lines as (u i ', v i ');

[0009] Translate all the projection points and the parallel lines where they are located until the translation reference point (u j ', v j ) is moved to the origin of the camera pixel coordinate system. Rotate the parallel lines horizontally with the origin of the camera pixel coordinate system as the rotation center. The rotation points corresponding to the projection points are (u i ”, v i ”); Among them, the translation reference point (u j ', v j ) can be any point selected from (u i ', v i ');

[0010] According to the affine transformation characteristics, correct the rotation points (u i ”, v i ”) to obtain the expression of (u i ”', v i ”');

[0011] With the origin of the pixel coordinate system as the rotation center, rotate the parallel lines in the opposite direction and translate all the corrected points until (u j ”', v j ) coincides with (u j ', v j '), and obtain the expression of the corrected points (u i ””, v i ””);

[0012] Construct the error function between (u i ””, v i ) and (u i , v i ). Take the derivative of the error function with respect to the correction amount and set it to 0 to solve for the correction amount, and then determine the expression of (ui”', vi”'). Finally, obtain the pixel coordinate values (u i ””, v i ””) of the charging hole after affine transformation correction.

[0013] A further technical solution is to determine a set of parallel lines according to the coordinates (u i , vi ) The process of determining a set of parallel lines is as follows:

[0014] According to the geometric relationship between the charging interface holes and the corresponding (u i , v i ) Fit p mutually parallel lines:

[0015] l1: v = ku + b + Δb

[0016] l2: v = ku + b

[0017] l3: v = ku + b + λ1Δb

[0018] ……

[0019] l p : v = ku + b + λ p-2 Δb

[0020] Where: u is the independent variable, v is the dependent variable, and k, b, and Δb are unknowns; λ1,..., λ p-2 Used to describe the relative relationship between the corresponding parallel lines and the parallel line l1, and its value is determined according to the geometric relationship between the charging holes;

[0021] Assume that the parallel lines l1, l2,..., l p Are respectively fitted by n1, n2,..., n p Points, establish the objective function D about k, b, and Δb 2 :

[0022]

[0023] The partial derivatives of the objective function with respect to k, b, and Δb are taken respectively, and the expression values are set to 0, and then the unknowns are solved according to λ1,..., λ p-2 、(u i , v i ) to determine p parallel lines.

[0024] A further technical solution is that the parallel lines are translated as a whole until (u j ', v j ) coincides with the origin of the camera pixel coordinate system, and the parallel lines are rotated to the horizontal with the origin of the pixel coordinate system as the rotation center, and the rotation angle β satisfies: β = arctan(k), where k is the slope of the parallel lines.

[0025] A further technical solution is that the (u i ”, v i ) and (u i ', v i ) satisfy:

[0026]

[0027] A further technical solution is to correct the rotation point (u i ”, v i ”), introduce a correction amount to obtain the expression of (u i ”', v i ”'), specifically:

[0028] Introduce k correction amounts du1,..., du k , and according to the positional relationship of each charging hole in the actual charging interface, correct (u i ”, v i ”) to obtain the expression of (u i ”', v i ”').

[0029] A further technical solution, the expression of (u i ””, v i ””) is:

[0030]

[0031] A further technical solution, the error function is

[0032] The beneficial effect of the present invention is that when the camera is facing the charging interface directly, the original coordinates (u i , v i ) of the center of the charging hole of the charging interface in the camera pixel coordinate system are obtained. At this time, the projective transformation can be approximated as an affine transformation. According to the fact that collinearity and distance ratio are maintained before and after the affine transformation, correct the rotation point (u i ', v i ') of the pixel coordinate projection point (u i ”, v i ”) to obtain the expression of (u i ”', v i ”'); with the origin of the pixel coordinate system as the rotation center, rotate the parallel lines in the opposite direction and translate all the corrected points until (u j ”', v j ”) coincides with (u j ', v j ') to obtain the expression of the transformed point (u i ””, v i ””); finally, by constructing an error function to minimize the error between the corrected pixel coordinates and the original pixel coordinates, solve the correction amount, and finally obtain the pixel coordinate value of the charging hole after affine transformation correction. The pixel coordinate value of the charging hole after correction obtained by using the method of the present invention can significantly improve the positioning accuracy of the charging interface. Brief Description of the Drawings

[0033] Figure 1 This is a model diagram of the AC charging interface described in the present invention;

[0034] Figure 2 This is a diagram showing the establishment of the coordinate system of the charging interface and the positional relationship between the charging holes described in the present invention;

[0035] Figure 3 This is a diagram showing the positional relationship of the charging holes in the camera image plane described in the present invention. Detailed Embodiment

[0036] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, but the protection scope of the present invention is not limited thereto.

[0037] A method for correcting the pixel coordinates of the charging holes of an electric vehicle based on affine transformation is as follows:

[0038] (1) Assume that there are n charging holes in the charging interface. When the camera is facing the charging interface directly, the image of the charging interface collected by the camera is transmitted to the computer, and the original coordinates (u i , v i ) of the centers of the charging holes in the camera pixel coordinate system are determined through the target detection algorithm. According to the coordinates (u i , v i ), a set of parallel lines is determined, where i = 1,..., n;

[0039] The process of determining a set of parallel lines according to the coordinates (u i , v i ) is as follows:

[0040] According to the geometric relationship between the holes of the charging interface and the corresponding (u i , v i ), p mutually parallel lines are fitted:

[0041] l1: v = ku + b + Δb

[0042] l2: v = ku + b

[0043] l3: v = ku + b + λ1Δb

[0044] ……

[0045] l p : v = ku + b + λ p-2 Δb

[0046] Where: u is the independent variable, v is the dependent variable, k, b, and Δb are unknowns; λ1, ……, λ p-2 is used to describe the relative relationship between the corresponding parallel lines and the parallel line l1, and its value is determined according to the geometric relationship between the charging holes;

[0047] Let the parallel lines \(l_1, l_2, \cdots, l\) p be respectively fitted by \(n_1, n_2, \cdots, n\) p points, and establish the objective function \(D\) about \(k\), \(b\), \(\Delta b\) 2 :

[0048]

[0049] The partial derivatives of the said objective function with respect to \(k\), \(b\), \(\Delta b\) are taken respectively, and the expression values are set to 0. Then, according to \(\lambda_1, \cdots, \lambda\) p-2 , \((u\) i , \(v\) i ) the unknowns are solved, and then \(p\) parallel lines are determined.

[0050] (2) Determine the projection points \((u\) i , \(v\) i ) of the corresponding parallel lines at \((u\) i ', \(v\) i ').

[0051] (3) Translate all the projection points and the parallel lines where they are located simultaneously until the translation reference point \((u\) j , \(v\) j ) is moved to the origin of the camera pixel coordinate system, and the parallel lines are rotated horizontally with the origin of the camera pixel coordinate system as the rotation center. The rotation points corresponding to the projection points are \((u\) i ”, \(v\) i ”); among them, the translation reference point \((u\) j , \(v\) j ) can be any point selected from \((u\) i , \(v\) i ');

[0052] The \((u\) i ”, \(v\) i ) and \((u\) i , \(v\) i ) satisfy:

[0053]

[0054] (4) According to the affine transformation characteristics, correct the rotation points \((u\) i ”, \(v\) i ) to obtain the expression of \((u\) i ”', \(v\) i ”');

[0055] Correct the rotation points \((u\) i ”, \(v\) i ) and introduce the correction amount to obtain \((u\) i ”', \(v\)i The expression of “') is specifically as follows:

[0056] Introduce k correction amounts du1, …, du k , and according to the positional relationship of each charging hole in the actual charging interface, correct (u i ”, v i ”) to obtain the expression of (u i ”', v i ”').

[0057] (5) With the origin of the pixel coordinate system as the rotation center, rotate the parallel lines in the opposite direction and translate all the corrected points until (u j ”', v j ”) coincides with (u j ', v j ') to obtain the expression of the corrected point (u i ””, v i ””);

[0058] (u i ””, v i ””) The expression is:

[0059]

[0060] (6) Construct the error function between (u i ””, v i ””) and (u i , v i ). Take the derivative of the error function with respect to the correction amount and set it to 0 to solve for the correction amount, and then determine the expression of (ui”', vi”'). Finally, obtain the pixel coordinate values (u i ””, v i ””) after affine transformation correction; the error function is

[0061] The present invention takes the AC charging interface of an electric vehicle as the research object. As Figure 1 shown, the AC charging interface has a 7-charging-hole structure, and the end faces of the holes are on the same plane. Define the origin of the coordinate system of the charging interface at the center of the end face of the charging interface. The coordinate system and the coordinate relationship between the charging holes are as Figure 2 shown. Among them, L1, L2, and L3 are AC power supplies, N is the neutral line, PE is the ground protection, CC is the charging connection confirmation, and CP is the control guide. The charging holes are distinct in features, and the sizes are divided into two categories: the outer diameters of the charging holes CP and CC are 9.7 mm, and the inner diameters are 3.5 mm; the outer diameters of the charging holes L1, L2, L3, N, and PE are 14.2 mm, and the inner diameters are 6.5 mm.

[0062] An Affine Transformation is any transformation that preserves collinearity and the ratio of distances. It is a special projective transformation composed of a linear transformation and a translation. Collinearity means that collinear points remain on a straight line after an affine transformation. The invariance of the distance ratio means that the ratio of line segments on parallel lines remains unchanged after an affine transformation. Affine transformations have strong practical applications, such as image registration, image rectification, texture rectification, and creating panoramic images, etc.

[0063] As Figure 3 shown, the origin of the camera pixel coordinate system u-v is O, and the center points of the charging holes are represented by points 1-7. When the camera is facing the charging interface directly, the image of the charging interface collected by the camera is transmitted to the computer in real time. The computer uses existing object detection algorithms to obtain the original coordinates of the 7 center points of the charging holes in the camera pixel coordinate system as (u i , v i )(i = 1,..., 7), which are represented by diamonds in Figure 3 . At this time, the origin of the coordinate system of the charging interface is located at the center of the camera image plane. The projective transformation can be approximated as an affine transformation. According to the properties of the affine transformation, the pixel coordinate values of the charging holes are corrected as follows: First, based on (u i , v i ), three mutually parallel lines l 12 , l 345 , l 67 are fitted, and the three parallel lines are defined by formula (1):

[0064]

[0065] where: u is the independent variable, v is the dependent variable, k, b, and △b are unknowns, and λ is used to describe the relative relationship between the parallel lines l 12 and l 67 . According to the geometric relationship between the charging holes ( Figure 2 ), λ = -0.0139 / 0.0112 can be obtained; according to the least squares method to fit the parallel lines, a target function D 2 about the three unknowns k, b, and △b of the parallel line equations is established:

[0066]

[0067] The partial derivatives of the target function with respect to k, b, and △b are taken respectively, and the expression values are set to 0, obtaining the following expressions:

[0068]

[0069] Substitute λ, (u i , v i ) into formula (3), the unknowns k, b, and △b can be solved, and then the three parallel lines l12 , l 345 , l 67 The specific expression of .

[0070] First adjust in the direction of perpendicular parallel lines: move point (u i ,v i ) is projected onto the corresponding parallel line, passing through point (u i ,v i ) respectively do l 12 , l 345 , l 67 Vertical line, such as Figure 3 As shown by the dashed line in the middle, the vertical line equation can be expressed as:

[0071]

[0072] Combining formulas (1) and (4), we can get (u i ,v i ) on the corresponding parallel line (u i ',v i '),like Figure 3 Shown as solid circle in the middle.

[0073] Then adjust in the parallel line direction (u axis direction): In this implementation, the translation reference point (u j ',v j ') is (u4', v4') (other points can also be selected as translation reference points j∈i, i=1,...,7), all projection points and parallel lines are translated at the same time until (u4', v4') is moved to the origin of the camera pixel coordinate system; with the origin of the pixel coordinate system as the rotation center, the translated parallel lines are rotated to the horizontal, and the rotation angle is β, which is determined by the slope k of the parallel lines, as shown in formula (5); after translation and rotation, the projection point (u i ',v i ') The corresponding rotation point is (u i ”,v i ”), the above process is shown in formula (6):

[0074] β=arctan(k) (5)

[0075]

[0076] Assume that point 4 remains unchanged after translation and rotation, or it can be any other point; correct the u coordinate of point 3 after translation and rotation, and add the correction amount du1; according to the characteristics of maintaining collinearity and distance ratio before and after affine transformation, point 4 can be obtained as the midpoint of point 3 and point 5, and correct the u coordinate of point 5 after translation and rotation; take the midpoint H of points 1 and 2, and add the correction amount du2; the corrected expression is shown in formula (7):

[0077]

[0078] Let the distance between point 3 and point 5 after two adjustments be d 35 , according to the characteristics of affine transformation that preserve collinearity and distance ratio, the distance d between point 1 and point 2 after two adjustments can be obtained 12 , the distance d between point 6 and point 7 after two adjustments 67 :

[0079]

[0080] Given d 12 , with H as the midpoint, thus the u - coordinates of point 1 and point 2 can be expressed as:

[0081]

[0082] According to the relationship between point I and point H, the u - coordinate of the midpoint I of point 6 and point 7 can be obtained:

[0083] u I = λu H (10)

[0084] Given d 67 , with I as the midpoint, thus the u - coordinates of point 6 and point 7 can be expressed as:

[0085]

[0086] The v - coordinates of point 1 to point 7 remain unchanged, expressed as:

[0087] v i ”’ = v”’ (12)

[0088] Taking the origin of the pixel coordinate system as the rotation center, rotating the parallel lines in the opposite direction by an angle β, and then translating all the corrected points until (u4”', v4”') coincides with the selected translation reference point (u4', v4'), the transformed points (u i ””, v i ””) are obtained:

[0089]

[0090] Construct the error function:

[0091]

[0092] To minimize the error between the corrected pixel coordinates and the original pixel coordinates, the error function err is differentiated with respect to du1 and du2 respectively and set to 0:

[0093]

[0094] Combining equations (7)-(11) and (13)-(15) to solve for du1 and du2, and substituting them back into equations (7), (9), and (11), and combining with equation (12), we get (u i ”', v i ”'). Then substituting into equation (13), we further obtain the pixel coordinates (u i ””, v i ””) of the charging hole after affine transformation correction.

[0095] Based on the same inventive concept as an affine transformation-based pixel coordinate correction method for an electric vehicle charging hole, the present application also provides an electronic device. The electronic device includes one or more processors and one or more memories. Computer-readable code is stored in the memories. When the computer-readable code is executed by one or more processors, the implementation of the affine transformation-based pixel coordinate correction method for an electric vehicle charging hole is performed. Among them, the memory may include a non-volatile storage medium and an internal memory; the non-volatile storage medium can store an operating system and computer-readable code. The computer-readable code includes program instructions. When the program instructions are executed, the processor can be made to execute any affine transformation-based pixel coordinate correction method for an electric vehicle charging hole. The processor is used to provide computing and control capabilities to support the operation of the entire electronic device. The memory provides an environment for the operation of the computer-readable code in the non-volatile storage medium. When the computer-readable code is executed by the processor, the processor can be made to execute any affine transformation-based pixel coordinate correction method for an electric vehicle charging hole.

[0096] It should be understood that the processor may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0097] In an embodiment of the present application, a computer-readable storage medium is also provided. The computer-readable storage medium stores computer-readable code. The computer-readable code includes program instructions. When the processor executes the program instructions, the affine transformation-based pixel coordinate correction method of the present application is implemented.

[0098] Among them, the computer-readable storage medium may be an internal storage unit of the electronic device described in the foregoing embodiments, such as the hard disk or memory of the computer device. The computer-readable storage medium may also be an external storage device of the electronic device, such as a plug-in hard disk, a SmartMedia Card (SMC), a Secure Digital (SD) card, a Flash Card, etc. equipped on the electronic device.

[0099] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some or all of the technical features; and 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, and they should all be covered by the scope of the claims and the description of the present invention.

Claims

1. A pixel coordinate correction method for an electric vehicle charging hole based on affine transformation, characterized in that: Suppose there are n charging holes in the charging interface. When the camera is facing the charging interface directly, the image of the charging interface collected by the camera is transmitted to the computer, and the original coordinates (u i , v i ) of the center of the charging hole in the camera pixel coordinate system are determined through the target detection algorithm. According to the coordinates (u i , v i ), a set of parallel lines are determined, where i = 1,..., n; Determine the projection points (u i , v i ) on the corresponding parallel lines as (u i ', v i '); Translate all projection points and the parallel lines they are on simultaneously until the translation reference point (u j ', v j ') is moved to the origin of the camera pixel coordinate system. Then, rotate the parallel lines around the origin of the camera pixel coordinate system until they are horizontal. The rotated point corresponding to the projection point is (u i ”, v i ”). Among them, the translation reference point (u j ', v j ') can be any point selected from (u i ', v i '). According to the characteristics of affine transformation, the rotation point (u i ”, v i ”) is corrected to obtain the expression of (u i ”', v i ”'); With the origin of the pixel coordinate system as the rotation center, rotate the parallel lines in the opposite direction and translate all the corrected points until (u j ”', v j ”') coincides with (u j ', v j '), and obtain the expression of the corrected point (u i ””, v i ””). The error function of the structure (u i "", v i ) and (u i , v i ) is such that the error function is differentiated with respect to the correction amount and set to 0 to solve for the correction amount, and then the expression of (u i "', v i "') is obtained. Finally, the pixel coordinate values (u i "", v i ) of the charging hole after affine transformation correction are obtained.

2. The method for correcting the pixel coordinates of the charging hole of an electric vehicle based on affine transformation according to claim 1, wherein, The process of determining a set of parallel lines according to the coordinates (u i , v i ) is as follows: According to the geometric relationship between the charging interface holes and the corresponding (u i , v i ), p mutually parallel lines are fitted: l1: v = ku + b + Δb l2: v = ku + b l3: v = ku + b + λ1Δb …… l p : v = ku + b + λ p-2 △b where: u is the independent variable, v is the dependent variable, and k, b, and Δb are unknowns; λ1, ……, λ p-2 used to describe the relative relationship between the corresponding parallel line and the parallel line l1, and its value is determined according to the geometric relationship between the charging holes; Let the parallel lines \(l_1, l_2, \cdots, l\) p be respectively fitted by \(n_1, n_2, \cdots, n\) p points, and establish the objective function \(D\) with respect to \(k\), \(b\), and \(\Delta b\) 2 : The objective function is respectively differentiated with respect to k, b, and △b, and the expression values are set to 0. Then, according to λ1, ……, λ p-2 , (u i , v i ) is used to solve the unknowns, and then p parallel lines are determined.

3. The method for correcting the pixel coordinates of an electric vehicle charging hole based on affine transformation according to claim 1, characterized in that, The parallel lines are translated as a whole until (u j ', v j ) coincides with the origin of the camera pixel coordinate system. Taking the origin of the pixel coordinate system as the rotation center, the parallel lines are rotated to be horizontal, and the rotation angle β satisfies: β = arctan(k), where k is the slope of the parallel lines.

4. The method for correcting pixel coordinates of an electric vehicle charging hole based on affine transformation according to claim 3, characterized in that, The said (u i ”, v i ”) and (u i ', v i ') satisfy:

5. The method for correcting the pixel coordinates of the charging hole of an electric vehicle based on affine transformation according to claim 4, wherein Correct the rotation point (u i ”, v i ”). Introduce the correction amount to obtain the expression of (u i ”', v i ”'), specifically as follows: Introduce k correction amounts du1, …, duk k , and according to the positional relationship of each charging hole in the actual charging interface, correct (u i ”, v i ”) to obtain the expression of (u i ”', v i ”').

6. The method for correcting the pixel coordinates of an electric vehicle charging hole based on affine transformation according to claim 5, characterized in that, (u i "", v i ") is expressed as:

7. The method for correcting pixel coordinates of an electric vehicle charging hole based on affine transformation according to claim 6, characterized in that The error function is

Images