High-precision extraction method and system for optical slice center based on weighted dimensionality reduction decomposition

Abstract

The present invention discloses a method and system for high-precision extraction of the center of an optical slice based on weighted dimensionality reduction decomposition, which relates to the field of calibration of line-structured light vision measurement systems in two / three-dimensional vision measurement. The method for high-precision extraction of the center of an optical slice based on weighted dimensionality reduction decomposition mainly includes: reducing the dimensionality of the "image layer space" and decomposing it to obtain a gray vector, analyzing and calculating the gray vector to obtain the sub-pixel coordinates of the decomposition direction of the "image layer space", matching the serial number in the vertical direction of the decomposition direction of the current gray vector to obtain the center coordinates of the optical slice corresponding to the gray vector, and traversing and calculating the center coordinates of the optical slices corresponding to all gray vectors to obtain the center of the optical slice of the light strip pixel-level image. Implementing the method and system for high-precision extraction of the center of an optical slice based on weighted dimensionality reduction decomposition provided by the present invention can improve the real-time performance and accuracy of the extraction of the center of the optical slice in a single-line laser scenario.

Description

Technical Field

[0001] The present invention relates to the field of calibration of line structured light vision measurement systems in two / three-dimensional vision measurement, and more specifically, to a method and system for high-precision extraction of the center of optical slices based on weighted dimensionality reduction decomposition. Background Art

[0002] The optical slice laser scanning three-dimensional measurement method is an important active vision technology. Based on the principle of three-dimensional laser scanning technology, it has the advantages of a wide measurement range, high measurement efficiency, and a small volume of equipment hardware, and is widely used in key national development fields such as aerospace, automotive intelligent manufacturing, and ship intelligent manufacturing.

[0003] The principle of optical slice laser scanning three-dimensional measurement is to project a single-line laser of an optical slice onto a heterogeneous machining surface to generate an optical slice contour representing the corresponding cross-section of the heterogeneous machining surface. After the imaging system acquires the optical slice contour, it first extracts the sub-pixel center coordinates of the optical slice contour in the image space, and maps the center coordinates of the optical slice in the image space to the physical space (calibration) through calibration to obtain the true height data of the entire optical slice contour. Specifically, through one-dimensional translation, high-precision three-dimensional point cloud data of the heterogeneous machining surface can be obtained by one-way scanning in the direction perpendicular to the projection plane of the single-line laser of the optical slice. Therefore, extracting the sub-pixel center coordinates of the optical slice contour in the image space is an important prerequisite for the optical slice laser scanning three-dimensional measurement method, and the calculation of the sub-pixel center coordinates of the optical slice contour in the image space with high efficiency, high precision, and high robustness is the key to ensuring the real-time performance, high precision, and high robustness of the line structured light vision measurement system.

[0004] Traditional methods for extracting the center of optical slices mainly include the Steger algorithm, etc. The Steger algorithm is based on the Hessian matrix, obtains the normal direction of the optical slice contour in the image space through the Hessian matrix, and then obtains the center point in the normal direction to obtain the sub-pixel position of the center of the optical slice contour. This method has high robustness and high precision, but it has a large amount of computation, is difficult to quickly extract the center line of the optical slice contour, and is difficult to meet the requirements of applications with high real-time performance such as rapid measurement and three-dimensional reconstruction.

[0005] J. Zhang proposed a method for high-precision extraction of the center line of structured light stripes for three-dimensional reconstruction. This method processes the image through a secondary optimization algorithm of threshold segmentation and gray center of gravity method and extracts its sub-pixel center. Based on J. Zhang's method, multi-line laser measurement can not only be realized, but also the center line of the structured light stripes can be extracted with high real-time performance. However, this method needs to process the image using Gaussian filtering, which is not only time-consuming but also has a negative edge blurring effect, and the edge pixel points often change, so it is difficult to achieve high-precision and high-stability extraction of the center coordinates of the light stripes.

[0006] The patent "Line Structured Light Center Extraction Method and Device Based on Gaussian Hyper-Laplacian Distribution" (Patent No.: ZL202311135784.4) proposes a line structured light center extraction method based on Gaussian hyper-Laplacian distribution. Gaussian filtering is used to process the structured light image, which can remove the noise affecting the fitting accuracy without destroying the original shape and structure of the line structured light. This method has a certain anti-noise ability and meets certain accuracy requirements. However, in the method of determining the fitting region within the light strip boundary region in this method, by expanding the points with the maximum gray value in each column by twice the width of the line structured light strip to both sides to determine the fitting interval, since there may be multiple and discontinuous maximum points of gray values in each column during the extraction of the structured light center, it is easy to produce large errors. Therefore, this method has poor anti-interference ability and low accuracy.

[0007] Scholars such as Y. Gong, G. Liu, H. Huang, and T. Song proposed an improved laser centerline extraction algorithm based on internal propulsion. This method first fits the tangent equation based on the center point to determine the normal vector of the laser line, then calculates the sub-pixel coordinates of the normal vector of the center point. Finally, linear smoothing processing is used to make the laser center point smoother. Compared with the gray center of gravity method, the root mean square error of this method is increased by 0.337 pixels, increasing the robustness of the centerline extraction. However, in the threshold-based light strip center point search algorithm proposed by these scholars, obtaining the center point coordinates through the center method will reduce the accuracy. Moreover, this method has a large amount of computation, and the problem of time consumption has not been well solved.

[0008] In summary, traditional centerline extraction methods are generally divided into two categories. The first category is the geometric center method represented by the threshold method. This type of method has strong real-time performance, strong anti-interference ability, and high accuracy, but it takes more time. Although later scholars carried out image stereo correction and preprocessed the image using Gaussian filtering to reduce errors, it is still impossible to achieve rapid detection of complex objects. The second type of method is the energy center method represented by the gray center of gravity method. This type of method is suitable for multi-line laser measurement scenarios and finds the structured light center through the contour information of the structured light. Therefore, this type of method is more applied to matching multi-line laser centers. It has high real-time performance but is greatly affected by noise and has low accuracy. Moreover, the conventional use of the gray center of gravity method is to process two-dimensional scenarios, while the multi-threshold sub-region weighted method mentioned in this embodiment processes the one-dimensional scenario after the two-dimensional scenario is dimensionally reduced and decomposed, improving the real-time performance of structured light center measurement. Later, some scholars proposed a secondary optimization algorithm for the gray center of gravity method, and the accuracy has been improved to a certain extent compared with the gray center of gravity method, but it still cannot meet the high-precision structured light center measurement. Therefore, the current line structured light vision measurement system has low real-time performance, low accuracy, and low robustness. Summary of the Invention

[0009] The object of the present invention is to provide a method and system for accurately extracting the center of an optical slice by weighted dimensionality reduction decomposition, which can improve the real-time performance and accuracy of extracting the center of an optical slice in a single-line laser scenario.

[0010] The present invention provides a method for accurately extracting the center of an optical slice based on weighted dimensionality reduction decomposition, including: S1: Obtain an image of the light stripe at the pixel level, and according to the image of the light stripe at the pixel level, obtain the dimensionality reduction decomposition direction of the "image layer space" of the image of the light stripe at the pixel level and the direction perpendicular to the dimensionality reduction decomposition direction of the "image layer space" of the image of the light stripe at the pixel level; S2: Grayscale the image of the light stripe at the pixel level, and decompose the image of the light stripe at the pixel level according to the dimensionality reduction decomposition direction of the "image layer space" of the image of the light stripe at the pixel level to obtain a grayscale vector of the "image layer space" of the image of the light stripe at the pixel level; S3: According to the grayscale vector of the "image layer space" of the image of the light stripe at the pixel level, obtain the sub-pixel coordinates of the decomposition direction of the "image layer space" of the grayscale vector; S4: According to the sub-pixel coordinates of the decomposition direction of the "image layer space" of the grayscale vector, match the serial number of the direction perpendicular to the decomposition direction of the current grayscale vector to obtain the coordinates of the center of the optical slice corresponding to the grayscale vector; S5: Traverse and calculate the coordinates of the center of the optical slice corresponding to all grayscale vectors to obtain the center of the optical slice of the image of the light stripe at the pixel level.

[0011] Further, step S1 of the above method for accurately extracting the center of an optical slice based on weighted dimensionality reduction decomposition specifically includes: S11: Obtain an image of the light stripe at the pixel level, and according to the image of the light stripe at the pixel level, obtain the main direction of the optical slice, the row direction of the "image layer space" of the image of the light stripe at the pixel level, and the column direction of the "image layer space" of the image of the light stripe at the pixel level; S12: According to the main direction of the optical slice, the row direction of the "image layer space" of the image of the light stripe at the pixel level, and the column direction of the "image layer space" of the image of the light stripe at the pixel level, obtain the angle between the main direction of the optical slice and the row direction of the "image layer space" and the angle between the main direction of the optical slice and the column direction of the "image layer space"; S13: When the angle between the main direction of the optical slice and the row direction of the "image layer space" is not less than the angle between the main direction of the optical slice and the column direction of the "image layer space", take the column direction of the image of the light stripe at the pixel level as the dimensionality reduction decomposition direction of the "image layer space" of the image of the light stripe at the pixel level, and take the row direction of the image of the light stripe at the pixel level as the direction perpendicular to the dimensionality reduction decomposition direction of the "image layer space" of the image of the light stripe at the pixel level; when the angle between the main direction of the optical slice and the row direction of the "image layer space" is less than the angle between the main direction of the optical slice and the column direction of the "image layer space", take the row direction of the image of the light stripe at the pixel level as the dimensionality reduction decomposition direction of the "image layer space" of the image of the light stripe at the pixel level, and take the column direction of the image of the light stripe at the pixel level as the direction perpendicular to the dimensionality reduction decomposition direction of the "image layer space" of the image of the light stripe at the pixel level.

[0012] Further, step S3 of the above method for accurately extracting the center of an optical slice based on weighted dimensionality reduction decomposition specifically includes: According to the grayscale vector of the "image layer space" of the image of the light stripe at the pixel level, obtain the sub-pixel coordinates of the decomposition direction of the "image layer space" of the grayscale vector, as shown in the formula:

[0013]

[0014] where P k is the sub-pixel coordinate of the decomposition direction of the "image layer space" of the k-th gray vector, n is the number of sub-pixel coordinates existing in the reserved area, and P ki is the i-th sub-pixel coordinate point of the decomposition direction of the "image layer space" of the k-th gray vector, and A ki is the weight corresponding to the gray value of the i-th sub-pixel coordinate point of the decomposition direction of the "image layer space" of the k-th gray vector, r ki is the gray value of the i-th sub-pixel coordinate point of the decomposition direction of the "image layer space" of the k-th gray vector, a is the first preset coefficient for calculating the gray value weight, b is the second preset coefficient for calculating the gray value weight, c is the third preset coefficient for calculating the gray value weight, Th is the gray baseline constant, T is the preset gray threshold, σ 2 (.) is the variance function, Max(.) is the maximum value function, and V G is the total gray value of the entire image, and R A (T) is the probability that a pixel point is assigned to class A, and V A (T) is the average gray value of the pixel points assigned to class A, and R B (T) is the probability that a pixel point is assigned to class B, and V B (T) is the average gray value of the pixel points assigned to class B, and R i is the probability that the gray value of a pixel point is i, and g i is the number of pixels with gray value i in the image.

[0015] Furthermore, step S4 of the above method for accurately extracting the optical slice center based on weighted dimensionality reduction decomposition specifically includes: According to the sub-pixel coordinates of the decomposition direction of the "image layer space" of the gray vector, match the serial number in the vertical direction of the decomposition direction of the current gray vector to obtain the optical slice center coordinates corresponding to the gray vector, as shown in the formula:

[0016] P k (R k , C k ) = (P k , k)

[0017] where P k (R k , C k ) are the optical slice center coordinates corresponding to the k-th gray vector, R k is the optical slice center decomposition direction coordinate of the "image layer space", C k is the coordinate in the vertical direction of the optical slice center decomposition direction of the "image layer space", and P kIt is the sub-pixel coordinate of the decomposition direction of the "image layer space" of the k-th gray vector, and k is the serial number of the vertical direction of the decomposition direction of the current gray vector.

[0018] Further, step S5 of the above-mentioned high-precision extraction method for the center of the optical slice based on weighted dimensionality reduction decomposition specifically includes: traversing and calculating the center coordinates of the optical slice corresponding to all gray vectors to obtain the center of the optical slice of the light strip pixel-level image, as shown in the formula:

[0019] {m} = {P k (R k , C k ) | k = 1, 2,..., H}

[0020] Among them, {m} is the center of the optical slice of the light strip pixel-level image, and H is the number of gray vectors.

[0021] The present invention also provides a system, including the following modules: an image acquisition module configured to: acquire a light strip pixel-level image, and obtain the dimensionality reduction decomposition direction of the "image layer space" of the light strip pixel-level image and the vertical direction of the dimensionality reduction decomposition direction of the "image layer space" of the light strip pixel-level image according to the light strip pixel-level image; a vector acquisition module configured to: grayscale the light strip pixel-level image, and decompose the light strip pixel-level image according to the dimensionality reduction decomposition direction of the "image layer space" of the light strip pixel-level image to obtain the gray vector of the "image layer space" of the light strip pixel-level image; a decomposition coordinate module configured to: obtain the sub-pixel coordinate of the decomposition direction of the "image layer space" of the gray vector according to the gray vector of the "image layer space" of the light strip pixel-level image; a coordinate matching module configured to: match the serial number of the vertical direction of the decomposition direction of the current gray vector according to the sub-pixel coordinate of the decomposition direction of the "image layer space" of the gray vector to obtain the center coordinate of the optical slice corresponding to the gray vector; an image optical slice center module configured to: traverse and calculate the center coordinates of the optical slice corresponding to all gray vectors to obtain the center of the optical slice of the light strip pixel-level image.

[0022] Further, the image acquisition module of the above system is specifically configured as follows: acquiring a light stripe pixel-level image, and obtaining the main direction of the light slice, the row direction of the "image layer space" of the light stripe pixel-level image, and the column direction of the "image layer space" of the light stripe pixel-level image based on the light stripe pixel-level image; obtaining the angle between the main direction of the light slice and the row direction of the "image layer space" and the angle between the main direction of the light slice and the column direction of the "image layer space" based on the main direction of the light slice, the row direction of the "image layer space" of the light stripe pixel-level image, and the column direction of the "image layer space" of the light stripe pixel-level image; when the angle between the main direction of the light slice and the row direction of the "image layer space" is not less than the angle between the main direction of the light slice and the column direction of the "image layer space", taking the column direction of the light stripe pixel-level image as the dimensionality reduction decomposition direction of the "image layer space" of the light stripe pixel-level image, and taking the row direction of the light stripe pixel-level image as the direction perpendicular to the dimensionality reduction decomposition direction of the "image layer space" of the light stripe pixel-level image; when the angle between the main direction of the light slice and the row direction of the "image layer space" is less than the angle between the main direction of the light slice and the column direction of the "image layer space", taking the row direction of the light stripe pixel-level image as the dimensionality reduction decomposition direction of the "image layer space" of the light stripe pixel-level image, and taking the column direction of the light stripe pixel-level image as the direction perpendicular to the dimensionality reduction decomposition direction of the "image layer space" of the light stripe pixel-level image.

[0023] Further, the decomposition coordinate module of the above system is specifically configured as follows: obtaining the sub-pixel coordinates of the decomposition direction of the "image layer space" of the gray vector based on the gray vector of the "image layer space" of the light stripe pixel-level image, as shown in the formula:

[0024]

[0025]

[0026] where P k is the sub-pixel coordinate of the decomposition direction of the "image layer space" of the k-th gray vector, n is the number of sub-pixel coordinates existing in the reserved area, P ki is the i-th sub-pixel coordinate point of the decomposition direction of the "image layer space" of the k-th gray vector, A ki is the weight corresponding to the gray value of the i-th sub-pixel coordinate point of the decomposition direction of the "image layer space" of the k-th gray vector, r ki is the gray value of the i-th sub-pixel coordinate point of the decomposition direction of the "image layer space" of the k-th gray vector, a is the first preset coefficient for calculating the gray value weight, b is the second preset coefficient for calculating the gray value weight, c is the third preset coefficient for calculating the gray value weight, Th is the gray baseline constant, T is the preset gray threshold, σ 2 (.) is the variance function, Max(.) is the maximum value function, V G is the total gray value of the entire image, R A (T) is the probability that a pixel point is classified into class A, V A (T) is the average gray value of the pixel points assigned to class A, RB (T) is the probability that a pixel is classified into class B, V B (T) is the average gray level of the pixels assigned to class B, R i is the probability that the gray level of a pixel is i, g i is the number of pixels with gray level i in the image.

[0027] Furthermore, the coordinate matching module of the above system is specifically configured as follows: according to the sub-pixel coordinates of the decomposition direction of the "image layer space" of the gray vector, match the serial number of the vertical direction of the decomposition direction of the current gray vector to obtain the optical slice center coordinates corresponding to the gray vector, as shown in the formula:

[0028] P k (R k ,C k )=(P k ,k)

[0029] where, P k (R k ,C k ) are the optical slice center coordinates corresponding to the k-th gray vector, R k is the decomposition direction coordinate of the optical slice center of the "image layer space", C k is the coordinate of the vertical direction of the decomposition direction of the optical slice center of the "image layer space", P k is the sub-pixel coordinate of the decomposition direction of the "image layer space" of the k-th gray vector, and k is the serial number of the vertical direction of the decomposition direction of the current gray vector.

[0030] Furthermore, the image optical slice center module of the above system is specifically configured as follows: traverse and calculate the optical slice center coordinates corresponding to all gray vectors to obtain the optical slice center of the light bar pixel-level image, as shown in the formula:

[0031] {m}={P k (R k ,C k )|k = 1, 2,..., H}

[0032] where, {m} is the optical slice center of the light bar pixel-level image, and H is the number of gray vectors.

[0033] Implementing the method and system for high-precision extraction of the optical slice center based on weighted dimensionality reduction decomposition provided by the present invention has the following beneficial effects:

[0034] Construct a dimensionality reduction decomposition model of the two-dimensional gray matrix of the "image layer space". By decomposing the two-dimensional gray matrix of the "image layer space" into an ordered "gray vector pool", the problem of extracting the two-dimensional coordinates of the optical slice center of the "image layer space" is transformed into the problem of high-precision calculation of the vector center point in the "gray vector pool" and vector serial number matching, reducing the difficulty of extracting the optical slice center;

[0035] Study the gray-scale distribution characteristic mechanism of each gray-scale vector in the "gray-scale vector pool", construct a high-precision calculation model for the gray-scale center point of the vector in the "gray-scale vector pool", and improve the accuracy of extracting the center of the optical slice;

[0036] Based on the orderliness of the "gray-scale vector pool", by circularly traversing the "gray-scale vector pool", the matching between the serial number of the vector in the "gray-scale vector pool" and the gray-scale center point of the vector can be realized. Compared with the operations of the two-dimensional matrix (convolution, thinning, morphological algorithm), the computational amount of extracting the center of the optical slice is reduced, thereby improving the real-time performance of the extraction. Brief Description of the Drawings

[0037] The present invention will be further described below in conjunction with the drawings and embodiments. In the drawings:

[0038] Figure 1 is a flowchart of the high-precision optical slice center extraction method based on weighted dimensionality reduction decomposition provided by the present invention;

[0039] Figure 2 is a schematic diagram for selecting the dimensionality reduction decomposition direction of the "image layer space" provided by the present invention;

[0040] Figure 3 is a schematic diagram of dimensionality reduction decomposition of the "image layer space" in the column direction (perpendicular to the main direction of the optical slice) provided by the present invention;

[0041] Figure 4 is a schematic diagram of the gray-scale distribution of the gray-scale vector provided by the present invention;

[0042] Among them, (a) is the gray-scale distribution schematic diagram of the gray-scale vector V 1 and (b) is the gray-scale distribution schematic diagram of the gray-scale vector V K ;

[0043] Figure 5 is a schematic diagram of the "gray-scale - weight mapping curve" provided by the present invention;

[0044] Figure 6 is a schematic diagram for realizing the matching method of the serial numbers of the gray-scale vectors in the "gray-scale vector pool" taking V K as an example provided by the present invention. Detailed Description of the Specific Embodiment

[0045] For a clearer understanding of the technical features, objectives, and effects of the present invention, the specific embodiments of the present invention will now be described in detail with reference to the drawings.

[0046] Figure 1 shows a schematic diagram of the high-precision optical slice center extraction method based on weighted dimensionality reduction decomposition in this embodiment. In this embodiment, the high-precision optical slice center extraction method based on weighted dimensionality reduction decomposition includes:

[0047] S1: Obtain the light stripe pixel-level image. According to the light stripe pixel-level image, obtain the "image layer space" dimensionality reduction decomposition direction of the light stripe pixel-level image and the direction perpendicular to the "image layer space" dimensionality reduction decomposition direction of the light stripe pixel-level image;

[0048] Specifically, step S1 includes: S11: Obtain the light stripe pixel-level image. According to the light stripe pixel-level image, obtain the main direction of the light slice, the "image layer space" row direction and the "image layer space" column direction of the light stripe pixel-level image; S12: According to the main direction of the light slice, the "image layer space" row direction and the "image layer space" column direction of the light stripe pixel-level image, obtain the angle between the main direction of the light slice and the "image layer space" row direction and the angle between the main direction of the light slice and the "image layer space" column direction; S13: When the angle between the main direction of the light slice and the "image layer space" row direction is not less than the angle between the main direction of the light slice and the "image layer space" column direction, use the column direction of the light stripe pixel-level image as the "image layer space" dimensionality reduction decomposition direction of the light stripe pixel-level image, and use the row direction of the light stripe pixel-level image as the direction perpendicular to the "image layer space" dimensionality reduction decomposition direction of the light stripe pixel-level image; when the angle between the main direction of the light slice and the "image layer space" row direction is less than the angle between the main direction of the light slice and the "image layer space" column direction, use the row direction of the light stripe pixel-level image as the "image layer space" dimensionality reduction decomposition direction of the light stripe pixel-level image, and use the column direction of the light stripe pixel-level image as the direction perpendicular to the "image layer space" dimensionality reduction decomposition direction of the light stripe pixel-level image;

[0049] S2: Grayscale the light stripe pixel-level image. According to the "image layer space" dimensionality reduction decomposition direction of the light stripe pixel-level image, decompose the light stripe pixel-level image to obtain the grayscale vector of the "image layer space" of the light stripe pixel-level image;

[0050] S3: According to the grayscale vector of the "image layer space" of the light stripe pixel-level image, obtain the sub-pixel coordinates of the decomposition direction of the "image layer space" of the grayscale vector;

[0051] Specifically, step S3 includes: According to the grayscale vector of the "image layer space" of the light stripe pixel-level image, obtain the sub-pixel coordinates of the decomposition direction of the "image layer space" of the grayscale vector, such as the formula:

[0052]

[0053] where, P k is the sub-pixel coordinate of the decomposition direction of the "image layer space" of the k-th grayscale vector, n is the number of sub-pixel coordinates existing in the reserved area, P ki is the i-th sub-pixel coordinate point of the decomposition direction of the "image layer space" of the k-th grayscale vector, A ki is the weight corresponding to the grayscale value of the i-th sub-pixel coordinate point of the decomposition direction of the "image layer space" of the k-th grayscale vector, rki is the gray value of the i-th sub-pixel coordinate point in the decomposition direction of the "image layer space" of the k-th gray vector, a is the first preset coefficient for calculating the gray value weight, b is the second preset coefficient for calculating the gray value weight, c is the third preset coefficient for calculating the gray value weight, Th is the gray baseline constant, T is the preset gray threshold, σ 2 (.) is the variance function, Max(.) is the maximum value function, V G is the total gray value of the entire image, R A (T) is the probability that a pixel point is assigned to class A, V A (T) is the average gray value of the pixel points assigned to class A, R B (T) is the probability that a pixel point is assigned to class B, V B (T) is the average gray value of the pixel points assigned to class B, R i is the probability that the gray value of a pixel point is i, g i is the number of pixels with gray value i in the image;

[0054] S4: According to the sub-pixel coordinates in the decomposition direction of the "image layer space" of the gray vector, match the serial number in the vertical direction of the decomposition direction of the current gray vector to obtain the optical slice center coordinates corresponding to the gray vector;

[0055] Specifically, step S4 includes: According to the sub-pixel coordinates in the decomposition direction of the "image layer space" of the gray vector, match the serial number in the vertical direction of the decomposition direction of the current gray vector to obtain the optical slice center coordinates corresponding to the gray vector, as shown in the formula:

[0056] P k (R k ,C k )=(P k ,k)

[0057] where, P k (R k ,C k ) are the optical slice center coordinates corresponding to the k-th gray vector, R k is the optical slice center decomposition direction coordinate of the "image layer space", C k is the vertical direction coordinate of the optical slice center decomposition direction of the "image layer space", P k is the sub-pixel coordinate in the decomposition direction of the "image layer space" of the k-th gray vector, k is the serial number in the vertical direction of the decomposition direction of the current gray vector;

[0058] S5: Traverse and calculate the optical slice center coordinates corresponding to all gray vectors to obtain the optical slice center of the light strip pixel-level image;

[0059] Specifically, step S5 includes: traversing and calculating the optical slice center coordinates corresponding to all grayscale vectors to obtain the optical slice center of the light strip pixel-level image, as shown in the formula:

[0060] {m} = {P k (R k , C k )|k = 1, 2,..., H}

[0061] where {m} is the optical slice center of the light strip pixel-level image, and H is the number of grayscale vectors.

[0062] This embodiment provides a system, including the following modules:

[0063] An image acquisition module, configured to: acquire a light strip pixel-level image, and based on the light strip pixel-level image, obtain the "image layer space" dimensionality reduction decomposition direction of the light strip pixel-level image and the direction perpendicular to the "image layer space" dimensionality reduction decomposition direction of the light strip pixel-level image;

[0064] Specifically, the image acquisition module of the above system is configured to: acquire a light strip pixel-level image, and based on the light strip pixel-level image, obtain the main optical slice direction, the "image layer space" row direction and the "image layer space" column direction of the light strip pixel-level image; based on the main optical slice direction, the "image layer space" row direction and the "image layer space" column direction of the light strip pixel-level image, obtain the angle between the main optical slice direction and the "image layer space" row direction and the angle between the main optical slice direction and the "image layer space" column direction; when the angle between the main optical slice direction and the "image layer space" row direction is not less than the angle between the main optical slice direction and the "image layer space" column direction, use the column direction of the light strip pixel-level image as the "image layer space" dimensionality reduction decomposition direction of the light strip pixel-level image, and use the row direction of the light strip pixel-level image as the direction perpendicular to the "image layer space" dimensionality reduction decomposition direction of the light strip pixel-level image; when the angle between the main optical slice direction and the "image layer space" row direction is less than the angle between the main optical slice direction and the "image layer space" column direction, use the row direction of the light strip pixel-level image as the "image layer space" dimensionality reduction decomposition direction of the light strip pixel-level image, and use the column direction of the light strip pixel-level image as the direction perpendicular to the "image layer space" dimensionality reduction decomposition direction of the light strip pixel-level image;

[0065] A vector acquisition module, configured to: grayscale the light strip pixel-level image, and based on the "image layer space" dimensionality reduction decomposition direction of the light strip pixel-level image, decompose the light strip pixel-level image to obtain the grayscale vectors of the "image layer space" of the light strip pixel-level image;

[0066] A decomposition coordinate module, configured to: based on the grayscale vectors of the "image layer space" of the light strip pixel-level image, obtain the sub-pixel coordinates of the decomposition direction of the grayscale vectors in the "image layer space";

[0067] Specifically, the decomposition coordinate module of the above system is configured to: obtain the sub-pixel coordinates of the decomposition direction of the "image layer space" of the gray vector according to the gray vector of the light strip pixel-level image, as shown in the formula:

[0068]

[0069]

[0070] where P k is the sub-pixel coordinate of the decomposition direction of the "image layer space" of the k-th gray vector, n is the number of sub-pixel coordinates existing in the reserved area, P ki is the i-th sub-pixel coordinate point of the decomposition direction of the "image layer space" of the k-th gray vector, A ki is the weight corresponding to the gray value of the i-th sub-pixel coordinate point of the decomposition direction of the "image layer space" of the k-th gray vector, r ki is the gray value of the i-th sub-pixel coordinate point of the decomposition direction of the "image layer space" of the k-th gray vector, a is the first preset coefficient for calculating the gray value weight, b is the second preset coefficient for calculating the gray value weight, c is the third preset coefficient for calculating the gray value weight, Th is the gray baseline constant, T is the preset gray threshold, σ 2 (.) is the variance function, Max(.) is the maximum value function, V G is the total gray value of the entire image, R A (T) is the probability that a pixel point is assigned to class A, V A (T) is the average gray value of the pixel points assigned to class A, R B (T) is the probability that a pixel point is assigned to class B, V B (T) is the average gray value of the pixel points assigned to class B, R i is the probability that the gray value of a pixel point is i, g i is the number of pixels with gray value i in the image;

[0071] The coordinate matching module is configured to: match the serial number in the vertical direction of the decomposition direction of the current gray vector according to the sub-pixel coordinates of the decomposition direction of the "image layer space" of the gray vector, and obtain the optical slice center coordinates corresponding to the gray vector;

[0072] Specifically, the coordinate matching module of the above system is configured to: match the serial number in the vertical direction of the decomposition direction of the current gray vector according to the sub-pixel coordinates of the decomposition direction of the "image layer space" of the gray vector, and obtain the optical slice center coordinates corresponding to the gray vector, as shown in the formula:

[0073] P k (R k , C k ) = (P k , k)

[0074] Among them, P k (R k , C k ) is the center coordinate of the optical slice corresponding to the k-th gray vector, R k is the decomposition direction coordinate of the center of the optical slice in the "image layer space", C k is the coordinate in the direction perpendicular to the decomposition direction of the center of the optical slice in the "image layer space", P k is the sub-pixel coordinate of the decomposition direction of the "image layer space" of the k-th gray vector, and k is the serial number in the direction perpendicular to the decomposition direction of the current gray vector;

[0075] Image optical slice center module, configured to: traverse and calculate the center coordinates of the optical slices corresponding to all gray vectors to obtain the center of the optical slice of the light strip pixel-level image;

[0076] Specifically, the image optical slice center module of the above system is configured to: traverse and calculate the center coordinates of the optical slices corresponding to all gray vectors to obtain the center of the optical slice of the light strip pixel-level image, as shown in the formula:

[0077] {m} = {P k (R k , C k ) | k = 1, 2,..., H}

[0078] Among them, {m} is the center of the optical slice of the light strip pixel-level image, and H is the number of gray vectors.

[0079] In some embodiments, the above method for high-precision extraction of the optical slice center based on weighted dimensionality reduction decomposition can also be implemented through the following steps:

[0080] Step 1. Selection of the dimensionality reduction decomposition direction of the "image layer space". Assume that the resolution of the "image layer space" is W × H:

[0081] 1) Determine the main direction of the optical slice;

[0082] 2) Analyze the angle α between the main direction of the optical slice and the row direction of the "image layer space";

[0083] 3) Analyze the angle β between the main direction of the optical slice and the column direction of the "image layer space";

[0084] 4) Compare the magnitudes of α and β. When α ≥ β, perform dimensionality reduction decomposition on the "image layer space" in the column direction; when α < β, perform dimensionality reduction decomposition on the "image layer space" in the row direction and grayscale the current image. Next, take the dimensionality reduction decomposition of the "image layer space" in the column direction as an example. Similarly, the dimensionality reduction decomposition of the "image layer space" in the row direction is similar. The schematic diagram of the selection of the dimensionality reduction decomposition direction of the "image layer space" is as Figure 2 shown;

[0085] Step 2. Perform dimensionality reduction decomposition on the "image layer space" in the column direction:

[0086] Assume that the resolution of the "image layer space" is W×H. Then, the "image layer space" can be decomposed in the column vector dimensionality reduction into a "gray vector pool" composed of a total of H gray vectors {V k |k = 1, 2... H}, as shown in Figure 3 ; Taking the gray vector V k as an example, the center coordinates of the optical slice corresponding to this gray vector are:

[0087] PA k (R k , C k ) = (P k , k)

[0088] In the formula, R k is the column direction coordinate of the center of the optical slice of the "image layer space"; C k is the row direction coordinate of the center of the optical slice of the "image layer space"; P k is the column direction sub-pixel coordinate of the "image layer space" of the gray vector V k in the "gray vector pool"; k is the vector serial number in the row direction of the gray vector V k ;

[0089] Step 3. Realize the high-precision calculation method for the center point of the gray vector in the "gray vector pool":

[0090] After the dimensionality reduction segmentation of the "image layer space", the extraction of the center of the optical slice of the "image layer space" is transformed from a two-dimensional matrix operation problem into the center calculation problem of the gray curves of H gray vectors. Considering the gradual change characteristics that the center of the gray distribution of the optical slice is the highest and the two sides are the lowest, and at the same time to suppress the gray interference of the ambient light, this embodiment proposes a vector adaptive threshold algorithm. Based on each vector, its gray baseline constant Th is set, and the calculation method of Th is as follows:

[0091] Let g i be the number of pixels with gray level i in the image. Then there are g 0 , g 1 , g 2 , then the probability that the gray level of any point in the gray scale diagram is i is:

[0092]

[0093] And there is:

[0094]

[0095] Let the threshold be T. The points in the image are classified into two categories A and B according to their corresponding gray levels, where A ∈ (0, T) and B ∈ (T + 1, 255). Then, if an arbitrary point is selected, the probability that this point is classified into category A is R A (T), and the average gray level of the points classified into category A is V A (T):

[0096]

[0097] Similarly, the probability that a point is classified into category B is R B (T), and the average gray level of the points classified into category B is V B (T):

[0098]

[0099] The total gray level value V of the entire image G :

[0100]

[0101] Therefore, it can be known that the variance is:[[]]END]]

[0102] σ 2 (T) = R A (T)(V A (T) - V G ) 2 + R B (T)(V B (T) - V G ) 2

[0103] Furthermore, the gray level baseline constant Th can be determined by the following formula.

[0104]

[0105] Furthermore, it is discussed in two cases:

[0106] 1) In the first case, there is no light slice distribution on the gray level vector:

[0107] Taking the gray level vector V Figure 3 as an example, 1 (a) is the schematic diagram of the gray level distribution of V Figure 4 ; Since there is no light slice gray level distribution in V 1 , and the entire gray level distribution is lower than the gray level baseline constant Th. In this case, it is considered that there is no effective center point on V 1 , that is, there is no light slice distribution on the gray level vector V 1 ; 1

[0108] 2) In the second case, there is a light slice distribution on the gray vector:

[0109] Taking Figure 3 the gray vector V K as an example, Figure 4 (b) is the schematic diagram of the gray distribution of the gray vector V K ; Since there is a light slice distribution on V K , it is considered that there is an effective center point on V k . Therefore, the images in the schematic diagram of the gray distribution of the gray vector V k with gray values greater than the gray baseline constant Th are retained, and the images with gray values less than Th are removed, and the processing is carried out through the following formula;

[0110]

[0111] In the formula, r ki is the gray value of the i-th sub-pixel coordinate point in the column direction of the "image layer space" in the schematic diagram of the gray distribution of the gray vector V K ;

[0112] According to the gray mapping principle, different weights A ki are set for different gray values r ki ; The higher the gray value, the higher the reliability of the sub-pixel coordinate point in the column direction of the "image layer space" it represents. Therefore, the higher the gray value, the larger the specified A ki . Therefore, in order to increase the proportion of high brightness, this embodiment proposes a "gray-weight mapping curve", as Figure 5 shown. This example in this embodiment is only one embodiment of the mapping, and this curve is not limited to this one and can be changed according to the scene and conditions. The function is as follows, and r ki ∈(Th, 255), where Th = 200 in this example:

[0113]

[0114] In the formula, A ki is the weight corresponding to the gray value of the i-th sub-pixel coordinate point in the column direction of the "image layer space" in the schematic diagram of the gray distribution of the gray vector V K ;

[0115] Furthermore, the sub-pixel coordinate P K in the column direction of the "image layer space" of the gray vector V k can be calculated through the following formula:

[0116]

[0117] In the formula, n is the number of sub-pixel coordinates in the retained area, and P ki is the gray vector V KIn the schematic diagram of the gray-scale distribution, the i-th sub-pixel coordinate point in the column direction of the "image layer space", P k is V K The sub-pixel coordinate R in the column direction of the optical slice center of the "image layer space" K ;

[0118] Step 4. Implementation of the method for matching the serial numbers of gray-scale vectors in the "gray-scale vector pool":

[0119] Based on Steps 1, 2, and 3, by means of loop traversal, the calculation of the center points of all gray-scale vectors in the "gray-scale vector pool" can be realized. Taking V k as an example, when obtaining the sub-pixel coordinate R k of the optical slice center of V k , the row-direction index number of the current gray-scale vector is matched, and the matching of the optical slice center coordinates (P k , k) is realized in combination with the following formula Figure 6 is the schematic diagram for implementing the method for matching the serial numbers of gray-scale vectors in the "gray-scale vector pool" with Vk as an example:

[0120] P k (R k , C k ) = (P k , k)

[0121] After traversing and calculating the H gray-scale vectors in the "gray-scale vector pool", the set {m} can be obtained, and finally a high-real-time extraction model of the optical slice center coordinates of the "image layer space" is constructed:

[0122] {m} = {P k (R k , C k ) | k = 1, 2,..., H}

[0123] Meanwhile, since the method of using the gray-scale baseline constant Th is adopted for the calculation of the center point, the influence of noise interference such as ambient light on the extraction accuracy of the optical slice center is suppressed.

[0124] The embodiments of the present invention have been described above in conjunction with the accompanying drawings. However, the present invention is not limited to the above specific embodiments. The above specific embodiments are merely illustrative rather than restrictive. Under the inspiration of the present invention, those of ordinary skill in the art can also make many forms without departing from the spirit and scope protected by the claims of the present invention. These all fall within the protection scope of the present invention.

Claims

1. A high-precision extraction method for the center of a light slice based on weighted dimensionality reduction decomposition, characterized in that: The following steps are involved: S1: Acquire a light strip pixel-level image, and obtain, according to the light strip pixel-level image, a dimensionality reduction decomposition direction of the "image layer space" of the light strip pixel-level image and a direction perpendicular to the dimensionality reduction decomposition direction of the "image layer space" of the light strip pixel-level image; S2: graying the light stripe pixel level image, decomposing the light stripe pixel level image according to the dimensionality reduction decomposition direction of the "image layer space" of the light stripe pixel level image, and obtaining the grayscale vector of the "image layer space" of the light stripe pixel level image; S3: according to the grayscale vector of the "image layer space" of the pixel-level image of the light strip, obtaining the sub-pixel coordinates of the decomposition direction of the grayscale vector of the "image layer space"; S4: According to the sub-pixel coordinates of the decomposition direction of the "image layer space" of the grayscale vector, the serial number of the vertical direction of the decomposition direction of the current grayscale vector is matched to obtain the center coordinates of the light slice corresponding to the grayscale vector; S5: traverse and calculate the light slice center coordinates corresponding to all grayscale vectors to obtain the light slice center of the light strip pixel-level image; Step S3 specifically includes: obtaining the sub-pixel coordinates of the decomposition direction of the "image layer space" of the grayscale vector according to the grayscale vector of the "image layer space" of the pixel-level image of the light strip, such as the formula: σ 2 (T)=R A (T)(V A (T)-V G ) 2 +R B (T)(V B (T)-V G ) 2 Among them, P k is the sub-pixel coordinate of the decomposition direction of the "image space" of the k-th grayscale vector, n is the number of sub-pixel coordinates in the retained area, P ki is the sub-pixel coordinate point of the i-th "image space" decomposition direction of the k-th grayscale vector, A ki is the weight corresponding to the gray value of the sub-pixel coordinate point in the i-th "image space" decomposition direction of the k-th gray vector, r ki is the gray value of the sub-pixel coordinate point in the i-th "image space" decomposition direction of the k-th gray vector, a is the first preset coefficient for calculating the gray value weight, b is the second preset coefficient for calculating the gray value weight, c is the third preset coefficient for calculating the gray value weight, Th is the gray baseline constant, T is the preset gray threshold, σ 2 (.) is the variance function, Max(.) is the maximum value function, V G is the total gray value of the entire image, R A (T) is the probability of a pixel being classified into class A, V A (T) is the average grayscale of pixels assigned to class A, R B (T) is the probability of a pixel being classified into class B, V B (T) is the average grayscale of the pixels assigned to class B, R i is the probability that the gray level of the pixel is i, g i is the number of pixels with gray level i in the image; Step S4 specifically includes: according to the sub-pixel coordinates of the decomposition direction of the "image layer space" of the grayscale vector, matching the serial number of the vertical direction of the decomposition direction of the current grayscale vector, and obtaining the light slice center coordinates corresponding to the grayscale vector, such as the formula: P k (R k ,C k )=(P k ,k) Among them, P k (R k ,C k ) is the center coordinate of the light slice corresponding to the kth grayscale vector, R k is the coordinate of the center decomposition direction of the light slice in the "image space", C k is the vertical coordinate of the decomposition direction of the light slice center in the "image space", P k is the sub-pixel coordinate of the decomposition direction of the "image layer space" of the kth grayscale vector, and k is the serial number in the vertical direction of the decomposition direction of the current grayscale vector.

2. The method for high-precision extraction of light slice centers based on weighted dimensionality reduction decomposition according to claim 1 is characterized in that: Step S1 specifically includes: S11: Acquire a light strip pixel-level image, and obtain a light slice main direction, an "image layer space" row direction, and an "image layer space" column direction of the light strip pixel-level image according to the light strip pixel-level image; S12: according to the main direction of the light slice, the row direction of the "image layer space" and the column direction of the "image layer space" of the light strip pixel-level image, obtaining the angle between the main direction of the light slice and the row direction of the "image layer space" and the angle between the main direction of the light slice and the column direction of the "image layer space"; S13: When the angle between the main direction of the light slice and the row direction of the "image layer space" is not less than the angle between the main direction of the light slice and the column direction of the "image layer space", the column direction of the light stripe pixel level image is taken as the dimensionality reduction decomposition direction of the "image layer space" of the light stripe pixel level image, and the row direction of the light stripe pixel level image is taken as the perpendicular direction of the dimensionality reduction decomposition direction of the "image layer space" of the light stripe pixel level image; when the angle between the main direction of the light slice and the row direction of the "image layer space" is less than the angle between the main direction of the light slice and the column direction of the "image layer space", the row direction of the light stripe pixel level image is taken as the dimensionality reduction decomposition direction of the "image layer space" of the light stripe pixel level image, and the column direction of the light stripe pixel level image is taken as the perpendicular direction of the dimensionality reduction decomposition direction of the "image layer space" of the light stripe pixel level image.

3. The method for extracting the center of a light slice with high precision based on weighted dimensionality reduction decomposition according to claim 1 is characterized in that: Step S5 specifically includes: traversing and calculating the light slice center coordinates corresponding to all grayscale vectors to obtain the light slice center of the light strip pixel-level image, as shown in the formula: {m}={Pk(Rk,Ck)|k=1,2,...,H} Wherein, {m} is the center of the light slice of the light bar pixel level image, and H is the number of grayscale vectors.

4. A system for extracting the center of a light slice, characterized in that The system includes the following modules: The image acquisition module is configured to: acquire the light strip pixel level image, and obtain the "image layer space" dimensionality reduction decomposition direction of the light strip pixel level image and the vertical direction of the "image layer space" dimensionality reduction decomposition direction of the light strip pixel level image according to the light strip pixel level image; A vector acquisition module is configured to: grayscale the light strip pixel level image, decompose the light strip pixel level image according to the dimensionality reduction decomposition direction of the "image layer space" of the light strip pixel level image, and obtain the grayscale vector of the "image layer space" of the light strip pixel level image; A decomposition coordinate module is configured to obtain the sub-pixel coordinates of the decomposition direction of the grayscale vector in the "image layer space" according to the grayscale vector of the "image layer space" of the pixel-level image of the light strip; A coordinate matching module is configured to: match the sub-pixel coordinates of the decomposition direction of the grayscale vector in the "image layer space" of the grayscale vector, and obtain the center coordinates of the light slice corresponding to the grayscale vector; The image light slice center module is configured as follows: traversing and calculating the light slice center coordinates corresponding to all grayscale vectors to obtain the light slice center of the light strip pixel-level image; The decomposition coordinate module is specifically configured as follows: according to the grayscale vector of the "image layer space" of the pixel-level image of the light strip, the sub-pixel coordinate of the decomposition direction of the grayscale vector of the "image layer space" is obtained, as shown in the formula: σ 2 (T)=R A (T)(V A (T)-V G ) 2 +R B (T)(V B (T)-V G ) 2 Among them, P k is the sub-pixel coordinate of the decomposition direction of the "image space" of the k-th grayscale vector, n is the number of sub-pixel coordinates in the retained area, P ki is the sub-pixel coordinate point of the i-th "image space" decomposition direction of the k-th grayscale vector, A ki is the weight corresponding to the gray value of the sub-pixel coordinate point in the i-th "image space" decomposition direction of the k-th gray vector, r ki is the gray value of the sub-pixel coordinate point in the i-th "image space" decomposition direction of the k-th gray vector, a is the first preset coefficient for calculating the gray value weight, b is the second preset coefficient for calculating the gray value weight, c is the third preset coefficient for calculating the gray value weight, Th is the gray baseline constant, T is the preset gray threshold, σ 2 (.) is the variance function, Max(.) is the maximum value function, V G is the total gray value of the entire image, R A (T) is the probability of a pixel being classified into class A, V A (T) is the average grayscale of pixels assigned to class A, R B (T) is the probability of a pixel being classified into class B, V B (T) is the average grayscale of the pixels assigned to class B, R i is the probability that the gray level of the pixel is i, g i is the number of pixels with gray level i in the image; The coordinate matching module is specifically configured as follows: according to the sub-pixel coordinates of the decomposition direction of the "image layer space" of the grayscale vector, the serial number of the vertical direction of the decomposition direction of the current grayscale vector is matched to obtain the light slice center coordinates corresponding to the grayscale vector, as shown in the formula: P k (R k ,C k )=(P k ,k) Among them, P k (R k ,C k ) is the center coordinate of the light slice corresponding to the kth grayscale vector, R k is the coordinate of the center decomposition direction of the light slice in the "image space", C k is the vertical coordinate of the decomposition direction of the light slice center in the "image space", P k is the sub-pixel coordinate of the decomposition direction of the "image layer space" of the kth grayscale vector, and k is the serial number in the vertical direction of the decomposition direction of the current grayscale vector.

5. The system for extracting the center of a light slice according to claim 4, characterized in that: The image acquisition module is specifically configured as follows: Acquire a light strip pixel-level image, and obtain a light slice main direction, an "image layer space" row direction, and an "image layer space" column direction of the light strip pixel-level image according to the light strip pixel-level image; According to the main direction of the light slice, the row direction of the "image layer space" and the column direction of the "image layer space" of the light strip pixel-level image, the angle between the main direction of the light slice and the row direction of the "image layer space" and the angle between the main direction of the light slice and the column direction of the "image layer space" are obtained; When the angle between the main direction of the light slice and the row direction of the "image layer space" is not less than the angle between the main direction of the light slice and the column direction of the "image layer space", the column direction of the light stripe pixel level image is taken as the dimensionality reduction decomposition direction of the "image layer space" of the light stripe pixel level image, and the row direction of the light stripe pixel level image is taken as the perpendicular direction of the dimensionality reduction decomposition direction of the "image layer space" of the light stripe pixel level image; when the angle between the main direction of the light slice and the row direction of the "image layer space" is less than the angle between the main direction of the light slice and the column direction of the "image layer space", the row direction of the light stripe pixel level image is taken as the dimensionality reduction decomposition direction of the "image layer space" of the light stripe pixel level image, and the column direction of the light stripe pixel level image is taken as the perpendicular direction of the dimensionality reduction decomposition direction of the "image layer space" of the light stripe pixel level image.

6. The system for extracting the center of a light slice according to claim 4, characterized in that: The image light slice center module is specifically configured as follows: traversing and calculating the light slice center coordinates corresponding to all grayscale vectors to obtain the light slice center of the light strip pixel-level image, as shown in the formula: {m}={Pk(Rk,Ck)|k=1,2,...,H} Wherein, {m} is the center of the light slice of the light bar pixel level image, and H is the number of grayscale vectors.

Images