A method and system for tracking a continuous groove based on a line structured light point cloud

Abstract

The present application belongs to the field of machine vision detection, and particularly discloses a track tracing method and system for continuous grooves based on line structured light point clouds. The method comprises the following steps: S100, using a line laser scanner to scan along the direction of the grooves to obtain point cloud data spliced by each laser line point cloud; S200, calculating features on each laser line, including features reflecting the size of height change and features of height mutation caused by horizontal lines; S300, judging groove feature points on the laser line according to a preset height difference range and mutation degree; and S400, searching from the starting point, forming continuous track points along the laser line from the starting point according to the continuity of the groove track and the feature points to obtain the track of the continuous grooves. The present application can realize track tracing at the same level as the sensor accuracy, has the advantages of high precision and real-time detection, and is suitable for track generation of various continuous grooves.

Description

Technical Field

[0001] This invention belongs to the field of machine vision, and more specifically, relates to a method and system for tracking the trajectory of continuous grooves based on line structured light point clouds. Background Technology

[0002] In industrial production and automated manufacturing, detecting and processing the trajectory of grooves (such as grooves formed at weld joints, grooves at product assembly joints, etc.) is an important task. Image-based processing methods suffer from insufficient image resolution and poor accuracy, occlusion at complex trajectory imaging positions, and poor robustness to deformation in multi-view template matching, which limits groove trajectory extraction and affects accuracy and reliability. Methods based on point cloud volume, centroid, PCA principal components, and the geometry of the circumscribed sphere rectangle struggle to accurately find the unique features that distinguish grooves from others. Methods based on feature descriptors such as PFH (Point Feature Histogram), FPFH (Fast Point Feature Histogram), and SHOT (Rotation Invariant Feature) suffer from high computational cost and poor real-time performance. Registration-based methods suffer from poor adaptability to deformation, require standard models, and lack generalization ability for similar features.

[0003] Therefore, there is an urgent need to propose a new method for generating the trajectory of continuous concave grooves, which can uniformly solve the problem of continuous groove trajectory extraction while taking into account both efficiency and accuracy. Summary of the Invention

[0004] To address the aforementioned deficiencies or improvement needs of existing technologies, this invention provides a method and system for tracking the trajectory of continuous grooves based on line structured light point clouds. The purpose is to solve the problems of low detection efficiency, lack of generalization, insufficient accuracy, and inability to meet the needs of automated production in existing technologies that generate continuous trajectories of concave grooves through teaching.

[0005] To achieve the above objectives, according to one aspect of the present invention, a trajectory tracking method for continuous grooves based on line structured light point clouds is provided, comprising the following steps:

[0006] Step S100: Use a line laser scanner to scan along the direction of the groove to obtain point cloud data composed of point clouds stitched together from each laser line point cloud;

[0007] Step S200: Calculate the features of each laser line, including features reflecting the magnitude of height changes and features indicating abrupt height changes in horizontal lines;

[0008] Step S300: Determine the groove feature points on the laser line according to the preset height difference range and abrupt change degree;

[0009] Step S400: Search from the starting point. Based on the continuity of the groove trajectory, based on the feature points, recursively form continuous trajectory points along the laser line from the starting point to obtain the trajectory of the continuous groove.

[0010] Further, step S200 includes:

[0011] Step S210: Calculate the height difference information of a laser line and calculate the peak point and peak value of the height difference information;

[0012] Step S220: Calculate the variance ratio information of a laser line and calculate the peak point and peak value of the variance ratio information.

[0013] Further, in step S210, the depth data along a laser line is first represented as a discrete sequence; for each point P on the laser line... i Calculate point P i Calculate the average depth of the left and right windows, and then find the absolute value ΔH of the difference between the average depths of the left and right windows. i ,Right now:

[0014] ΔH i =|mean(P i-n ,…,P i-1 )-mean(P i+1 ,…,P i+n )|

[0015] Where n represents the window size, mean() represents calculating the average depth, and P i-n P represents the point in the left window. i+n The point in the right window represents the peak point, which corresponds to the maximum value in the height difference information. The peak value is the height difference value corresponding to the peak point. By calculating the peak value of the height difference information sequence and its corresponding peak point index, the size of the height difference on the laser line is determined to reflect the contour of the groove.

[0016] In step S220, the sum of the variances of the left and right windows is calculated and compared with the variance of the bilateral windows, that is:

[0017]

[0018] Here, sum() represents summation, and Var() represents variance.

[0019] Further, step S300 includes:

[0020] Step S310: Traverse the peak index of height difference. If its height difference range is greater than or equal to the preset lower limit of height difference and less than or equal to the preset upper limit of height difference, and there exists a variance ratio peak index and the absolute difference of height difference peak index that is less than or equal to the preset variance and the maximum peak offset distance of height, then the height difference peak index is considered to be a groove point.

[0021] Step S320: The lower limit of variance search starts from the initial feature variance ratio and decreases sequentially according to the variance ratio decrease amount until the set of groove points is non-empty and the feature variance ratio is greater than or equal to the lower limit of the feature variance ratio.

[0022] Further, step S310 includes:

[0023] The initial feature variance ratio and variance ratio reduction are set. The initial feature variance ratio is set based on the pre-judgment of the groove imaging quality to ensure that normal grooves can be accurately identified under the condition of the initial feature variance ratio.

[0024] When the feature variance ratio of the search is greater than or equal to 1, if the groove point index was not found when searching the data on the laser line in the previous time, the feature variance ratio of the search is reduced by the decrease amount.

[0025] The search is performed based on the updated feature variance ratio. If the smallest feature variance ratio also fails to find the groove index, it is determined that there is no groove on this laser line.

[0026] Further, step S400 includes:

[0027] Step S410: Based on the initial point on the initial line, find the groove point on the next laser line within the search range;

[0028] Step S420: Accurately locate the groove point, update the trajectory point and search for the next laser line until the termination line; specifically, given the index of a point on the laser line, the search range of the next laser line includes the index of the previous point ± the search range parameter; whenever a groove is located on a new laser line, the trajectory point is rotated to the higher side of the groove so that it falls into the minimum point formed by the groove, and the index of this trajectory point is updated to the index of the previous point.

[0029] Further, step S420 includes:

[0030] Step S421: Take the left and right window points of the currently determined groove point, with the x-index of the current point being 0. Convert the depth sequence information into point cloud information of the actual interval, and represent it in the xz coordinate system:

[0031] x = idx * x_resolution

[0032] z = height

[0033] Where idx represents the index of the point on the current laser line, x_resolution represents the resolution of the point on the current laser line, and height represents the z-coordinate of the point on the current laser line.

[0034] Step S422: Rotate the sequence corresponding to xz by 45° around the origin towards the higher side of the groove, thus ensuring that the groove point becomes a local minimum point;

[0035] Step S423: Find the index k corresponding to the local minimum in the sequence corresponding to xz;

[0036] Step S424, update the current trajectory point index as follows:

[0037] idx′=idx-window+k

[0038] Where idx' represents the updated index of the point on the current laser line, and window represents the size of the window for precise positioning;

[0039] Step S425: Update the previous index to idx′.

[0040] Furthermore, it also includes step S500, which involves shifting the continuous trajectory points along a fixed direction by a certain distance or shifting them along the normal of the continuous trajectory points by a fixed distance.

[0041] Further, step S500 includes:

[0042] Calculate the curvature ρ and normal C of the current trajectory;

[0043]

[0044] Where m represents the number of points in the neighborhood of the current point, i = 1 to m, Pi represents the i-th point in the neighborhood of the current point, P0 represents the current point, x”, y”, z” represent the second derivatives of the xyz coordinates of the current point, x’, y’, z’ represent the first derivatives of the xyz coordinates of the current point, and T represents the transpose;

[0045] The trajectory point is offset by a fixed distance along the normal direction on the lower side of the groove to obtain new two-dimensional coordinates; on the depth map, the depth corresponding to the new two-dimensional coordinates is interpolated, and the new coordinates are used as the offset trajectory point.

[0046] According to another aspect of the invention, a trajectory tracking system for continuous grooves based on line structured light point clouds is provided, comprising a memory, a processor, and a computer program stored in the memory, the processor executing the computer program to implement the steps of the trajectory tracking method for continuous grooves based on line structured light point clouds as described in any of the preceding claims.

[0047] In summary, the technical solutions conceived in this invention, compared with the prior art, can achieve the following beneficial effects:

[0048] (1) The method of the present invention identifies the location of the abrupt change point by calculating the features that reflect the magnitude of the height change and the features that generate height abrupt changes in the horizontal line, thereby ensuring the accurate analysis of the depth information of each laser line and thus identifying the key feature points of the groove; then, by searching from the starting point, according to the continuity of the groove trajectory, the feature points can be used to form a continuous trajectory point by recursively following the laser line from the starting point, and a complete trajectory line can be tracked while scanning.

[0049] (2) The present invention ensures that the trajectory is continuously generated from the starting laser line to the ending laser line, and the density of the trajectory points is determined by the search interval. The search interval multiplied by the x-axis laser point resolution equals the density of the trajectory points. Therefore, the density of the trajectory points can be adjusted by adjusting the search interval, thereby achieving trajectory tracking at the same level as the sensor accuracy. It has the advantages of high accuracy and real-time detection, and is suitable for trajectory generation of various continuous grooves. Attached Figure Description

[0050] Figure 1 This is a schematic diagram summarizing the three-dimensional structure and algorithm of the groove provided in a preferred embodiment of the present invention;

[0051] Figure 2 This is a schematic diagram of a discrete data sequence formed by depth information on a laser line according to a preferred embodiment of the present invention;

[0052] Figure 3 This is provided by a preferred embodiment of the present invention. Figure 2 A schematic diagram showing the height difference information and peak height difference corresponding to discrete data;

[0053] Figure 4 This is provided by a preferred embodiment of the present invention. Figure 2 A schematic diagram showing the variance ratio information and peak variance ratio for discrete data;

[0054] Figure 5 This is a schematic diagram of the grooves on multiple laser lines provided in a preferred embodiment of the present invention;

[0055] Figure 6 This is a flowchart of the core algorithm provided in a preferred embodiment of the present invention;

[0056] Figure 7 This is a schematic diagram of the precision positioning groove point provided in a preferred embodiment of the present invention;

[0057] Figure 8 This is a flowchart of the complete process provided in a preferred embodiment of the present invention;

[0058] Figure 9 This is a diagram illustrating the effect of trajectory offset provided in a preferred embodiment of the present invention. Detailed Implementation

[0059] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0060] The rapid development of machine vision technology has greatly improved productivity, enabling more intelligent operation of auxiliary execution units. However, existing groove extraction strategies suffer from several problems: image processing-based methods suffer from poor accuracy due to insufficient image resolution, occlusion issues when capturing complex trajectories, and poor robustness to deformation in multi-view template matching, thus limiting the accuracy and reliability of groove trajectory extraction. Point cloud geometry-based methods, such as volume, centroid, PCA principal components, and circumscribed sphere rectangles, struggle to accurately find unique groove features; feature descriptor-based methods, such as PFH, FPFH, and SHOT, are computationally intensive and have poor real-time performance; while registration-based methods have poor adaptability to deformation, require standard models, and lack generalization ability for similar features.

[0061] To address the shortcomings of existing technologies, such as Figures 1-9 As shown, this invention provides a trajectory tracking method for continuous grooves based on line structured light point clouds, aiming to solve the problems of low detection efficiency, poor generalization, and insufficient accuracy when generating continuous groove trajectories through teaching. This method has advantages such as strong ability to handle deformation problems, high speed, high accuracy, and no need for standard model assistance, meeting the needs of automated production.

[0062] Step S100: Use a line laser scanner to scan along the direction of the groove to obtain point cloud data composed of point clouds stitched together from each laser line.

[0063] Specifically, in this embodiment, the process first requires aligning the scanner with the starting position of the groove and setting scanning parameters, including laser intensity, scanning speed, and resolution. During scanning, the line laser scanner moves along the direction of the concave groove, and the laser line sequentially scans each cross-section of the groove. When the laser line contacts the groove wall, reflection occurs, and these reflected lights are received by sensors mounted on the scanner. By processing the received reflected light signals, point cloud data corresponding to each laser line can be obtained. Point cloud data is a set of three-dimensional coordinate points, each point recording its specific position in space. To obtain complete point cloud data of the groove, the point cloud data obtained from scanning each laser line needs to be stitched together. Figure 5 The stitching process shown relies on a precise positioning system to ensure that the position and direction of each laser line are accurately recorded and aligned.

[0064] Step S200: Calculate the features of each laser line, including features reflecting the magnitude of height changes and features indicating abrupt height changes in horizontal lines.

[0065] To extract features from each laser line, each point is first processed as follows: Figure 2 As shown. Specifically, height difference information is obtained by calculating the difference between the average values ​​of the data in the left and right windows for each point, such as... Figure 3 As shown, this difference reflects the height variation of the point. Furthermore, the sum of the variances of the left and right windows for each point is calculated and compared to the ratio of the variances of both windows. This ratio is used to identify points of abrupt height changes, such as... Figure 4 As shown. Then, the peak value of the height difference information and the corresponding peak index are analyzed to determine the position of the point in the height change. Similarly, the peak value of the variance ratio and the corresponding peak index are calculated to identify the location of the abrupt change point. This process ensures accurate analysis of the depth information of each laser line, thereby identifying the key feature points of the groove.

[0066] In a preferred embodiment, step S200 includes:

[0067] Step S210: Calculate the height difference information of a laser line and calculate the peak point and peak value of the height difference information;

[0068] Step S220: Calculate the variance ratio information of a laser line and calculate the peak point and peak value of the variance ratio information.

[0069] Specifically, the depth data along a laser line is first represented as a discrete sequence. For each point P on the laser line... i We calculate the average depth of its left and right windows, and then find the absolute value of the difference between them, ΔH. i ,Right now:

[0070] ΔH i =|mean(P i-n ,…,P i-1 )-mean(P i+1 ,…,P i+n )|

[0071] Where n represents the window size, mean() represents calculating the average depth, and P i-n P represents the point in the left window. i+n This represents a point in the right window. A peak point corresponds to the maximum value in the height difference information, while the peak value is the height difference corresponding to that peak point. By calculating the peak value of the height difference information sequence and its corresponding peak point index, we can determine the magnitude of the height difference on the laser line, reflecting the contour of the groove.

[0072] Similarly, for calculating the variance ratio, we can use the variance ratio of the depth data within the window to measure height abrupt changes. Specifically, we calculate the sum of the variances of the left and right windows and compare it with the variance of the bilateral window, i.e.:

[0073]

[0074] Where n represents the size of the feature extraction window, sum() represents summation, and Var() represents variance calculation.

[0075] Step S300: Determine the groove feature points on the laser line based on the preset height difference range and abrupt change degree.

[0076] During the trajectory point generation process, we first set the range of groove height difference, namely the minimum and maximum height difference, as well as the range of groove variance ratio, including the upper and lower limits of the variance ratio. Simultaneously, we also preset the maximum peak offset distance for variance and height, taking into account the potential error between the actual groove height difference and the peak index of the variance ratio.

[0077] For each laser line's depth information sequence, the peak value of the height difference information is first identified, ensuring it falls within a preset height difference range. If this condition is met, the index corresponding to that peak value is marked as a possible groove point. Next, each possible trajectory point index is examined to determine if it is a true groove point index. This step involves comparing it with the variance ratio peak index. If the absolute difference between a variance ratio peak index and the peak index of the height difference information is less than the preset maximum offset distance between variance and height, then that possible groove point index is marked as a groove point index.

[0078] In a preferred embodiment, step S300 includes:

[0079] Step S310: Traverse the peak index of height difference. If its height difference range is greater than or equal to the preset lower limit of height difference and less than or equal to the preset upper limit of height difference, and there exists a variance ratio peak index and the absolute difference of height difference peak index that is less than or equal to the preset variance and the maximum peak offset distance of height, then the height difference peak index is considered to be a groove point.

[0080] Step S320: The lower limit of the variance search should be gradually decreased from the initial feature variance ratio by the decreasing variance ratio until the set of groove points is non-empty and the feature variance ratio is greater than or equal to the lower limit of the feature variance ratio.

[0081] More preferably, step S310 includes:

[0082] The initial feature variance ratio and variance ratio reduction are set. The initial variance ratio is set based on the pre-judgment of the groove imaging quality to ensure that normal grooves can be accurately identified under the condition of the initial feature variance ratio.

[0083] When the feature variance ratio of the search is greater than or equal to 1, if the groove point index was not found in the previous search of the laser line data, the feature variance ratio of the search is reduced by the decrease amount.

[0084] The search is performed based on the updated feature variance ratio. If the smallest feature variance ratio also fails to find the groove index, it is determined that there is no groove on this laser line.

[0085] Step S400: Search from the starting point. Based on the continuity of the groove trajectory, form continuous trajectory points by recursively following the laser line from the starting point according to the feature points.

[0086] This process ensures that the trajectory is continuously generated from the starting laser line to the ending laser line. The density of trajectory points is determined by the search interval. The search interval multiplied by the x-axis laser point resolution equals the density of trajectory points. The density of trajectory points can be adjusted by adjusting the search interval.

[0087] In a preferred embodiment, step S400 includes:

[0088] Step S410: Based on the initial point on the initial line, find the groove point on the next laser line within the search range;

[0089] Step S420: Precisely locate the groove point, update the trajectory point, and search for the next laser line until the termination line.

[0090] Specifically, given the index of a point on the laser line, the search range of the next laser line includes the previous point index ± the search range parameter. This parameter considers the continuity of the grooves; continuous groove lines ensure that the difference in groove point index values ​​between laser lines is less than or equal to the search range between lines. Whenever a groove is located on a new laser line, it is necessary to rotate the trajectory point towards the higher side of the groove so that it falls into the minimum point formed by the groove. At the same time, the index of this trajectory point is updated to the previous point index.

[0091] In a preferred embodiment, step S420 includes:

[0092] Step S421: Take the left and right window points of the currently determined groove point, with the x-index of the current point being 0. Convert the depth sequence information into point cloud information of the actual interval, and represent it in the xz coordinate system:

[0093] x = idx * x_resolution

[0094] z = height

[0095] Where idx represents the index of the point on the current laser line, x_resolution represents the resolution of the point on the current laser line, and height represents the z-coordinate of the point on the current laser line.

[0096] Step S422: Rotate the sequence corresponding to xz by 45° around the origin towards the higher side of the groove, thus ensuring that the groove point becomes a local minimum point;

[0097] Step S423: Find the index k corresponding to the local minimum in the sequence corresponding to xz;

[0098] Step S424, update the current trajectory point index as follows:

[0099] idx′=idx-window+k

[0100] Where idx' represents the updated index of the point on the current laser line, and window represents the size of the window for precise positioning;

[0101] Step S425, update the previous index to idx′, such as Figure 7 As shown, the precise positioning effect is demonstrated in detail.

[0102] These steps ensure that a continuous trajectory is generated on the laser line. The precise positioning method guarantees the accuracy of trajectory generation, the limitation of the search range ensures the continuity of the trajectory, and the setting of the search interval ensures that the density of the trajectory is controllable.

[0103] Step S500, according to actual needs, offset the trajectory by a fixed distance along a fixed direction or by a fixed distance along the normal of the trajectory, including:

[0104] Calculate the curvature ρ and normal C of the current trajectory;

[0105]

[0106]

[0107] Where m represents the number of points in the neighborhood of the current point, i = 1 to m, Pi represents the i-th point in the neighborhood of the current point, P0 represents the current point, x”, y”, z” represent the second derivatives of the xyz coordinates of the current point, x’, y’, z’ represent the first derivatives of the xyz coordinates of the current point, and T represents the transpose;

[0108] The trajectory point is offset by a fixed distance along the normal direction on the lower side of the groove to obtain new two-dimensional coordinates; on the depth map, the depth corresponding to the new two-dimensional coordinates is interpolated, and the new coordinates are used as the offset trajectory point.

[0109] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A trajectory tracking method for continuous grooves based on line structured light point cloud, characterized in that, The method comprises the following steps: Step S100, scanning along the direction of the groove using a line laser scanner to obtain point cloud data spliced from each laser line point cloud; Step S200 ， Features on each laser line are calculated, including features that reflect the magnitude of height changes and features that produce height discontinuities in the horizontal line. Step S300 ， According to the preset height difference range and mutation degree, the groove feature points on the laser line are judged. Step S400, searching from the starting point, forming a continuous trajectory point from the starting point along the laser line according to the continuity of the trajectory of the groove according to the feature point to obtain the trajectory of the continuous groove; Step S400 comprises: Step S410, searching for a groove point on the next laser line in the search range according to the initial point on the initial line; Step S420, accurately positioning the groove point, updating the trajectory point to search for the next laser line until the terminal line; specifically, given the index of a point on the laser line, the search range of the next laser line includes the index of the previous point ± a search range parameter; whenever a groove is located on a new laser line, the trajectory point is made to fall into the minimum point formed by the groove by rotating to the higher side of the groove, and at the same time, the index of this trajectory point is updated to the index of the previous point; Step S420 comprises: Step S421, taking the current judged groove point left and right window point, the x index of the current point is 0, converting the depth sequence information into actual interval point cloud information, in Coordinate system representation: Wherein, idx represents the index corresponding to the point on the current laser line, represents the resolution of the point on the current laser line, and height represents the z coordinate value of the point on the current laser line. Step S422, the The corresponding sequence is rotated by 45° around the origin to the higher side of the groove, so that the groove point becomes a local minimum point. Step S423, searching for the index k corresponding to the local minimum value in the sequence corresponding to x-z; Step S424, updating the current trajectory point index to: Wherein, idx' represents the updated index of the point on the current laser line, and window represents the window size of accurate positioning; Step S425, update the last point index as .

2. The method of claim 1, wherein, Step S200 comprises: Step S210, calculating the height difference information of a laser line and calculating the peak point and peak value of the height difference information; Step S220, calculating the variance ratio information of a laser line and calculating the variance ratio information In step S220, the sum of the variances of the left and right windows is calculated and compared with the variance of the bilateral window, that is: Peak point and peak value.

3. The method of claim 2, wherein, In step S210, the depth data along a laser line is first represented as a discrete sequence; for each point on the laser line... Calculation points Calculate the average depth of the left and right windows, and then find the absolute value of the difference between the average depths of the left and right windows. ,Right now: wherein n represents the size of the window, mean() represents the average value of the depth, a point representing a left window, a point representing a right window; the peak point corresponds to the maximum value in the height difference information, and the peak value is the difference value corresponding to the peak point; by calculating the peak value of the height difference information sequence and the index of the peak point corresponding thereto, the size of the height difference on the laser line is determined to reflect the profile of the groove; Wherein, sum() represents summation, and Var() represents variance. Step S300 comprises:

4. The method of claim 3, wherein, Step S310, traversing the height difference peak index, if the difference range is greater than or equal to the preset lower limit of the height difference and less than or equal to the preset upper limit of the height difference, and there is a variance ratio peak index whose absolute difference with the height difference peak index is less than or equal to the preset maximum peak distance of the sum of the variances and the height, it is considered that the height difference peak index is a groove point; Step S320, the lower limit of the variance search is sequentially decreased by a variance ratio decrement from the initial feature variance ratio until the groove point set is non-empty and the feature variance ratio is greater than or equal to the lower limit of the feature variance ratio. Step S310 comprises:

5. The method of claim 4, wherein, Setting the initial feature variance ratio and the variance ratio decrement, the setting of the initial feature variance ratio is based on the pre-judgment of the imaging quality of the groove, so as to ensure that the normal groove can be accurately identified under the condition of the initial feature variance ratio; When the searched feature variance ratio is greater than or equal to 1, if no groove point index is found when searching the data on the laser line last time, the searched feature variance ratio is decreased by the decrement; Searching according to the updated search feature variance ratio, if the smallest feature variance ratio cannot find the groove point index, it is judged that there is no groove point on this laser line. Further comprising step S500, offsetting the continuous trajectory point by a distance along a fixed direction or offsetting the continuous trajectory point by a fixed distance along the normal direction of the continuous trajectory point.

6. The method of claim 1, wherein, Step S500 comprises:

7. The method of claim 6, wherein, ​ calculating a curvature of the current trajectory and normal ; wherein, m denotes the number of points within the neighborhood of the current point, i =1~m, P i denotes the i-th point within the neighborhood of the current point, P 0denotes the current point, x’’ , y’’ , z’’ denote the second derivatives of the xyz coordinates of the current point, respectively, x’ , y’ , z’ denote the first derivatives of the xyz coordinates of the current point, respectively, and T denotes the transpose. The trajectory point is offset by a fixed distance along the normal direction along the lower side of the groove to obtain a new two-dimensional coordinate; and the depth corresponding to the new two-dimensional coordinate is interpolated on the depth map, and the new coordinate is taken as the offset trajectory point. 8.A trajectory tracking system for continuous grooves based on line structured light point cloud, comprising a memory, a processor and a computer program stored on the memory, characterized in that, The processor executes the computer program to implement the steps of the trajectory tracking method based on the line structured light point cloud continuous groove of any one of claims 1-7.

Images