Autonomous path planning method and system for steel bar binding robot, storage medium and product

Through the depth filtering algorithm and YOLOv5 network identifying the intersection of steel bars, designing a low-complexity path planning algorithm, solving the problem of time-consuming and low efficiency of steel bar binding robot path planning, and achieving efficient and intelligent steel bar binding operations to adapt to different environments.

CN119085649APending Publication Date: 2024-12-06HUAZHONG UNIV OF SCI & TECH
View PDF 0 Cites 7 Cited by

Patent Information

Application Number
CN202411198002.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-29
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

The independent path planning of existing reinforced bar binding robots is time-consuming and low-efficiency, and has poor adaptability to different environments, making it difficult to achieve intelligence.

Method used

The deep filtering algorithm is used to process the steel mesh images, combine the YOLOv5 network to identify and locate the unbound steel bar intersections, design a low-complexity path planning algorithm, and plan the binding path through positive and reverse order methods to achieve autonomous and intelligent steel bar binding by robots.

Benefits of technology

The binding efficiency of steel bar binding robots has been greatly improved, the path planning time has been reduced, high adaptability and automated operations to different environments have been achieved, and labor costs have been reduced.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119085649A_ABST
    Figure CN119085649A_ABST
Patent Text Reader

Abstract

The invention belongs to the field of steel bar binding, and particularly discloses a steel bar binding robot autonomous path planning method and system, a storage medium and a product, and the method comprises the following steps: processing an initial steel bar mesh image through a depth filtering algorithm, and obtaining a surface layer steel bar mesh image; visually identifying and classifying bound and unbound reinforcing steel bar cross points in the surface layer reinforcing steel bar mesh image, and determining coordinates of the unbound reinforcing steel bar cross points; and finally, performing path planning through an algorithm to obtain a reasonable steel bar binding path. The invention can provide a reasonable and efficient autonomous path planning method for a reinforcing steel bar binding robot used in building construction, and the robot can autonomously bind reinforcing steel bar intersection points along a reasonable path on a construction site of a reinforcing steel bar project by using the autonomous path planning method; the planning speed is high, the path is reasonable, the intelligent degree and the working efficiency of the steel bar binding robot can be improved, and the labor cost is greatly reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of steel bar binding, and more specifically, relates to an autonomous path planning method, system, storage medium and product of a steel bar binding robot. Background Art

[0002] At present, the steel bar engineering of construction is still highly dependent on manual work, which is labor-intensive, inefficient, and risky. It cannot be mass-produced and industrialized. These problems seriously restrict the development of the construction industry. Steel bar binding is one of the important processes in reinforced concrete construction technology. When vibrating concrete, in order to ensure that the steel bars do not shift and the force position remains unchanged, the steel bars need to be tied and fixed. When workers perform binding operations, they will first lay the steel bars as required; then manually use binding hooks or steel bar binding guns to bind the cross parts. This process requires workers to bend over for a long time, and the labor intensity of workers is high. Long-term work is prone to fatigue, and problems such as unreliable binding and missing binding will occur, affecting the quality of concrete structures. Therefore, using intelligent steel bar binding robots to replace workers to perform operations is expected to be one of the effective ways to solve the above problems.

[0003] However, the current autonomous path planning of the steel bar tying robot during the tying process is time-consuming, inefficient, poorly adaptable to different environments, and requires the operator to have a certain programming foundation, making it difficult to meet the requirements of intelligence. Therefore, studying a method for autonomous path planning of steel bar tying robots is of great significance to the transformation, upgrading and development of industries such as construction, improving the production efficiency and quality of engineering construction, and reducing the labor intensity of workers. Summary of the invention

[0004] In response to the above defects or improvement needs of the prior art, the present invention provides a method, system, storage medium and product for autonomous path planning of a rebar tying robot, with the aim of enabling the robot to autonomously and intelligently perform tying work at rebar intersections, thereby solving the problem that current tying path planning is time-consuming and inefficient.

[0005] To achieve the above object, according to a first aspect of the present invention, a method for autonomous path planning of a steel bar tying robot is proposed, comprising the following steps:

[0006] S1, processing the initial steel mesh image by a depth filtering algorithm to obtain a surface steel mesh image;

[0007] S2, identifying the intersection points of the tied and untied steel bars in the surface steel mesh image, and determining the coordinates of the intersection points of the untied steel bars to form an array of points to be tied;

[0008] S3. Planning the binding path for the binding point array, including:

[0009] S31, starting from the first lashing point in the first row of the lashing point array, lashing all lashing points in the first row in positive order;

[0010] S32, respectively calculating the distance between the end lashing position of the current row and the first and last lashing points of the next row, and taking the lashing point with the smaller distance as the starting lashing point of the next row;

[0011] S33, taking the next row as the current row, if the starting tying point of the current row is the first tying point to be tied in the current row, tying all the tying points to be tied in the current row in forward order; if the starting tying point of the current row is the last tying point to be tied in the current row, tying all the tying points to be tied in the current row in reverse order;

[0012] S34. Repeat steps S32 and S33 until the binding path planning for all the binding points is completed.

[0013] As a further preferred embodiment, step S2, identifying the intersections of tied and untied steel bars in the surface steel mesh image, comprises:

[0014] The pre-trained weight model is used to identify the intersections of tied and untied steel bars in the surface steel mesh image, and an identification box is drawn at the identification position.

[0015] As further preferred, the pre-training process of the weight model includes:

[0016] Constructing a training set: marking the intersections of tied and untied steel bars on the pre-acquired surface steel mesh image, and drawing an identification box with the intersections of the steel bars as the center;

[0017] The YOLOv5 network is trained using the training set, and the trained YOLOv5 network is used as a weight model.

[0018] As further preferred, step S2, determining the coordinates of the intersection points of the untied steel bars, comprises:

[0019] For the intersection of untied steel bars, the average value of the coordinates of the upper left corner and the lower right corner of the corresponding identification frame is calculated to obtain the coordinates of the center point of the identification frame, which is the coordinates of the intersection of the untied steel bars.

[0020] As a further preferred embodiment, step S1, processing the steel mesh image by a deep filtering algorithm, comprises:

[0021] The pixels in the initial steel mesh image whose depth values ​​are greater than the set threshold are set to black for filtering, thereby obtaining an image that only retains the surface steel mesh.

[0022] According to a second aspect of the present invention, there is provided a rebar tying robot autonomous path planning system, comprising a processor, wherein the processor is used to execute the above-mentioned rebar tying robot autonomous path planning method.

[0023] According to a third aspect of the present invention, there is provided a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the above-mentioned autonomous path planning method for a steel bar tying robot.

[0024] According to a fourth aspect of the present invention, there is provided a computer program product, which includes a computer program, and when the computer program is executed by a processor, the above-mentioned autonomous path planning method for a steel bar tying robot is implemented.

[0025] In general, the above technical solution conceived by the present invention has the following technical advantages compared with the prior art:

[0026] 1. The present invention obtains the surface steel bar image by deeply filtering the steel mesh image, quickly eliminates the interference of the lower steel bars on visual recognition, and then locates the untied steel bar intersections. Finally, a reasonable and fast steel bar tying path planning is given, so that the robot can autonomously and intelligently perform the tying work of the steel bar intersections, solving the problems of current path planning that is time-consuming, inefficient, poorly adaptable to different environments, and difficult to achieve intelligence.

[0027] 2. The present invention designs a low-complexity path planning algorithm, which plans the most reasonable steel bar tying path at a very fast speed. The path planning efficiency is very high and the time consumption is short, which greatly improves the tying efficiency of the steel bar tying robot. At the same time, the total time consumed by the robot for tying steel bars along the path is minimized, and the tying efficiency is the highest.

[0028] 3. The present invention automates the entire process of deep filtering of steel mesh, intelligent identification, classification and positioning of steel bar intersections based on the YOLOv5 algorithm, and path planning, without the need for manual programming or intervention according to the working environment, thereby replacing the need for manual operation and programming, improving the robot binding efficiency and environmental adaptability, and reducing labor costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 This is a diagram showing the effect of depth filtering of a steel mesh in path planning in an embodiment of the present invention;

[0030] Figure 2 This is a rendering of the effect of identifying and classifying steel bar intersections in path planning in an embodiment of the present invention;

[0031] Figure 3 This is a rendering of the effect of locating the intersection of steel bars in the path planning of an embodiment of the present invention;

[0032] Figure 4 It is a path planning effect diagram of a steel mesh screen without tied points in the path planning of an embodiment of the present invention;

[0033] Figure 5 It is a path planning effect diagram of a steel mesh screen containing tied points in the path planning of an embodiment of the present invention;

[0034] Figure 6 This is a flow chart of the autonomous path planning method of the steel bar tying robot according to an embodiment of the present invention. DETAILED DESCRIPTION

[0035] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0036] An embodiment of the present invention provides an autonomous path planning method for a steel bar tying robot, such as Figure 6 As shown, the following steps are included:

[0037] S1. Steel mesh image depth filtering: The initial steel mesh image is processed by a depth filtering algorithm to obtain a surface steel mesh image that only retains the surface steel bars, such as Figure 1 shown.

[0038] Specifically, the depth filtering algorithm is as follows: pixels whose depth values ​​are greater than a set threshold are set to black for filtering, and only the surface layer steel mesh is retained to eliminate the interference of the lower layer steel bars on visual recognition.

[0039] S2. Use the YOLOv5 algorithm to identify and classify steel bar binding nodes: Use the weight model to identify the intersections of bound and unbound steel bars in the surface steel mesh image, and draw an identification box at the identification position, such as Figure 2 As shown; then determine the coordinates of the intersection points of the untied steel bars to form an array of points to be tied.

[0040] Specifically, the pre-training process of the weight model includes:

[0041] Construct a training set: pre-acquire the surface steel mesh image in the same way as in S1, mark the intersections of the tied and untied steel bars thereon, and draw an identification box with the intersections of the steel bars as the center. Train the YOLOv5 network using the training set, and use the trained YOLOv5 network as the weight model.

[0042] The weight model file can be used repeatedly to identify and classify steel bar intersections, and when the quality of the imported training photos and their annotations is good, the recognition and classification accuracy of the weight model file can be higher than 99%.

[0043] Specifically, the coordinates of the intersection points of the untied steel bars are determined to form an array of points to be tied, including:

[0044] According to the coordinates of the upper left corner and the lower right corner of the identification frame, the coordinates of the center point of the identification frame are obtained by calculating the average value of the two, which is the positioning coordinates of the corresponding steel bar intersection point, such as Figure 3 shown.

[0045] When implemented by an algorithm, the intersections need to be stored in an array, so the number needs to be calculated based on the coordinates of the intersections to determine the sequence number stored in the array. In this embodiment, the untied steel bar intersections are numbered in the form of (x, y), where x is the sequence number of the row where the steel bar intersection is located (increasing from bottom to top), and y is the sequence number of the column where the steel bar intersection is located (increasing from left to right), so that each untied steel bar intersection has a unique number. In this way, the bottom row of steel bar intersections in the image can be the first row, because the tying robot is generally closer to the bottom row, so the bottom row is used as the first row for tying.

[0046] More specifically, when numbering, the horizontal coordinate of the intersection is divided by an interval (roughly the number of pixels between adjacent steel bars in the image) and used as the y in the number (x, y), that is, the serial number of the column where the intersection is located. The vertical coordinate of the intersection is taken as the opposite number and divided by the same interval, and then added to the total number of rows of steel bars in the picture as the x in the number (x, y), that is, the serial number of the row where the intersection is located.

[0047] S3. Planning the binding path for the binding point array, including:

[0048] S31, starting from the first lashing point in the first row of the lashing point array, lashing all lashing points in the first row in positive order;

[0049] S32, respectively calculating the distance between the end lashing position of the current row and the first and last lashing points of the next row, and taking the lashing point with the smaller distance as the starting lashing point of the next row;

[0050] S33, taking the next row as the current row, if the starting tying point of the current row is the first tying point to be tied in the current row, then tying all the tying points to be tied in the current row in positive order (i.e., from the first tying point to be tied in the current row to the last tying point); if the starting tying point of the current row is the last tying point to be tied in the current row, then tying all the tying points to be tied in the current row in reverse order (i.e., from the last tying point to be tied in the current row to the first tying point);

[0051] S34. Repeat steps S32 and S33 until all the binding points to be bound are completed, corresponding to the path of the steel bar binding gun and the actions to be performed, and the path planning is completed. Figure 4 This is the path planning result when the steel bar intersections are not tied. Figure 5 This is the path planning result when there are already tied points. Since the tying robot is in the middle row at the beginning, it is moved to the first row first. Figure 4 and Figure 5 In the figure, cp indicates the point to be bound, and bd indicates the point already bound.

[0052] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An autonomous path planning method for a steel bar tying robot. It is characterized in that The steps include: S1, processing the initial steel mesh image by a depth filtering algorithm to obtain a surface steel mesh image; S2, identifying the intersection points of the tied and untied steel bars in the surface steel mesh image, and determining the coordinates of the intersection points of the untied steel bars to form an array of points to be tied; S3. Planning the binding path for the binding point array, including: S31, starting from the first lashing point in the first row of the lashing point array, lashing all lashing points in the first row in positive order; S32, respectively calculating the distance between the end lashing position of the current row and the first and last lashing points of the next row, and taking the lashing point with the smaller distance as the starting lashing point of the next row; S33, taking the next row as the current row, if the starting tying point of the current row is the first tying point to be tied in the current row, tying all the tying points to be tied in the current row in forward order; if the starting tying point of the current row is the last tying point to be tied in the current row, tying all the tying points to be tied in the current row in reverse order; S34. Repeat steps S32 and S33 until the binding path planning for all the binding points is completed.

2. The autonomous path planning method for a steel bar tying robot according to claim 1, It is characterized in that Step S2, identifying the intersections of tied and untied steel bars in the surface steel mesh image, includes: The pre-trained weight model is used to identify the intersections of tied and untied steel bars in the surface steel mesh image, and an identification box is drawn at the identification position.

3. The autonomous path planning method for a steel bar tying robot according to claim 2, It is characterized in that The pre-training process of the weight model includes: Constructing a training set: marking the intersections of tied and untied steel bars on the pre-acquired surface steel mesh image, and drawing an identification box with the intersections of the steel bars as the center; The YOLOv5 network is trained using the training set, and the trained YOLOv5 network is used as a weight model.

4. The autonomous path planning method for a steel bar tying robot according to claim 2, It is characterized in that Step S2, determining the coordinates of the intersection points of the untied steel bars, comprises: For the intersection of untied steel bars, the average value of the coordinates of the upper left corner and the lower right corner of the corresponding identification frame is calculated to obtain the coordinates of the center point of the identification frame, which is the coordinates of the intersection of the untied steel bars.

5. The autonomous path planning method for a steel bar tying robot according to any one of claims 1 to 4, It is characterized in that Step S1, processing the steel mesh image by a deep filtering algorithm, including: The pixels in the initial steel mesh image whose depth values ​​are greater than the set threshold are set to black for filtering, thereby obtaining an image that only retains the surface steel mesh.

6. An autonomous path planning system for a steel bar tying robot, It is characterized in that It includes a processor, and the processor is used to execute the autonomous path planning method of the steel bar binding robot as described in any one of claims 1-5.

7. A computer-readable storage medium having a computer program stored thereon, It is characterized in that When the computer program is executed by the processor, the autonomous path planning method for the steel bar binding robot is implemented as described in any one of claims 1-5.

8. A computer program product, It is characterized in that It includes a computer program, which, when executed by a processor, implements the autonomous path planning method for a steel bar binding robot as described in any one of claims 1 to 5.

Citation Information

Cited By

  • Steel bar binding method, device and equipment and storage medium

    CN120071029A

  • Mechanical arm steel bar binding point visual identification method based on multi-algorithm fusion

    CN120472133A

  • A Visual Recognition Method for Rebar Binding Points in Robotic Arms Based on Multi-Algorithm Fusion

    CN120472133B

  • Steel bar binding machine operation control method and device based on intelligent path planning

    CN120993824A

  • Automatic steel bar binding machine based on machine vision

    CN121187170A