Methods and apparatus for robots to autonomously enter packaging containers

Abstract

The application relates to a method, device, computer equipment and storage medium for robot autonomous entry into a packaging container. The method comprises the following steps: acquiring inner wall positioning information set of the packaging container, acquiring packaging container position information by using a positioning label recognition algorithm based on the inner wall positioning information set, and controlling the robot to move to the position of the packaging container according to the packaging container position information; acquiring contour data set of the packaging container, acquiring a packaging container relative position calibration set according to the contour data set by using a clustering algorithm and a feature matching algorithm, and controlling the robot to enter the packaging container according to the packaging container relative position calibration set; acquiring bottom positioning information set of the packaging container, acquiring packaging container pose information by using a positioning label recognition algorithm based on the bottom positioning information set, and controlling the robot to adjust the attitude of the robot according to the packaging container pose information. By using the method, a more efficient, intelligent and automatic robot packaging process can be realized.

Description

Technical Field

[0001] This application relates to the field of robotics, and in particular to a method, apparatus, computer equipment, and storage medium for a robot to autonomously enter a packaging container. Background Technology

[0002] Traditional quadruped robot packaging processes typically require manual intervention and operation, resulting in low packaging efficiency and susceptibility to human error. Furthermore, robots are usually large and heavy, making them difficult to handle and deploy, further increasing the time and complexity of the packaging process. Therefore, improving packaging efficiency and quality, and developing automated and intelligent systems, has become particularly important. Summary of the Invention

[0003] Therefore, it is necessary to provide a method, apparatus, computer equipment, and storage medium for a robot to autonomously enter a packaging container to address the above-mentioned technical problems.

[0004] In a first aspect, this application provides a method for a robot to autonomously enter a packaging container, the method comprising:

[0005] The robot obtains the inner wall positioning information set of the packaging container, uses a positioning tag recognition algorithm based on the inner wall positioning information set to obtain the packaging container's position information, and controls the robot to move to the location of the packaging container according to the packaging container's position information.

[0006] Obtain the contour dataset of the packaging container, and use clustering and feature matching algorithms to obtain the relative position calibration set of the packaging container based on the contour dataset. Control the robot to enter the packaging container according to the relative position calibration set of the packaging container.

[0007] The robot obtains the bottom positioning information set of the packaging container, uses a positioning tag recognition algorithm based on the bottom positioning information set to obtain the pose information of the packaging container, and controls the robot to adjust its own posture according to the pose information of the packaging container.

[0008] In one embodiment, obtaining the location information of the packaging container using a positioning tag recognition algorithm based on the inner wall positioning information set includes:

[0009] Preprocess the inner wall positioning information set;

[0010] Feature extraction is performed based on the preprocessed inner wall positioning information set to obtain a feature point set;

[0011] Pose calculation is performed based on the feature point set to obtain the location information of the packaging container.

[0012] In one embodiment, preprocessing the inner wall positioning information set includes:

[0013] A denoising algorithm is used to filter the inner wall positioning information set, and the filtered inner wall positioning information set is then binarized.

[0014] In one embodiment, feature extraction based on the preprocessed inner wall positioning information set to obtain a feature point set includes:

[0015] The preprocessed inner wall positioning information set is subjected to contour detection to filter out the positioning label contour set;

[0016] The pre-generated set of location tag templates is matched and fitted with the set of location tag contours to obtain the set of fitted location tag contours.

[0017] The feature point set is obtained based on the contour fitting set of the positioning label.

[0018] In one embodiment, calculating the pose based on the feature point set to obtain the packaging container position information includes:

[0019] The feature point set is paired with the corresponding point set in 3D space, and the pose estimation algorithm is called to obtain the robot pose dataset. The position information of the packaging container is obtained based on the robot pose dataset.

[0020] In one embodiment, obtaining the relative position calibration set of the packaging container based on the contour dataset using clustering and feature matching algorithms includes:

[0021] The contour dataset is filtered, and a point cloud clustering algorithm is used to divide the filtered contour dataset into different clusters, and the contour cluster set is extracted.

[0022] An iterative nearest point algorithm is used to match a pre-established set of packaging container outline templates with a set of outline clusters to obtain a set of relative position calibrations for the packaging containers.

[0023] Secondly, this application also provides a device for a robot to autonomously enter a packaging container, the device comprising:

[0024] The packaging container positioning module is used to acquire the inner wall positioning information set of the packaging container, and to use a positioning tag recognition algorithm to obtain the position information of the packaging container based on the inner wall positioning information set, and to control the robot to move to the position of the packaging container according to the position information of the packaging container.

[0025] The packaging container entry module is used to obtain the contour dataset of the packaging container. It uses clustering and feature matching algorithms to obtain the relative position calibration set of the packaging container based on the contour dataset, and controls the robot to enter the packaging container according to the relative position calibration set of the packaging container.

[0026] The packaging container pose adjustment module is used to obtain the bottom positioning information set of the packaging container, and to obtain the pose information of the packaging container using a positioning tag recognition algorithm based on the bottom positioning information set, thereby controlling the robot to adjust its own posture according to the pose information of the packaging container.

[0027] Thirdly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to perform the following steps:

[0028] Obtain the inner wall positioning information set of the packaging container; use a positioning tag recognition algorithm based on the inner wall positioning information set to obtain the packaging container's position information; control the robot to move to the location of the packaging container according to the packaging container's position information;

[0029] Obtain the contour dataset of the packaging container; use clustering and feature matching algorithms to obtain the relative position calibration set of the packaging container based on the contour dataset; control the robot to enter the packaging container according to the relative position calibration set of the packaging container;

[0030] Obtain the bottom positioning information set of the packaging container; use a positioning tag recognition algorithm based on the bottom positioning information set to obtain the pose information of the packaging container; control the robot to adjust its own posture according to the pose information of the packaging container.

[0031] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, performs the following steps:

[0032] Obtain the inner wall positioning information set of the packaging container; use a positioning tag recognition algorithm based on the inner wall positioning information set to obtain the packaging container's position information; control the robot to move to the location of the packaging container according to the packaging container's position information;

[0033] Obtain the contour dataset of the packaging container; use clustering and feature matching algorithms to obtain the relative position calibration set of the packaging container based on the contour dataset; control the robot to enter the packaging container according to the relative position calibration set of the packaging container;

[0034] Obtain the bottom positioning information set of the packaging container; use a positioning tag recognition algorithm based on the bottom positioning information set to obtain the pose information of the packaging container; control the robot to adjust its own posture according to the pose information of the packaging container.

[0035] The aforementioned method, apparatus, computer equipment, and storage medium for a robot to autonomously enter a packaging container involve: acquiring a set of positioning information from the inner wall of the packaging container; using a positioning tag recognition algorithm based on the inner wall positioning information to obtain the packaging container's position information; controlling the robot to move to the location of the packaging container according to the packaging container's position information; acquiring a contour dataset of the packaging container; using clustering and feature matching algorithms to obtain a relative position calibration set of the packaging container based on the contour dataset; controlling the robot to enter the packaging container according to the relative position calibration set of the packaging container; acquiring a set of bottom positioning information from the packaging container; using a positioning tag recognition algorithm based on the bottom positioning information to obtain the packaging container's pose information; and controlling the robot to adjust its own posture according to the packaging container's pose information. This reduces the need for manual adjustments, improves the accuracy and speed of the packaging process, and achieves a more efficient, intelligent, and automated packaging process. Attached Figure Description

[0036] Figure 1 This is a flowchart illustrating a method for a robot to autonomously enter a packaging container in one embodiment;

[0037] Figure 2 The right and front views are of the robot and packaging container in one embodiment;

[0038] Figure 3 This is a left view of the robot and packaging container in one embodiment;

[0039] Figure 4 This is a schematic diagram of the robot and packaging container in the far-field state in one embodiment;

[0040] Figure 5 This is a schematic diagram of the robot and packaging container in a near-field state in one embodiment;

[0041] Figure 6 This is a schematic diagram of the packaging container in a packaged state according to one embodiment;

[0042] Figure 7 This is a schematic diagram of a robot lying prone inside a packaging container in one embodiment.

[0043] Figure 8 This is a flowchart illustrating a method for a robot to autonomously enter a packaging container in another embodiment;

[0044] Figure 9 This is a schematic diagram of a module for a device in one embodiment that allows a robot to autonomously enter a packaging container;

[0045] Figure 10 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0046] Traditional quadruped robot packaging processes typically require manual intervention and operation, including manual handling and positioning of the robot. This results in low packaging efficiency and a high risk of human error. Furthermore, the robots are usually large and heavy, making them difficult to handle and deploy, further increasing the time and complexity of the packaging process. Therefore, improving packaging efficiency and quality, and developing automated and intelligent systems, has become particularly important.

[0047] The robot in this application uses a monocular camera on its head to identify positioning tags on the inner wall of the packaging box, thereby autonomously triggering its entry into the box. Combined with a LiDAR scanner on its back that scans the box's contour, it achieves the ability to autonomously enter the box. To ensure precise positioning, the robot is equipped with a monocular camera on its underside to identify positioning tags on the bottom of the box. By aligning with these tags, the robot can accurately position itself inside the box. A grooved guide structure is also designed on the bottom of the box to guide the robot into the fixed groove even if there is a positional deviation, improving the accuracy and reliability of the packaging. The packaging lid has a fixed structure; when the lid is closed, the fixed structure on the lid fits tightly against the robot's back, providing stable support and ensuring the safety and reliability of the packaging. Compared to traditional robotic packaging systems, this reduces the need for manual adjustments, improves the accuracy and speed of the packaging process, and enables a more efficient, intelligent, and automated packaging process.

[0048] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0049] In one embodiment, such as Figure 1 As shown, a method for a robot to autonomously enter a packaging container is provided. This embodiment illustrates the application of this method to a terminal. It is understood that this method can also be applied to a server, or to a system including both a terminal and a server, and implemented through interaction between the terminal and the server. Specifically, the target robot is a quadrupedal robot dog. Of course, the target robot can also be a humanoid robot or other multi-legged robots, etc. In this embodiment, the method includes the following steps:

[0050] Step 102: Obtain the inner wall positioning information set of the packaging container, and use the positioning tag recognition algorithm to obtain the position information of the packaging container based on the inner wall positioning information set. Control the robot to move to the location of the packaging container according to the position information of the packaging container.

[0051] For example, the robot includes a quadruped robot, and the packaging container includes a box and a carton.

[0052] In one embodiment, the robot's head is equipped with a first camera device. Through this camera, the robot can acquire image information of positioning tags on the inner wall of the packaging container, i.e., the positioning information set of the inner wall of the packaging container. Further, the robot uses a positioning tag recognition algorithm to perform image processing and analysis on the inner wall positioning information set to obtain the position and orientation information of the positioning tags, thereby achieving the recognition, positioning, and orientation calculation of the positioning tags, providing a basis for the robot to enter and align with the packaging container. The first camera device includes a monocular camera.

[0053] Specifically, image processing and analysis using a positioning tag recognition algorithm includes the robot using image processing algorithms to analyze and process the image information of the positioning tag, calculating the two-dimensional coordinates of the key feature points of the positioning tag in the image, and combining this with the known three-dimensional feature point coordinates. A pose estimation algorithm is then used to solve for the rotation matrix and translation vector of the first camera device, thereby determining the position and pose of the positioning tag in space. Optionally, the pose estimation algorithm includes the Perspective-n-Point (PnP) algorithm.

[0054] Furthermore, once the first camera device recognizes the positioning tag, the robot will trigger the function of entering the packaging container. Based on the position and posture information of the positioning tag, the robot can determine the position of the packaging container, i.e., the packaging container position information, and thus calculate the position of the target point directly in front of the packaging container. The motion control system controls the robot to move along the shortest path to the target point to the front of the packaging container, and then turn its back to the packaging container so that it can subsequently align with the packaging container in real time using the back scanning device and move backward into it.

[0055] Step 104: Obtain the contour dataset of the packaging container. Use clustering and feature matching algorithms to obtain the relative position calibration set of the packaging container based on the contour dataset. Control the robot to enter the packaging container according to the relative position calibration set of the packaging container.

[0056] In one embodiment, a scanning device is mounted on the back of the robot to scan the contours of three sides of the inner wall of the packaging container, acquiring a contour dataset of the packaging container. The output data of the scanning device includes point cloud data, representing the distance between the inner wall of the packaging container and the scanning device, and the corresponding scanning angle. By comparing and matching the scanned contour data, precise calibration of the robot's position with the packaging container's position is achieved, ensuring accurate alignment of the robot with the packaging container when reversing. The scanning device includes a single-line LiDAR.

[0057] Step 106: Obtain the bottom positioning information set of the packaging container, and use the positioning tag recognition algorithm to obtain the pose information of the packaging container based on the bottom positioning information set. Control the robot to adjust its own posture according to the pose information of the packaging container.

[0058] In one embodiment, a second camera is installed on the robot's abdomen, specifically below its belly, to capture positioning information from a location tag on the bottom of the packaging container. The robot then uses image processing and recognition algorithms to identify, locate, and calculate the posture of the positioning tag, enabling precise prone control. The second camera includes a monocular camera.

[0059] Specifically, the second camera acquires the bottom positioning information of the packaging container, and the robot uses a positioning tag recognition algorithm to obtain the container's position information. This algorithm is identical to the one used to process the positioning information set of the inner wall of the packaging container acquired by the first head camera. Once the positioning tag is recognized, the robot needs to align and calibrate its position based on the tag's location and orientation information. The motion control system can adjust the robot's trajectory, joint angles, and speed based on the real-time positioning tag information, enabling the robot to smoothly and accurately lower itself and perfectly align with the positioning tag on the bottom of the packaging container.

[0060] In the aforementioned method for a robot to autonomously enter a packaging container, the robot uses a head-mounted camera to identify positioning tags on the inner wall of the packaging container, obtaining a set of positioning information for the inner wall. Based on this information, a positioning tag recognition algorithm is used to obtain the container's position information, triggering the autonomous entry function. The motion control system then controls the robot to move to the container's location based on the positioning information. A rear-mounted scanning device scans the container's contour to obtain a contour dataset. Clustering and feature matching algorithms are used to obtain a relative position calibration set based on the contour dataset. The robot then enters the container according to this calibration set, achieving autonomous entry. A second camera located on the robot's underside identifies positioning tags on the bottom of the packaging container, obtaining a set of bottom positioning information. Based on this bottom positioning information, a positioning tag recognition algorithm is used to obtain the container's pose information. The robot then adjusts its posture according to the container's pose information, precisely positioning itself inside the container by aligning with the positioning tags.

[0061] In one embodiment, obtaining the location information of a packaging container using a positioning tag recognition algorithm based on the inner wall positioning information set includes: preprocessing the inner wall positioning information set; extracting features based on the processed inner wall positioning information set to obtain a feature point set; and calculating the pose based on the feature point set to obtain the location information of the packaging container.

[0062] Specifically, image preprocessing is performed on the image information of the positioning tags, which constitutes the inner wall positioning information set. Image preprocessing includes noise removal and binarization. Feature extraction is then performed on the preprocessed binary image, including contour detection, template matching, and feature point extraction to obtain a feature point set. The positional information of the feature points in the image is crucial for pose estimation; these feature points have predefined corresponding 3D coordinates in 3D space. Pose calculation is then performed based on the feature point set, including 2D (Two-Dimensional) to 3D (Three-Dimensional) correspondence and pose resolution to obtain the packaging container's position information.

[0063] In this embodiment, the inner wall positioning information set is preprocessed; feature extraction is performed based on the processed inner wall positioning information set to obtain a feature point set; pose calculation is performed based on the feature point set to obtain the packaging container position information, providing data support for the robot to trigger the function of entering the packaging container.

[0064] In one embodiment, preprocessing the inner wall positioning information set includes: filtering the inner wall positioning information set using a denoising algorithm, and binarizing the filtered inner wall positioning information set.

[0065] In one embodiment, noise in the image is reduced by applying a denoising algorithm, and then the color image of the filtered inner wall positioning information set is converted into a binary image, thereby simplifying the image processing process.

[0066] In this embodiment, a denoising algorithm is used to filter the inner wall positioning information set. Noise removal helps improve the accuracy of subsequent processing steps, especially in edge detection and feature extraction tasks. Binarizing the filtered inner wall positioning information set provides better contrast and feature visibility.

[0067] In one embodiment, feature extraction based on the preprocessed inner wall positioning information set to obtain the feature point set includes: performing contour detection on the preprocessed inner wall positioning information set to filter out the positioning label contour set; matching and fitting the pre-generated positioning label template set with the positioning label contour set to obtain the positioning label contour fitting set; and obtaining the feature point set based on the positioning label contour fitting set.

[0068] In one embodiment, contour detection is performed on the preprocessed binary image, and then a polygon approximation algorithm is used to examine the shape, area, and other features of the contours. The extracted contours can be filtered to retain those that are likely location tags, thus obtaining a location tag contour set. This retained set of location tag contours is then matched with a predefined set of location tag templates, and the least squares method is used to fit the templates to obtain a fitted set of location tag contours. The location tag templates are pre-generated and contain the shape and encoding information of the location tags; each location tag has a unique corresponding encoding information. A corner detection algorithm is then used to extract a set of feature points from the fitted contours (i.e., the fitted set of location tag contours).

[0069] In this embodiment, the positioning label contour set is filtered out by performing contour detection on the preprocessed inner wall positioning information set; the positioning label template set is matched and fitted with the positioning label contour set to obtain the positioning label contour fitting set; the feature point set is obtained based on the positioning label contour fitting set to provide a data basis for calculating and obtaining the packaging container position information, so as to realize the function of robot autonomously entering the packaging container.

[0070] In one embodiment, the process of calculating the pose based on the feature point set to obtain the packaging container position information includes: pairing the feature point set with the corresponding point set in three-dimensional space, calling the pose estimation algorithm to obtain the robot pose dataset, and obtaining the packaging container position information based on the robot pose dataset.

[0071] In one embodiment, pose calculation based on the feature point set includes 2D-3D correspondence and pose resolution. Specifically, 2D-3D correspondence involves pairing the feature point set in the image with its corresponding points in three-dimensional space. For pose calculation, at least three pairs of 2D-3D paired feature points are required.

[0072] Furthermore, the PnP algorithm is invoked based on the pairing of feature points and their corresponding points to estimate the camera's rotation and translation vectors. The algorithm aims to minimize the projection error, specifically the reprojection error of feature points in 3D space onto the image plane. This process can be solved using an iterative optimization method (Levenberg-Marquardt). After the algorithm iterates, the estimated rotation and translation vectors are obtained, representing the camera's pose, i.e., the robot pose dataset. The results are then transformed to the camera's coordinate system. This involves applying an inverse transformation to the pose information in the camera's coordinate system to obtain the location and pose information of the positioning tag in 3D space, i.e., the packaging container's position information.

[0073] In this embodiment, the feature point set is paired with the corresponding point set in three-dimensional space, and the pose estimation algorithm is called to obtain the robot pose dataset. Based on the robot pose dataset, the position information of the packaging container is obtained. According to the position information of the packaging container, the robot can determine the position of the packaging container, thereby calculating the position of the target point directly in front of the packaging container. The robot then moves to the front of the packaging container along the shortest path to the target point and turns its back to the packaging container so that it can subsequently rely on the rear LiDAR to align with the packaging container in real time and move backward into it.

[0074] In one embodiment, obtaining the relative position calibration set of the packaging container based on the contour dataset using clustering and feature matching algorithms includes: filtering the contour dataset and using a point cloud clustering algorithm to divide the filtered contour dataset into different clusters to extract the contour cluster set; and using an iterative nearest point algorithm to match a pre-established packaging container contour template set with the contour cluster set to obtain the relative position calibration set of the packaging container.

[0075] In one embodiment, a contour scanning and position calibration algorithm is used to obtain relative position calibration data between the robot and the packaging container based on contour data. Specifically, the contour data processing system filters the received point cloud data from the scanning device to remove noise and outliers, resulting in clear point cloud data. A clustering algorithm is then used to divide the point cloud data into different clusters, extracting contour cluster sets, each representing an independent object contour. The extracted contours are matched against a pre-established set of packaging container contour templates. The ICP (Iterative Closest Point) algorithm is used to match the packaging container contour and calculate the differences in position and orientation between the robot and the packaging container, determining their relative position and obtaining a relative position calibration set for the packaging container. The motion control system adjusts the robot's posture and position to ensure accurate alignment with the inner wall of the packaging container. During the calibration process, the calibration results are monitored in real time to obtain relative position calibration data, allowing for fine-tuning and correction to improve the accuracy and stability of the calibration.

[0076] Specifically, point cloud clustering algorithms are used to segment point cloud data into different clusters, starting from the first point in the point cloud. For each point, its neighboring points are checked against a threshold for the search neighborhood size to see if they meet the clustering criteria. Points that meet the criteria are added to a new cluster, and this process is repeated recursively, checking neighboring points to see if they also meet the criteria and adding them to the same cluster. This process is repeated until all points have been traversed, and all points are assigned to their respective clusters, each representing an independent object or part. This step will aid in subsequent contour matching and position calibration of the inner walls of packaging containers.

[0077] The Intercalary Point Collation (ICP) algorithm is an iterative point cloud registration algorithm used to find the optimal rigid transformation (rotation and translation) between two sets of point clouds, making corresponding points in the two sets as close as possible. A pre-established set of packaging container outline templates is considered the "reference point cloud," and the clustered point cloud is considered the "point cloud to be registered." For each point in the point cloud to be registered, its nearest neighbor in the reference point cloud is found, establishing corresponding point pairs. The rigid transformation that makes the corresponding point pairs closest is calculated using the least squares method, aligning the point cloud to be registered as closely as possible with the reference point cloud. Specifically, the ICP algorithm minimizes the distance error between corresponding point pairs to solve for the optimal rotation and translation.

[0078] In this embodiment, the contour dataset is filtered and then divided into different clusters using a point cloud clustering algorithm to extract the contour cluster set. The iterative nearest point algorithm is used to match the pre-established packaging container contour template set with the contour cluster set to obtain the packaging container relative position calibration set. The motion control system adjusts the robot's motion trajectory and speed according to the packaging container relative position calibration set and the real-time relative position calibration data. The point cloud clustering algorithm and the ICP algorithm are used to make the robot smoothly back into the packaging container.

[0079] Combination Figure 2 As shown in Figure 2, a right view and a front view of the robot and packaging container according to an embodiment of the present invention are presented. Figure 2 (a) is the right view. Figure 2 (b) is the front view. In some embodiments, the robot 202 is in a standing quadruped posture, at which time it autonomously navigates the packaging map using the LiDAR mapping and positioning function. The packaging container includes a pressing surface 204, a size positioning label 206, front and rear guide surfaces 208, and left and right guide surfaces 210, wherein the size positioning label is located on the inner wall and bottom of the packaging container, the front and rear guide surfaces include a front guide surface and a rear guide surface, and the left and right guide surfaces include a left guide surface and a right guide surface.

[0080] Combination Figure 3 As shown in Figure 3, a left view of the robot and packaging container in one embodiment of the present invention is presented, wherein both the robot 202 and the packaging container 302 are... Figure 2 The robot 202 shown is identical to the packaging container.

[0081] Combination Figure 4 As shown in Figure 4, a schematic diagram of the robot and packaging container in a far-field state according to an embodiment of the present invention is presented, wherein... Figure 4 (a) is the right view. Figure 4 (b) Main view, Figure 4(c) is the left view. Robot 202 acquires image information of the positioning tag on the inner wall of packaging container 302 through a first camera device on its head; this is the positioning information set of the inner wall of packaging container 302. Robot 202 uses image processing algorithms to analyze and process the image information of the positioning tag, calculating the two-dimensional coordinates of the key feature points of the positioning tag in the image. Combining this with the known three-dimensional feature point coordinates, a pose estimation algorithm is used to solve for the rotation matrix and translation vector of the first camera device, thereby determining the position and pose of the positioning tag in space. Once the camera device recognizes the positioning tag, robot 202 will trigger the function of entering packaging container 302. Based on the position and pose information of the positioning tag, robot 202 can determine the position of packaging container 302, i.e., the position information of the packaging container, and thus calculate the position of the target point directly in front of packaging container 302. The motion control system controls robot 202 to move along the shortest path to the target point to the front of packaging container 302, and then turn its back to packaging container 302, so that it can subsequently align with packaging container 302 in real time using a back-scanning device and retreat into it.

[0082] Combination Figure 5 As shown in Figure 5, a schematic diagram of the robot and packaging container in the arrival state according to an embodiment of the present invention is presented. Figure 5 (a) is the right view. Figure 5 (b) Main view, Figure 5 (c) is a left view, showing the robot 202 aligning with the packaging container 302 in real time using a back-mounted scanning device and then backing into the packaging container 302. Specifically, the robot 202, equipped with a back-mounted scanning device, scans the contours of three sides of the inner wall of the packaging container 302 to obtain a contour dataset of the packaging container 302. The output data of the scanning device includes point cloud data, representing the distance between the inner wall of the packaging container 302 and the scanning device, and the corresponding scanning angle. The robot 202 uses contour scanning and position calibration algorithms to obtain a relative position calibration set between the robot 202 and the packaging container 302 based on the contour data. The motion control system adjusts the motion trajectory and speed of the robot 202 based on the relative position calibration set and the real-time acquired relative position calibration data, enabling it to smoothly back into the packaging container 302.

[0083] Combination Figure 6 As shown in Figure 6, a three-view diagram of the packaging container in the packaged state according to an embodiment of the present invention is presented. Figure 6 (a) is the right view. Figure 6 (b) Main view, Figure 6 (c) is the left view. A groove guide structure is designed at the bottom of the packaging container 302 to guide the robot 202 to accurately align and position itself when it lies down inside the packaging container 302. A fixing structure is designed on the top cover of the packaging container 302 to further stabilize the position of the robot 202 during the packaging process, while preventing shaking and loosening.

[0084] Combination Figure 7 As shown in Figure 7, a schematic diagram of a robot lying prone in a packaging container according to an embodiment of the present invention is presented. Figure 7 (a) is the right view. Figure 7 (b) is the left view. The robot 202 enters the packaging container 302, obtains the bottom positioning information set of the packaging container 302, uses the positioning tag recognition algorithm based on the bottom positioning information set to obtain the pose information of the packaging container, and adjusts its own posture according to the pose information of the packaging container under the guidance of the groove structure to achieve lying down inside the packaging container 302.

[0085] In another embodiment, such as Figure 8 As shown, a method for a robot to autonomously enter a packaging container is provided, including:

[0086] Step 802: Obtain the inner wall positioning information set of the packaging container.

[0087] Step 804: The inner wall positioning information set is filtered using a denoising algorithm, and the filtered inner wall positioning information set is then binarized.

[0088] Step 806: Perform contour detection on the preprocessed inner wall positioning information set and filter out the positioning label contour set.

[0089] Step 808: Match and fit the pre-generated location tag template set with the location tag contour set to obtain the location tag contour fitting set.

[0090] Step 810: Obtain the feature point set based on the positioning label contour fitting set.

[0091] Step 812: Pair the feature point set with the corresponding point set in the three-dimensional space, and call the pose estimation algorithm to obtain the robot pose dataset. Based on the robot pose dataset, obtain the position information of the packaging container.

[0092] Step 814: Control the robot to move to the location of the packaging container based on the packaging container location information.

[0093] Step 816: Obtain the outline dataset of the packaging container.

[0094] Step 818: Filter the contour dataset and use a point cloud clustering algorithm to divide the filtered contour dataset into different clusters and extract the contour cluster set.

[0095] Step 820: The iterative nearest point algorithm is used to match the pre-established packaging container outline template set with the outline cluster set to obtain the packaging container relative position calibration set.

[0096] Step 822: Control the robot to enter the packaging container according to the relative position calibration set of the packaging container.

[0097] Step 824: Obtain the bottom positioning information set of the packaging container, and use the positioning tag recognition algorithm to obtain the pose information of the packaging container based on the bottom positioning information set. Control the robot to adjust its own posture according to the pose information of the packaging container.

[0098] In this embodiment, a head-mounted camera and a positioning tag recognition algorithm are introduced to enable the robot to autonomously trigger entry into the packaging container. A back-mounted scanning device then scans the inner wall contour of the packaging container and calibrates its position, helping the robot accurately enter the container. Compared to traditional robotic packaging requiring manual intervention and operation, this method offers higher autonomy and intelligence, improving the efficiency and flexibility of the packaging process. Furthermore, the back-mounted scanning device and a camera located on the underside of the robot's body enable precise positioning of the robot within the packaging container. The scanning device scans the inner wall contour of the packaging container and calibrates it against the robot's own position, while the camera on the underside of the robot identifies the positioning tag on the bottom of the packaging container, ensuring precise alignment and positioning of the robot within the container, thus improving packaging accuracy and stability. Through the design and coordination of the head-mounted camera, back-mounted scanning device, underside camera, and the internal structure of the packaging container, the robot can achieve autonomous entry into the packaging container, accurate positioning, and stable positioning.

[0099] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0100] Based on the same inventive concept, this application also provides an apparatus for implementing the above-described method of robot autonomous entry into a packaging container. The solution provided by this apparatus is similar to the implementation described in the above-described method. Therefore, the specific limitations of one or more apparatus embodiments for robot autonomous entry into packaging containers provided below can be found in the limitations of the method for robot autonomous entry into packaging containers described above, and will not be repeated here.

[0101] In one embodiment, such as Figure 9As shown, a device for a robot to autonomously enter a packaging container is provided, comprising: a packaging container positioning module 902, a packaging container entry module 904, and a packaging container pose adjustment module 906, wherein:

[0102] The packaging container positioning module 902 is used to acquire the inner wall positioning information set of the packaging container, and to acquire the packaging container position information using a positioning tag recognition algorithm based on the inner wall positioning information set, and to control the robot to move to the location of the packaging container according to the packaging container position information.

[0103] The packaging container entry module 904 is used to obtain the contour dataset of the packaging container. It uses clustering and feature matching algorithms to obtain the relative position calibration set of the packaging container based on the contour dataset, and controls the robot to enter the packaging container according to the relative position calibration set of the packaging container.

[0104] The packaging container pose adjustment module 906 is used to obtain the bottom positioning information set of the packaging container, and to obtain the pose information of the packaging container by using a positioning tag recognition algorithm based on the bottom positioning information set, and to control the robot to adjust its own posture according to the pose information of the packaging container.

[0105] In one embodiment, the packaging container positioning module 902 includes:

[0106] The preprocessing module is used to preprocess the inner wall positioning information set;

[0107] The feature extraction module is used to extract features based on the preprocessed inner wall positioning information set to obtain a set of feature points;

[0108] The pose calculation module is used to calculate the pose based on the feature point set and obtain the position information of the packaging container.

[0109] In one embodiment, the preprocessing module includes:

[0110] The Gaussian filtering and binarization module is used to filter the inner wall positioning information set using a denoising algorithm and then perform binarization processing on the filtered inner wall positioning information set.

[0111] In one embodiment, the feature extraction module includes:

[0112] The contour detection module is used to perform contour detection on the preprocessed inner wall positioning information set and filter out the positioning label contour set.

[0113] The matching and fitting module is used to match and fit the pre-generated set of positioning label templates with the set of positioning label contours to obtain a set of fitted positioning label contours.

[0114] The feature point set acquisition module is used to obtain the feature point set based on the contour fitting set of the positioning label.

[0115] In one embodiment, the pose calculation module includes:

[0116] The packaging container location information acquisition module is used to pair the feature point set with the corresponding point set in three-dimensional space, call the pose estimation algorithm to obtain the robot pose dataset, and obtain the packaging container location information based on the robot pose dataset.

[0117] In one embodiment, the packaging container inlet module 904 includes:

[0118] The contour cluster set extraction module is used to filter the contour dataset and use a point cloud clustering algorithm to divide the filtered contour dataset into different clusters and extract the contour cluster set.

[0119] The packaging container relative position calibration set acquisition module is used to match a pre-established packaging container contour template set and contour cluster set using an iterative nearest point algorithm to obtain the packaging container relative position calibration set.

[0120] The various modules in the aforementioned device for the robot to autonomously enter the packaging container can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0121] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 10 As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores data acquired and generated during the method of autonomous robot entry into packaging containers. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When executed by the processor, the computer program implements a method for autonomous robot entry into packaging containers.

[0122] Those skilled in the art will understand that Figure 10The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0123] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0124] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0125] In one embodiment, a packaging container that a robot autonomously enters is provided, the packaging container comprising:

[0126] The groove structure, designed based on the robot's shape data, is used to guide the robot to align and position itself when adjusting its posture inside the packaging container.

[0127] A fixed structure, designed based on data from the robot's back, is used to stabilize the robot's position during the packaging process.

[0128] Specifically, a grooved guide structure is designed at the bottom of the packaging container to guide the robot to accurate alignment and positioning when it lies down inside the container. The groove structure is designed according to the robot's size and shape, using a combination of straight lines and curves to ensure the robot fits snugly into the groove. Even if there is a slight positional deviation, the robot can still be guided into the groove for accurate alignment. A fixing structure is designed on the top of the packaging container to further stabilize the robot's position during the packaging process, while preventing shaking and loosening. After the packaging lid is closed, the fixing structure on the lid fits tightly against the robot's back, providing stable support and ensuring the safety and reliability of the packaging.

[0129] In this embodiment, a fixing structure is introduced inside the packaging container to ensure the robot's stability within the container. The guide groove structure at the bottom of the packaging container ensures the robot can perfectly fit into the groove, achieving accurate positioning. The fixing structure on the top of the packaging container fits snugly against the robot's back, providing stable support and preventing the robot from shaking or loosening during packaging, thus improving the safety and reliability of the packaging process.

[0130] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0131] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0132] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0133] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method of robotically autonomously entering a packaging container, characterized in that, The method comprises: obtaining an inner wall positioning information set of the packaging container, obtaining packaging container position information based on the inner wall positioning information set using a positioning label recognition algorithm, and controlling the robot to move to the position of the packaging container according to the packaging container position information; obtaining a contour data set of the packaging container, obtaining the packaging container relative position calibration set according to the contour data set using a clustering algorithm and a feature matching algorithm, and controlling the robot to enter the packaging container according to the packaging container relative position calibration set; obtaining a bottom positioning information set of the packaging container, obtaining packaging container pose information based on the bottom positioning information set using a positioning label recognition algorithm, and controlling the robot to adjust its own pose according to the packaging container pose information.

2. The method of claim 1, wherein, The method comprises: preprocessing the inner wall positioning information set; performing feature extraction based on the preprocessed inner wall positioning information set to obtain a feature point set; performing pose calculation according to the feature point set to obtain packaging container position information.

3. The method of claim 2, wherein, The method comprises: performing filtering on the inner wall positioning information set using a denoising algorithm, and performing binary processing on the filtered inner wall positioning information set.

4. The method of claim 2, wherein, The method comprises: performing contour detection on the preprocessed inner wall positioning information set, and screening out a positioning label contour set; matching and fitting a pre-generated positioning label template set with the positioning label contour set to obtain a positioning label contour fitting set; obtaining a feature point set based on the positioning label contour fitting set.

5. The method of claim 2, wherein, The method comprises: pairing the feature point set with a corresponding point set in a three-dimensional space, calling a pose estimation algorithm to obtain a robot pose data set, and obtaining packaging container position information based on the robot pose data set.

6. The method of claim 1, wherein, The method comprises: performing filtering processing on the contour data set, and dividing the filtered contour data set into different clusters using a point cloud clustering algorithm to extract a contour cluster set; matching a pre-established packaging container contour template set with the contour cluster set using an iterative closest point algorithm to obtain a packaging container relative position calibration set.

7. A device for a robot to autonomously enter a packaging container, characterized in that The device comprises: a packaging container positioning module for obtaining an inner wall positioning information set of the packaging container, obtaining packaging container position information based on the inner wall positioning information set using a positioning label recognition algorithm, and controlling the robot to move to the position of the packaging container according to the packaging container position information; a packaging container entering module for obtaining a contour data set of the packaging container, obtaining the packaging container relative position calibration set according to the contour data set using a clustering algorithm and a feature matching algorithm, and controlling the robot to enter the packaging container according to the packaging container relative position calibration set; a packaging container pose adjusting module for obtaining a bottom positioning information set of the packaging container, obtaining packaging container pose information based on the bottom positioning information set using a positioning label recognition algorithm, and controlling the robot to adjust its own pose according to the packaging container pose information. The package container pose adjustment module is configured to acquire a set of bottom positioning information of the package container, acquire package container pose information by using a positioning label recognition algorithm based on the set of bottom positioning information, and control the robot to adjust its own pose according to the package container pose information.

8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that, The computer program, when executed by the processor, implements the steps of the method of any one of claims 1 to 6.

9. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program, when executed by the processor, implements the steps of the method of any one of claims 1 to 6.

Images