Three-dimensional point cloud processing method and system for anti-wrinkle interference clothes for massage robot, terminal and medium

Abstract

The invention discloses a three-dimensional point cloud processing method, system, terminal and medium for anti-wrinkle interference clothes for a massage robot, and relates to the technical field of robot perception and point cloud processing, and the method comprises the steps: obtaining the three-dimensional point cloud data of the surface of the clothes based on a depth camera, and carrying out the cutting processing, and obtaining the point cloud data after cutting; preprocessing the cut point cloud data to obtain target point cloud data, and performing cascade filtering processing to obtain optimized point cloud data; and based on the optimized point cloud data, calculating a normal vector of each point, and determining to carry out massage path planning so as to control the massage robot to execute a massage action based on the planned massage path. The method can remarkably reduce the influence of clothes wrinkles on massage direction recognition of the massage robot, and can adapt to interference of clothes made of different materials. In addition, according to the cascade filtering processing method, the calculation burden is greatly reduced while the effectiveness of normal vector estimation is ensured.

Description

Technical Field

[0001] This invention relates to the field of robot perception and point cloud processing technology, and in particular to a three-dimensional point cloud processing method, system, terminal and medium for anti-wrinkle interference clothing for massage robots. Background Technology

[0002] With the fast pace of modern life, neck, shoulder, back, and leg pain, muscle strain, and other problems are becoming increasingly common. Traditional manual physiotherapy services face challenges such as low efficiency, high cost, and lack of standardization. Physiotherapy massage robots have emerged to address this need, integrating artificial intelligence, biomechanics, and sensor technology to provide intelligent solutions for personal health management and rehabilitation.

[0003] When a massage robot performs its tasks, the vision system needs to acquire point clouds of the user's body surface using a depth camera. However, dynamic wrinkles created by the user's clothing can distort the point cloud surface, causing the normal vector calculation to deviate from the actual curvature of the human body. To address this, existing technologies combine the vision system with real-time perception of the human body contour. By constructing a 3D point cloud of the human body, the normal vectors of corresponding acupoints can be quickly estimated. However, in practical applications, point cloud data is often interfered with by clothing wrinkles, introducing significant errors. To denoise the point cloud, existing technologies only use statistical filtering to remove some noise points that constitute clothing wrinkles. However, this may lead to raised areas of wrinkles being misidentified as outliers due to abrupt changes in distance, resulting in holes in the point cloud. Furthermore, these methods only remove discrete points without reconstructing surface continuity, leaving the normal vector directions still inconsistent.

[0004] Therefore, existing technologies still have shortcomings. Summary of the Invention

[0005] The technical problem to be solved by this invention is to provide a three-dimensional point cloud processing method, system, terminal, and medium for anti-wrinkle interference clothing for massage robots, addressing the aforementioned deficiencies of the prior art. The technical solution adopted by this invention is as follows: In a first aspect, the present invention provides a three-dimensional point cloud processing method for anti-wrinkle interference clothing for massage robots, wherein the method includes: The 3D point cloud data of the clothing surface is acquired using a depth camera, and the 3D point cloud data is cropped to obtain cropped point cloud data. The cropped point cloud data is preprocessed to obtain target point cloud data, and the target point cloud data is then subjected to cascaded filtering to obtain optimized point cloud data after cascaded filtering. Based on the optimized point cloud data, the normal vector of each point is calculated, and a massage path is planned based on the calculated normal vector to control the massage robot to perform massage actions based on the planned massage path.

[0006] In one implementation, the three-dimensional point cloud data is cropped to obtain cropped point cloud data, including: Based on key point recognition technology, the location of acupoints on the human body is determined; The three-dimensional point cloud data is divided and cropped using the acupoints as the center to obtain cropped point cloud data.

[0007] In one implementation, the three-dimensional point cloud data is cropped to obtain cropped point cloud data, including: From the perspective of the depth camera, the three-dimensional point cloud data is cropped at a preset fixed size to obtain cropped point cloud data.

[0008] In one implementation, the cropped point cloud data is preprocessed to obtain target point cloud data, including: Statistical analysis of the local density of the cropped point cloud data, identification and removal of abnormal point clouds; The voxel downsampling method is used to divide the point cloud space of the cropped point cloud data after removing abnormal point clouds into uniform voxel grids, and only one representative point is retained in each voxel grid to obtain the target point cloud data.

[0009] In one implementation, the target point cloud data is subjected to cascaded filtering to obtain optimized point cloud data after cascaded filtering, including: The target point cloud data is subjected to adaptive bilateral filtering to obtain the first point cloud data; The first point cloud data is subjected to curvature-sensing filtering to obtain the second point cloud data; The second point cloud data is subjected to adaptive bilateral filtering again to obtain the optimized point cloud data.

[0010] In one implementation, the target point cloud data is subjected to adaptive bilateral filtering to obtain first point cloud data, including: Calculate the average distance between each sampling point and its nearest neighbor in the target point cloud data; The filtering parameters are automatically adjusted based on the average spacing, and the calculation methods for spatial weights and curvature weights are improved. The first point cloud data is obtained by using a joint weight composed of spatial weight and curvature weight, and setting a mixing factor to update the point position.

[0011] In one implementation, the first point cloud data undergoes curvature-sensing filtering to obtain second point cloud data, including: Calculate the curvature of each sampling point in the first point cloud data, and adjust the filtering range based on the calculated curvature to determine the curvature adaptive weight; The final weights are determined based on spatial weights and curvature adaptive weights, and the point positions are updated using the final weights to obtain the second point cloud data.

[0012] Secondly, embodiments of the present invention also provide a three-dimensional point cloud processing system for anti-wrinkle interference clothing for massage robots, wherein the system is used to implement the steps of the above-mentioned three-dimensional point cloud processing method for anti-wrinkle interference clothing for massage robots, and the system includes: The point cloud cropping module is used to acquire three-dimensional point cloud data of the clothing surface based on a depth camera, and to crop the three-dimensional point cloud data to obtain cropped point cloud data. A cascaded filtering module is used to preprocess the cropped point cloud data to obtain target point cloud data, and to perform cascaded filtering on the target point cloud data to obtain optimized point cloud data after cascaded filtering. The normal vector estimation module is used to calculate the normal vector of each point based on the optimized point cloud data, and to plan the massage path based on the calculated normal vector, so as to control the massage robot to perform massage actions based on the planned massage path.

[0013] Thirdly, embodiments of the present invention also provide a terminal, wherein the terminal includes a memory, a processor, and a three-dimensional point cloud processing program for anti-wrinkle interference clothing for massage robots stored in the memory and executable on the processor. When the processor executes the three-dimensional point cloud processing program for anti-wrinkle interference clothing for massage robots, it implements the steps of the three-dimensional point cloud processing method for anti-wrinkle interference clothing for massage robots according to any of the above solutions.

[0014] Fourthly, embodiments of the present invention also provide a computer-readable storage medium, wherein the computer-readable storage medium stores a three-dimensional point cloud processing program for anti-wrinkle interference clothing for massage robots, the three-dimensional point cloud processing program for anti-wrinkle interference clothing for massage robots implementing the steps of the three-dimensional point cloud processing method for anti-wrinkle interference clothing for massage robots as described in any of the above-described schemes on the computer-readable storage medium.

[0015] Beneficial Effects: Compared with existing technologies, this invention provides a 3D point cloud processing method for wrinkle-resistant clothing used in massage robots. First, 3D point cloud data of the clothing surface is acquired using a depth camera, and this data is then cropped to obtain cropped point cloud data. Next, the cropped point cloud data is preprocessed to obtain target point cloud data, which is then subjected to cascaded filtering to obtain optimized point cloud data. Then, based on the optimized point cloud data, the normal vector of each point is calculated, and massage path planning is performed based on the calculated normal vectors to control the massage robot to perform massage actions according to the planned path. This invention can significantly reduce the impact of clothing wrinkles on the massage direction recognition of the massage robot and can adapt to interference from different clothing materials. The cascaded filtering method used in this invention significantly reduces the computational burden while ensuring the effectiveness of normal vector estimation. Furthermore, this invention is applicable not only to the original complete point cloud but also to the cropped point cloud, and has a strong object generalization ability. It overcomes the contradiction between edge preservation and noise suppression in traditional methods, and provides high-confidence geometric data support for massage path planning and safety force control. Attached Figure Description

[0016] Figure 1 This is a flowchart of a preferred embodiment of the three-dimensional point cloud processing method for anti-wrinkle interference clothing for massage robots provided by the present invention.

[0017] Figure 2 This is a technical roadmap for a three-dimensional point cloud processing method for anti-wrinkle interference clothing for massage robots, provided in an embodiment of the present invention.

[0018] Figure 3 The original point cloud image is shown in the three-dimensional point cloud processing method for anti-wrinkle interference clothing for massage robots provided in the embodiments of the present invention.

[0019] Figure 4 The flowchart illustrates the preprocessing and cascaded filtering processes in the three-dimensional point cloud processing method for anti-wrinkle interference clothing for massage robots provided in this embodiment of the invention.

[0020] Figure 5 This is a schematic diagram of the target point cloud data obtained after preprocessing in the three-dimensional point cloud processing method for anti-wrinkle interference clothing for massage robots provided in an embodiment of the present invention.

[0021] Figure 6 This is a schematic diagram of the optimized point cloud data obtained after cascaded filtering in the three-dimensional point cloud processing method for anti-wrinkle interference clothing for massage robots provided in an embodiment of the present invention.

[0022] Figure 7 A point cloud diagram of the human body when it is not wearing clothes.

[0023] Figure 8 This is a schematic diagram of the reference normal vector of the point cloud when verifying the normal vector error.

[0024] Figure 9 This is a schematic diagram of the point cloud normal vector distribution in the wrinkled clothing area during normal vector error verification.

[0025] Figure 10 This is a schematic diagram of the point cloud normal vector distribution after cascaded filtering.

[0026] Figure 11 This is a schematic diagram of a three-dimensional point cloud processing system for anti-wrinkle interference clothing for massage robots, provided in an embodiment of the present invention.

[0027] Figure 12 A schematic diagram of a terminal provided in an embodiment of the present invention. Detailed Implementation

[0028] To make the objectives, technical solutions, and effects of this invention clearer and more explicit, 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 of the invention and are not intended to limit the invention.

[0029] The flowchart shown in the attached diagram is for illustrative purposes only and does not necessarily include all content, operations, or steps, nor does it require execution in the described order. For example, some operations or steps can be broken down, combined, or partially merged, so the actual execution order may change depending on the actual situation.

[0030] It should be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise. It should be understood that, in order to clearly describe the technical solutions of the embodiments of the present invention, the terms "first" and "second" are used in the embodiments of the present invention to distinguish identical or similar items with essentially the same function and effect. For example, the first control information and the second control information are only used to distinguish different control information and do not limit their order. Those skilled in the art will understand that the words "first" and "second" do not limit the quantity or the order of execution, and that the words "first" and "second" do not necessarily imply that they are different. It should also be understood that the terms "and / or" as used in this specification and the appended claims refer to any combination of one or more of the associated listed items and all possible combinations, and include such combinations.

[0031] To address the problems of existing technologies, this embodiment provides a 3D point cloud processing method for wrinkle-resistant clothing used in massage robots. The method described in this embodiment can significantly reduce the impact of clothing wrinkles on the massage direction recognition of the massage robot. In specific applications, this embodiment first acquires 3D point cloud data of the clothing surface using a depth camera, and then crops the 3D point cloud data to obtain cropped point cloud data. Next, the cropped point cloud data is preprocessed to obtain target point cloud data, and then cascaded filtering is performed on the target point cloud data to obtain optimized point cloud data after cascaded filtering. Then, based on the optimized point cloud data, the normal vector of each point is calculated, and massage path planning is performed based on the calculated normal vectors to control the massage robot to perform massage actions based on the planned massage path. This invention can adapt to interference from clothing of different materials and effectively avoids blurring of the human body contour. The cascaded filtering method used in this invention significantly reduces the computational burden while ensuring the effectiveness of normal vector estimation. Furthermore, this invention is applicable not only to the original complete point cloud but also to the cropped point cloud, and has a strong object generalization ability. It overcomes the contradiction between edge preservation and noise suppression in traditional methods, and provides high-confidence geometric data support for massage path planning and safety force control.

[0032] The 3D point cloud processing method for anti-wrinkle clothing for massage robots in this embodiment can be applied to a terminal, which can be an intelligent terminal such as a massage robot or a computer. Specifically, as shown below... Figure 1 As shown, the three-dimensional point cloud processing method for anti-wrinkle interference clothing for massage robots includes the following steps: Step S100: Acquire three-dimensional point cloud data of the clothing surface based on a depth camera, and perform cropping processing on the three-dimensional point cloud data to obtain cropped point cloud data.

[0033] Combination Figure 2 As shown in the figure, this embodiment can collect point clouds using a depth camera, such as... Figure 3 As shown in the original point cloud image, the clothing was then inspected, and 3D point cloud data of the clothing surface was acquired, recording the spatial coordinate information of the clothing surface. This provides the raw input for all subsequent data processing. In this embodiment, by acquiring the three-dimensional point cloud data of the clothing surface, a complete point cloud covering the massage area can be obtained, ensuring the spatial integrity of the raw data and laying the foundation for subsequent precise processing.

[0034] After acquiring the 3D point cloud data of the clothing surface, this embodiment can perform cropping processing on the 3D point cloud data to obtain cropped point cloud data. This embodiment can use region cropping or default cropping to perform cropping processing in order to remove redundant areas in the point cloud that are not related to massage (such as background, non-target body parts), focusing on the core processing area. Specifically, when using region cropping, this embodiment first determines the location of human acupoints based on key point recognition technology. Then, the 3D point cloud data is divided and cropped with the location of the human acupoint as the center to obtain cropped point cloud data. For example, a circular or rectangular area is divided with the location of the human acupoint as the center, and then the point cloud data located in the circular or rectangular area of ​​the 3D point cloud data is retained. In this way, point cloud data that fits the needs of the massage scene can be obtained, accurately retaining the core area related to acupoints and reducing the amount of subsequent calculations. When using default cropping, this embodiment can crop the 3D point cloud data with a preset fixed size from the perspective of the depth camera to obtain cropped point cloud data. Specifically, the projection of the 3D point cloud data onto the camera's Z-axis coordinate is:

[0035] in, and table Indicates the coordinate range of the cropping area. This represents the intrinsic parameters of the depth camera, therefore the 3D point cloud data. The cropping area is: .in, The preset effective depth range is defined as the cropping region. It is evident that the cropping region is jointly determined by the depth camera's intrinsic parameters and the preset effective depth range. Based on the cropped point cloud data obtained using this method, the point cloud processing range can be reduced, irrelevant point cloud data can be removed, and the efficiency of subsequent algorithms can be improved.

[0036] Step S200: Preprocess the cropped point cloud data to obtain target point cloud data, and perform cascade filtering on the target point cloud data to obtain optimized point cloud data after cascade filtering.

[0037] To suppress noise points in clothing folds, existing technologies employ Gaussian filtering based on a weighted average of neighboring pixels or median filtering using the median. However, these methods easily misidentify high-frequency folds as noise and over-smooth edge corners, potentially damaging the texture of thin clothing. Traditional bilateral filtering, which combines spatial and color domain weights in RGB images, can smooth textures while preserving edges, but it doesn't adapt to the characteristics of 3D point clouds, and color domain weights cannot be defined during point cloud processing.

[0038] To address the problems in the prior art, this embodiment preprocesses the cropped point cloud data after obtaining it, combining... Figure 4 As shown in the figure. This embodiment can statistically analyze the local density of the cropped point cloud data, identify and remove abnormal point clouds, such as removing discrete point clouds that are far from normal point groups, removing flying point noise, eliminating the interference of random noise on subsequent filtering, and avoiding normal vector estimation errors caused by noise. Next, this embodiment can use a voxel downsampling method to divide the point cloud space of the cropped point cloud data after removing abnormal point clouds into uniform voxel grids, and retain only one representative point in each voxel grid, such as retaining the point at the center of the voxel grid or the point closest to the center, to obtain the target point cloud data. This embodiment can reduce the amount of point cloud data, reduce redundancy, and improve the computational efficiency of subsequent cascaded filtering processing by using the voxel downsampling method, while retaining the overall geometric structure of the point cloud and avoiding feature loss due to oversimplification. The target point cloud data obtained after preprocessing is as follows. Figure 5 As shown in the figure. The preprocessing in this embodiment can also be replaced by end-to-end point cloud segmentation technology.

[0039] Furthermore, after preprocessing, this embodiment performs cascade filtering on the target point cloud data, combined with... Figure 4 As shown, the cascaded filtering process includes: adaptive bilateral filtering, curvature-aware filtering, and a further adaptive bilateral filtering. Specifically, in this embodiment, the target point cloud data is first subjected to adaptive bilateral filtering to obtain first point cloud data. Then, the first point cloud data is subjected to curvature-aware filtering to obtain second point cloud data. Finally, the second point cloud data is subjected to adaptive bilateral filtering again to obtain the optimized point cloud data.

[0040] In practical applications, when performing adaptive bilateral filtering, this embodiment first calculates the average distance between each sampling point and its nearest neighbor in the target point cloud data. The formula for calculating the average distance is as follows:

[0041] in Indicates the current sampling point. express The nearest point, The total amount of the midpoint cloud, This represents the number of dynamically sampled points, taking the total amount of the point cloud. The smaller value than 100 is used to ensure computational efficiency. In this embodiment, the calculated average spacing provides a basis for the local density of the point cloud to automatically adjust the filtering parameters, ensuring that the filtering parameters match the actual characteristics of the point cloud.

[0042] Next, the filtering parameters are automatically adjusted based on the average spacing, as follows:

[0043]

[0044] in, Indicates the width of the spatial domain filter. The base value is used to prevent average spacing. Too small Indicates the width of the color gamut filter. This represents the scaling factor between the color gamut and the spatial domain. By adaptively adjusting the filtering parameters, it avoids the problem of insufficient or excessive filtering in certain areas caused by fixed parameters, thus adapting to point cloud regions with different wrinkle densities. It increases the scaling factor in sparse point cloud regions. Expanding the neighborhood preserves fold continuity, while reducing fold density in high-density areas. Focusing on specific areas, it enhances noise reduction in wrinkled regions.

[0045] To calculate neighborhood points For the center point Spatial influence weighting, this embodiment improves spatial weighting and curvature weight The calculation method is as follows:

[0046]

[0047] This embodiment strengthens the influence of nearby points by adding a squared term to the Euclidean distance, and makes the weight of distant points approach 0 through exponential decay. The rate at which weights decay can be controlled.

[0048] Spatial weight and curvature weight Joint weights for:

[0049] This embodiment uses a joint weight composed of spatial weight and curvature weight, and sets a mixing factor to update the point positions, thus obtaining the first point cloud data. The first point cloud data obtained after point position updating is as follows:

[0050] Among them, the mixing factor Setting it to 0.3 preserves the features of the original point cloud. This allows for the elimination of large-scale wrinkles while retaining the original geometric features of the human clothing surface (such as skeletal protrusions), avoiding feature loss due to over-smoothing. Furthermore, during the design of the blending factor, parameters can be dynamically adjusted, changing the fixed blending factor to a curvature-dependent one, using weak preservation in high curvature regions and strong preservation in low curvature regions, or directly employing a lightweight neural network residual learning mechanism. In other implementations, conditional filtering methods can be used to replace the adaptive bilateral filtering process.

[0051] Furthermore, when performing curvature-sensing filtering on the first point cloud data, this embodiment can first calculate the curvature of each sampling point in the first point cloud data. Curvature The calculation formula is as follows:

[0052] in, express of The nearest neighbor point, Represents the normal vector of the nearest point. The center point normal vector is used. By calculating the curvature, areas with high curvature (such as clothing folds) can be accurately identified, providing a basis for dynamically adjusting the filtering range.

[0053] Next, this embodiment adjusts the filtering range based on the calculated curvature, that is, the adjusted filtering width of the color gamut. for:

[0054] Therefore, the curvature adaptive weights can be determined. for:

[0055] The final weights are determined based on spatial weights and curvature adaptive weights. as follows:

[0056] In this embodiment, the final weights can be used to update the points and obtain the optimized point cloud data. Based on the curvature-aware filtering process in this embodiment, the filter width can be precisely matched with the curvature of the folded area, which helps to eliminate the deviation of the normal vector direction caused by folds. This facilitates strong filtering of high curvature areas to minimize clothing folds, while only noise removal is performed on low curvature areas, preserving the overall contour features.

[0057] Furthermore, in this embodiment, after adaptive bilateral filtering and curvature-aware filtering, the second point cloud data is further subjected to adaptive bilateral filtering again. This second adaptive bilateral filtering helps to repair the minor distortions remaining from the first two stages, maintain the topological continuity of the point cloud, avoid point cloud fragmentation caused by local optimization, achieve global consistency optimization, and ensure that the overall point cloud is consistent with the true curvature of the human body. It can be seen that the cascaded filtering process in this embodiment removes large-scale noise and eliminates large-volume folded structures through the first stage of adaptive bilateral filtering. Then, the second stage of curvature-aware filtering performs curvature-aware local optimization, sharpens high-intensity fold edges, and corrects the normal vector direction. Finally, the third stage of adaptive bilateral filtering performs global optimization again, repairing the minor oscillations introduced in the second stage and maintaining topological continuity. This cascaded filtering process achieves synergistic optimization of fold suppression and restoration of the true curvature of the human body. The optimized point cloud data obtained after cascaded filtering is shown below. Figure 6 As shown in the image.

[0058] Step S300: Based on the optimized point cloud data, calculate the normal vector of each point, and plan the massage path based on the calculated normal vector, so as to control the massage robot to perform massage actions based on the planned massage path.

[0059] The optimized point cloud data obtained after the above cascaded filtering process is a high-quality point cloud, free from wrinkle interference. At this point, the normal vector of each point can be calculated, which reflects the orientation of the point cloud surface. Calculating the normal vector helps to eliminate the normal vector error caused by clothing wrinkles, providing an accurate surface orientation basis for subsequent massage path planning, and ensuring that the massage head of the massage robot fits the human body surface.

[0060] This embodiment can also verify the directional effectiveness of acupoint massage by calculating the normal vector error before and after cascaded filtering. The calculation formula is as follows: Single point error :

[0061] in, It is an inverse cosine function. The reference normal vector of the human body point cloud. This represents the estimated normal vector of the point cloud in the wrinkled clothing region.

[0062] Average angular error :

[0063] Human body point cloud when the user is not wearing clothes (e.g.) Figure 7The normal vector generated (as shown) is used as the reference normal vector. Point cloud scanning is performed under the same pose as the point cloud to be processed, and the estimated normal vector is obtained through cascaded filtering in this embodiment. When calculating the average angle error, the clothing fold area and its corresponding reference normal vector are used for calculation. The error is calculated in the clothing fold area using the reference vector and the estimated normal vector. The reference point cloud normal vector is shown below. Figure 8 As shown, the distribution of the point cloud normal vectors of the original wrinkled clothing is as follows: Figure 9 As shown, with an error of 30.06°, the point cloud normal vector distribution after cascade filtering in this embodiment is as follows. Figure 10 As shown, the error is only 8.97°. This demonstrates that the cascaded filtering process in this embodiment can maintain computational efficiency while suppressing the impact of clothing wrinkles on the point cloud.

[0064] Furthermore, this embodiment can also use the optimized point cloud data or texture features obtained after the above-mentioned cascaded filtering process to identify and segment the clothing-covered area, excluding non-clothing areas (such as exposed skin and hair). This helps to clearly define the target working area of ​​the massage robot and avoid interference from non-target areas on subsequent massage path planning. Next, massage path planning is performed based on the calculated normal vector, clothing-covered area, and acupoint locations to obtain the massage path of the massage robot's massage head. The planned massage path must conform to the filtered human body surface geometry to ensure uniform massage intensity and coverage of target acupoints or target areas. For example, the massage path can be a circular path around the acupoint or a linear path along the muscle direction. Finally, the massage robot is controlled to perform massage actions based on the planned massage path, such as the massage head performing pressing, kneading, and other operations according to the massage path. As can be seen, this embodiment can interface the results of point cloud processing with actual massage operations, facilitating the massage robot to achieve intelligent massage based on point cloud optimization.

[0065] Based on the above embodiments, the present invention also provides a three-dimensional point cloud processing system for anti-wrinkle interference clothing for massage robots, the system being used to implement the steps in the above method embodiments. Specifically, as Figure 11As shown, the 3D point cloud processing system for wrinkle-resistant clothing for massage robots includes: a point cloud cropping module 10, a cascaded filtering module 20, and a normal vector estimation module 30. Specifically, the point cloud cropping module 10 acquires 3D point cloud data of the clothing surface using a depth camera and crops the 3D point cloud data to obtain cropped point cloud data. The cascaded filtering module 20 preprocesses the cropped point cloud data to obtain target point cloud data and performs cascaded filtering on the target point cloud data to obtain optimized point cloud data after cascaded filtering. The normal vector estimation module 30 calculates the normal vector of each point based on the optimized point cloud data and plans a massage path based on the calculated normal vector to control the massage robot to perform massage actions based on the planned massage path.

[0066] The working principle of each functional module in the three-dimensional point cloud processing system for anti-wrinkle interference clothing for massage robots in this embodiment is the same as the principle of each step in the above method embodiment, and will not be repeated here.

[0067] The modules in the aforementioned 3D point cloud processing system for anti-wrinkle interference clothing for massage robots can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of the terminal in hardware form or independent of it, or stored in the memory of the terminal in software form, so that the processor can call and execute the corresponding operations of each module.

[0068] Based on the above embodiments, the present invention also provides a terminal, the principle block diagram of which can be as follows: Figure 12 As shown. The terminal may include one or more processors 100 ( Figure 12 (Only one is shown in the image), memory 101, and computer program 102 stored in memory 101 and executable on one or more processors 100. For example, a 3D point cloud processing program for anti-wrinkle interference clothing for massage robots. When one or more processors 100 execute computer program 102, they can implement various steps in the embodiments of the 3D point cloud processing method for anti-wrinkle interference clothing for massage robots. Alternatively, when one or more processors 100 execute computer program 102, they can implement the functions of various modules / units in the embodiments of wearable transcranial electrical stimulation devices based on electrode combinations, without limitation herein.

[0069] In one embodiment, the processor 100 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0070] In one embodiment, memory 101 may be an internal storage unit of an electronic device, such as a hard drive or RAM. Memory 101 may also be an external storage device of the electronic device, such as a plug-in hard drive, smart media card (SMC), secure digital (SD) card, flash card, etc. Furthermore, memory 101 may include both internal and external storage units. Memory 101 is used to store computer programs and other programs and data required by the terminal. Memory 101 can also be used to temporarily store data that has been output or will be output.

[0071] Those skilled in the art will understand that Figure 12 The block diagram shown is merely a partial structural diagram related to the present invention and does not constitute a limitation on the terminal to which the present invention is applied. A specific terminal may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0072] Those skilled in the art will understand that all or part of the processes in the methods of 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 of the above methods. Any references to memory, storage, operational databases, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual operating data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0073] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for processing a three-dimensional point cloud of a wrinkle-resistant garment for a massage robot, the method comprising: The method comprises: obtaining three-dimensional point cloud data of a clothing surface based on a depth camera, and performing clipping processing on the three-dimensional point cloud data to obtain clipped point cloud data; performing preprocessing on the clipped point cloud data to obtain target point cloud data, and performing cascade filtering processing on the target point cloud data to obtain optimized point cloud data after cascade filtering; based on the optimized point cloud data, calculating the normal vector of each point, and based on the calculated normal vector, planning a massage path to control the massage robot to perform a massage action based on the planned massage path. 2.The wrinkle-resistant clothes' 3D point cloud processing method for a massage robot according to claim 1, wherein, The clipping processing on the three-dimensional point cloud data to obtain the clipped point cloud data comprises: determining the position of the human body acupoint based on a key point recognition technology; dividing and clipping the three-dimensional point cloud data based on the human body acupoint position as the center to obtain the clipped point cloud data. 3.The wrinkle-resistant clothes' 3D point cloud processing method for a massage robot according to claim 1, wherein, The clipping processing on the three-dimensional point cloud data to obtain the clipped point cloud data comprises: under the angle of view of the depth camera, clipping the three-dimensional point cloud data with a preset fixed size to obtain the clipped point cloud data. 4.The wrinkle-resistant clothes' 3D point cloud processing method for a massage robot according to claim 1, wherein, The preprocessing on the clipped point cloud data to obtain the target point cloud data comprises: statistically analyzing the local density of the clipped point cloud data, identifying and removing abnormal point clouds; using a voxel downsampling method, dividing the point cloud space of the clipped point cloud data after removing the abnormal point clouds into uniform voxel grids, and retaining only one representative point in each voxel grid to obtain the target point cloud data. 5.The wrinkle-resistant interference clothing three-dimensional point cloud processing method for a massage robot according to claim 1, wherein, The cascade filtering processing on the target point cloud data to obtain the optimized point cloud data after cascade filtering comprises: performing adaptive bilateral filtering processing on the target point cloud data to obtain first point cloud data; performing curvature-aware filtering processing on the first point cloud data to obtain second point cloud data; performing adaptive bilateral filtering processing on the second point cloud data again to obtain the optimized point cloud data. 6.The wrinkle-resistant interference clothing three-dimensional point cloud processing method for a massage robot according to claim 5, wherein, The adaptive bilateral filtering processing on the target point cloud data to obtain the first point cloud data comprises: calculating the average distance between each sampling point and the nearest point in the target point cloud data; automatically adjusting the filtering parameters according to the average distance, and improving the calculation method of the spatial weight and the curvature weight; using the joint weight composed of the spatial weight and the curvature weight, and setting a mixing factor to update the point position to obtain the first point cloud data. 7.The wrinkle-resistant interference clothing three-dimensional point cloud processing method for a massage robot according to claim 6, wherein, The curvature-aware filtering processing on the first point cloud data to obtain the second point cloud data comprises: calculating the curvature of each sampling point in the first point cloud data, and adjusting the filtering range based on the calculated curvature to determine the curvature adaptive weight; determining the maximum weight based on the spatial weight and the curvature adaptive weight, and updating the point position using the maximum weight to obtain the second point cloud data. 8.A three-dimensional point cloud processing system for anti-wrinkle interference clothing for a massage robot, characterized by, The system is used to implement the steps of the three-dimensional point cloud processing method for anti-wrinkle interference clothing of a massage robot according to any one of claims 1-7, and the system comprises: a point cloud clipping module for obtaining three-dimensional point cloud data of a clothing surface based on a depth camera, and performing clipping processing on the three-dimensional point cloud data to obtain clipped point cloud data; A cascade filtering module is configured to preprocess the cropped point cloud data to obtain target point cloud data, and perform cascade filtering on the target point cloud data to obtain optimized point cloud data after cascade filtering. A normal vector estimation module is configured to calculate a normal vector of each point based on the optimized point cloud data, and perform massage path planning based on the calculated normal vector to control the massage robot to perform a massage action based on the planned massage path.

9. A terminal, characterized by comprising: The terminal comprises a memory, a processor, and a three-dimensional point cloud processing program for anti-wrinkle interference clothes of a massage robot stored in the memory and executable on the processor. When the processor executes the three-dimensional point cloud processing program for anti-wrinkle interference clothes of a massage robot, the steps of the three-dimensional point cloud processing method for anti-wrinkle interference clothes of a massage robot according to any one of claims 1-7 are implemented.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a three-dimensional point cloud processing program for anti-wrinkle interference clothes of a massage robot. The three-dimensional point cloud processing program for anti-wrinkle interference clothes of a massage robot implements the steps of the three-dimensional point cloud processing method for anti-wrinkle interference clothes of a massage robot according to any one of claims 1-7 on the computer readable storage medium.

Images