CN112315257B - Mattress system based on optimal transmission - Google Patents
Mattress system based on optimal transmission Download PDFInfo
- Publication number
- CN112315257B CN112315257B CN202011210097.0A CN202011210097A CN112315257B CN 112315257 B CN112315257 B CN 112315257B CN 202011210097 A CN202011210097 A CN 202011210097A CN 112315257 B CN112315257 B CN 112315257B
- Authority
- CN
- China
- Prior art keywords
- mattress
- human body
- posture
- model
- sleeper
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A47—FURNITURE; DOMESTIC ARTICLES OR APPLIANCES; COFFEE MILLS; SPICE MILLS; SUCTION CLEANERS IN GENERAL
- A47C—CHAIRS; SOFAS; BEDS
- A47C31/00—Details or accessories for chairs, beds, or the like, not provided for in other groups of this subclass, e.g. upholstery fasteners, mattress protectors, stretching devices for mattress nets
- A47C31/12—Means, e.g. measuring means for adapting chairs, beds or mattresses to the shape or weight of persons
- A47C31/123—Means, e.g. measuring means for adapting chairs, beds or mattresses to the shape or weight of persons for beds or mattresses
-
- A—HUMAN NECESSITIES
- A47—FURNITURE; DOMESTIC ARTICLES OR APPLIANCES; COFFEE MILLS; SPICE MILLS; SUCTION CLEANERS IN GENERAL
- A47C—CHAIRS; SOFAS; BEDS
- A47C27/00—Spring, stuffed or fluid mattresses or cushions specially adapted for chairs, beds or sofas
- A47C27/12—Spring, stuffed or fluid mattresses or cushions specially adapted for chairs, beds or sofas with fibrous inlays, e.g. made of wool, of cotton
- A47C27/122—Spring, stuffed or fluid mattresses or cushions specially adapted for chairs, beds or sofas with fibrous inlays, e.g. made of wool, of cotton with special fibres, such as acrylic thread, coconut, horsehair
-
- A—HUMAN NECESSITIES
- A47—FURNITURE; DOMESTIC ARTICLES OR APPLIANCES; COFFEE MILLS; SPICE MILLS; SUCTION CLEANERS IN GENERAL
- A47C—CHAIRS; SOFAS; BEDS
- A47C27/00—Spring, stuffed or fluid mattresses or cushions specially adapted for chairs, beds or sofas
- A47C27/14—Spring, stuffed or fluid mattresses or cushions specially adapted for chairs, beds or sofas with foamed material inlays
-
- A—HUMAN NECESSITIES
- A47—FURNITURE; DOMESTIC ARTICLES OR APPLIANCES; COFFEE MILLS; SPICE MILLS; SUCTION CLEANERS IN GENERAL
- A47C—CHAIRS; SOFAS; BEDS
- A47C31/00—Details or accessories for chairs, beds, or the like, not provided for in other groups of this subclass, e.g. upholstery fasteners, mattress protectors, stretching devices for mattress nets
- A47C31/008—Use of remote controls
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01L—MEASURING FORCE, STRESS, TORQUE, WORK, MECHANICAL POWER, MECHANICAL EFFICIENCY, OR FLUID PRESSURE
- G01L1/00—Measuring force or stress, in general
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Abstract
The invention discloses a mattress system based on optimal transmission, which comprises a human body posture monitoring module, a mattress supporting matrix, a signal acquisition module, a storage and control module and the like; on the basis of human body three-dimensional modeling, a three-dimensional human body model is flattened to a two-dimensional plane model through conformal parameterization, and then optimal transmission is utilized to realize mapping from conformal to conformal rate and conformal area. The mapping of the area can be better by the deep learning to discern every position, map the three-dimensional model with the two-dimensional texture after the mark backward finally, combine the three-dimensional model after the mark with the corresponding mattress support point of mattress support matrix just can adjust the pressure of each position, be applicable to the sleeper of various sizes. When the posture of the sleeper changes, the mattress control system can monitor the position and the sleeping posture of the human body in real time, automatically follow the human body to correspondingly adjust the supporting degree of each part of the mattress supporting matrix, maintain the optimal comfortable degree and realize the technical effect of 'the bed moves along with the human body'.
Description
Technical Field
The invention relates to the field of intelligent home furnishing, in particular to application of a full-discrete optimal transmission method in computer graphic images in an intelligent mattress control system.
Background
With the continuous improvement of living standard and health consciousness, the requirements of people on sleep quality are higher and higher. And the sleep quality has a direct relation with the body health, and modern medical research shows that the sleep support is closely related to the health of human skeleton, muscle, blood circulation and other aspects, the sleep quality is related to the mattress, and more than 90 percent of sleep problems are inappropriate sleep support related to the root mattress. The traditional mattress uses springs, sponge, latex or air bags and the like as main supporting materials, and has the defects that the fixed structure cannot adapt to the sleeping posture which is changed at any moment when people get, and the pressure distribution of the mattress is not humanized enough, so that the situations of ' waist pain when people get up, ' shoulder pain when people get up when getting up ' and ' neck pain after sleeping ' and the like can occur when people get up in the morning.
Disclosure of Invention
In order to solve the problems of pressure distribution of the existing mattress, the invention provides a mattress system based on optimal transmission, which realizes scientific modeling of dynamic pressure distribution.
In order to achieve the purpose, the invention adopts the technical scheme that:
a mattress system based on optimal transmission comprises a human body posture monitoring module, a mattress supporting matrix, a signal acquisition module, a storage and control module, a display module and a power supply module; the mattress supporting matrix is arranged at the bottom of the mattress, is controlled by the storage and control module, is used for adjusting the supporting degree of each part of the mattress to realize the adjustment of hardness, and consists of a plurality of supporting pieces which are distributed in a matrix form and can adjust the elongation;
the human body posture monitoring module is used for monitoring the contact position of a human body and the mattress and is used for judging the relative position relation between the posture of the sleeper and the mattress supporting matrix; the human body posture monitoring module comprises a camera capable of monitoring the whole mattress and a pressure sensor matrix for detecting the pressure of a human body on the mattress; the camera comprises an auxiliary fixing device which can drive the camera body to rotate around the mattress;
the input end of the signal acquisition module is connected with the output end of the human body posture monitoring module and is used for acquiring human body position posture data and pressure distribution data generated by a human body to the mattress from the human body posture monitoring module;
the input end of the storage and control module is connected with the output end of the signal acquisition module, and the output end of the storage and control module is connected with the input end of the mattress support matrix and is used for storing a sleep mode, outputting a control instruction to the mattress support matrix and outputting display data to the display module;
the power supply module is used for supplying a low-voltage direct-current power supply for the operation of each module in the mattress system;
on the basis of the modules, a non-rigid body deformation estimation is added on the basis of Kinect Fusion through human engineering mechanics modeling, a real-time scanning splicing algorithm is realized to carry out three-dimensional human body reconstruction on images continuously shot and obtained by a camera to obtain a model, and then the position of the contact of a human body and a mattress is obtained and is used as the basis for judging the relative position relation between the posture of a sleeper and a mattress support matrix; combining the pressure distribution data with the model to construct a special sleep model of the sleeper; when the posture of the sleeper changes, the mattress control system can monitor the position and the sleeping posture of the sleeper in real time, automatically adjust the supporting degree of each part of the mattress along with the posture of the sleeper, and maintain the optimal comfortable degree; calculating an initial mapping for optimal transmission; giving a target measure of optimal transmission; calculating on the optimal transmission of a target parameter domain with constant curvature to obtain final guaranteed measurement parametric mapping;
the method is characterized by comprising the following steps:
step 1: preprocessing depth data, mapping the camera internal parameters and the depth frames to a camera space function to convert a depth map into 3D point cloud, and then calculating a normal vector of each point; wherein the camera intrinsic parameters comprise fx, fy, cx, cy and 3 radial distortion parameters, and the camera space function is provided by the SDK of Kinect 2.0;
step 2: performing camera tracking, namely performing ICP (inductively coupled plasma) matching on the 3D point cloud converted by the current frame and the predicted 3D point cloud generated by the existing model so as to obtain the pose of the current camera;
and 3, step 3: depth data fusion, namely fusing the 3D point cloud of the current frame into the existing model by using a TSDF point cloud fusion algorithm according to the calculated pose of the current camera, realizing the reconstruction of a TSDF human surface model and obtaining a three-dimensional human body model;
and 4, step 4: scene rendering, namely predicting the environmental point cloud observed by the current camera by using a ray tracing computer graphics method and combining the existing model and the pose of the current camera; the obtained environmental point cloud is used for displaying and is also provided for the step 2 for ICP matching;
and 5: carrying out Ricci Flow conformal parameterization on the reconstructed three-dimensional human body model to obtain a two-dimensional plane model, and then calculating on a target parameter domain with constant curvature through optimal transmission to obtain final guaranteed measurement parameterized mapping; then, recognizing and marking different parts by using deep learning, and then remapping marked textures to a three-dimensional human body model;
step 6: the storage and control module adjusts the mattress support matrix step by step according to the pressure data acquired by combining the three-dimensional human body model, so that the overlarge pressure generated by the reaction of the original mattress on a certain part of the human body can be dispersed, the original suspended part of the human body can be supported more sufficiently, and the human body is in the most comfortable state; the storage and control module stores height change data of a mattress support matrix in the current sleeping posture, and the height change data and the human body position posture data jointly form a sleeping mode in the posture;
and 7: when the sleeper changes the sleeping posture, repeating the steps 3-6 to correct the sleeping posture so as to maintain the optimal comfortable degree;
and 8: when the sleeper leaves the bed to do other activities, the mattress system is switched to a standby mode; when the sleeper lies on the bed again, the position and posture information of the human body is collected again, the mode data of the corresponding sleep mode is called out, and the step 1 is executed again.
Preferably, in step 1, when the mattress system starts to be started, the human body posture monitoring module starts to monitor the human body position posture of the sleeper, a camera serving as a single depth sensor Kinect winds around the sleeper lying on the mattress, the signal acquisition module acquires the human body position posture data, and the mattress system adopts a Dynamic Fusion three-dimensional human body model reconstruction technology in computer graphics to reconstruct the human body surface model in real time.
Preferably, the mattress system adjusts the pressure distribution data acquired by the human body model and the signal acquisition module, and the suspended part of the human body can be fully supported by the mattress, namely the pressure generated by the protruded part of the human body on the mattress is uniformly dispersed, so that the body of a sleeper is in the most comfortable posture on the mattress; the storage and control module automatically stores height change data of the support piece in the current sleeping posture, and the height change data and the human body position posture data jointly form a sleeping mode in the posture.
Preferably, the supporting element is a hydraulic spring, the hydraulic spring can change the elongation under the drive of an electric signal, and the hardness of the corresponding position of the mattress is changed through the supporting force of the hydraulic spring; the hydraulic spring can tolerate the fluctuation of the reaction force of the mattress within the range of the set interval value, when the reaction force exceeds the range of the set interval value, the hydraulic spring can adjust the elongation amount according to the reaction force, the original elongation amount is recovered when the reaction force returns to the range of the set interval value, and a locking device for preventing the hydraulic spring from being damaged by too large external force is arranged in the hydraulic spring.
Preferably, the storage and control module may record a plurality of sleep modes, including a deep sleep mode, a light sleep mode, a night sound sleep mode, a rest in the middle of the afternoon, and a massage sleep mode.
Preferably, the storage and control module can store sleep mode data corresponding to a plurality of different sleeping postures; when the sleeper changes the sleeping posture, the storage and control module automatically matches the corresponding comfortable sleeping posture, when the sleeping posture which the sleeper is accustomed to is not matched, the storage and control module marks that the sleeper is not in a normal state, and the storage and control module sends a reminding signal to the display module.
Preferably, the mattress is made of coconut palm, memory cotton or latex.
Preferably, the ICP matching comprises the steps of: (1) Matching, for each point in the input source point cloud, a closest point in the reference point cloud; (2) Estimating a combination of rotation and translation by minimizing a root mean square point-to-point distance metric to optimally align each source point with the matching points obtained in step (1); (3) transforming the source point using the obtained transform; and (4) reselecting the association points and iterating the process.
Preferably, the TSDF surface model reconstruction includes the steps of: (1) Converting coordinates of the point cloud model after ICP matching from a world coordinate system to a camera coordinate system through a rotation matrix and a translation matrix; (2) Converting point cloud coordinates under a camera coordinate system into a screen image coordinate system through an internal reference matrix of the Kinect camera, and obtaining the depth H of each point; (3) The distance between the point cloud of different frames and the camera screen can be obtained by comparing the image depth H with the coordinate Z of the corresponding coordinate point; (4) Defining the distance as D, giving a weight W, and calculating and updating D and W under different frames; (5) Acquiring intersecting surface vertexes under multiple point clouds by using a preset distance truncation range; (6) And iterating the process to obtain final model surface point clouds to form a human surface model, and finishing the reconstruction of the TSDF human surface model.
The Kinect Fusion, which is a technology for performing real-time three-dimensional reconstruction by using the depth data of a low-cost single depth sensor (Kinect), provides a method for reconstructing a scene in real time by combining a GPU (graphics processing Unit) and a TSDF (time dependent dynamic distribution) model. The Kinect Fusion succeeds in realizing dense real-time three-dimensional reconstruction for the first time by utilizing the high parallelism of the GPU, and the reconstruction effect is very stable and excellent, and the relationship among the four modules is shown in fig. 1. Dynamic Fusion is further realized on the basis of Kinect Fusion. The Dynamic Fusion solves the problem of how to perform real-time surface reconstruction under the condition that the object to be measured moves simultaneously. Even if the surface of the object has deformation, the surface shape of the object can be accurately reduced in real time. The Dynamic Fusion transforms the dynamically changing scene (object to be reconstructed) acquired each frame into a canonical space (world space) through some transformation, i.e., creates or updates a static object surface model in the space. While each frame has a corresponding volumetric warp field, the model in canonical space can be restored to live frame (camera view space).
Further, the technical scheme of the optimal transmission adopted by the invention is as follows:
the optimal transport problem was initially used to investigate how to transfer a substance from a source site to a target site with minimal cost. This problem can be described as:
and satisfy
Wherein the content of the first and second substances,
the function c (x, y) represents the transmission cost from location x to location y; the functions u and v represent the amount of soil heap at some point within the source and target regions over region X. It can be seen that the goal of the optimal transfer problem is to find a mapping T X → X that will transfer the heap from the source area to the target area and minimize the amount of work in the heap transfer process.
The source and target may also be a set of a series of points in the course of the transmission. The function v is then no longer a function over a continuous area but a function over a finite set of points. However, when the value of the function is not 0 only at a limited position but is 0 at all other positions, the integral of the function over the area X will be 0. Instead, a measure substitution function may be used to represent the amount of material, and each point in the target may be given a dirac measure weighted by the amount of material transferred to the target point. The optimal transmission problem can then also be written as
The corresponding mass conservation condition is
v＝T # μ
The optimal transmission-based test-preserving parameterization method needs to give an initial mapping and target measurement at each vertex, and parameterization results are different according to different given measurements. The so-called measurement-preserving parameterization is to construct a plurality of different global parameterizations by controlling the measurement of the curved surface to respectively complete different tasks, such as discrete sampling by controlling the curvature measurement to construct the curvature-preserving parameterization, and texture reconstruction by controlling the surface element measurement to construct the area-preserving parameterization.
The method combines the two models together, so that the geometric shape of the model is considered while the integrated texture ensures that the area distortion is small, more weight can be obtained even in places with large wrinkle change, and the target measurement is specified according to the one-ring area and the Gaussian curvature at each vertex of the original model.
γ i ＝αK i +βA i
Wherein K i Is a vertex v i Gaussian curvature of (A) i Is a vertex v i The one-ring area, alpha and beta are coefficients, which can be adjusted according to actual conditions. Both α and β are set to 1 in the present invention. And calculating on the optimal transmission of the target parameter domain of the constant curvature to obtain the final guaranteed measurement parametric mapping. The overall process of calculating the warranty parameterization by optimal transmission is shown in fig. 5.
The method comprises the following specific steps:
(1) Calculating an initial mapping for optimal transmission;
(2) Giving a target measure of optimal transmission;
(3) And calculating on the optimal transmission of the target parameter domain of the constant curvature to obtain the final guaranteed measurement parametric mapping.
Through the reasonable arrangement of each module of mattress system, under sleeper's pressure effect, the pressure variation of mattress obtains closed loop feedback to the effort of the corresponding position of real-time adjustment mattress and human body forms the intelligent automatic pressure adjustment of mattress. The posture and the pressure distribution of the sleeper on the mattress can be quickly monitored in real time, and meanwhile, the pressure at each part on the mattress can be fed back and adjusted, so that the satisfied and comfortable pressure distribution state of the sleeper is achieved.
On the basis of human body three-dimensional modeling, a three-dimensional human body model is flattened to a two-dimensional plane model through conformal parameterization, and then optimal transmission is utilized to realize mapping from conformal to conformal rate and conformal area. The mapping of the preserved areas can be better recognized by deep learning of each part, such as the neck, the elbow and the like. And finally, reversely mapping the marked two-dimensional texture to the three-dimensional model, and combining the marked three-dimensional model with corresponding mattress supporting points of the mattress supporting matrix to adjust the pressure of each part, so that the method is suitable for sleepers with various body types. When the posture of the sleeper changes, the mattress control system can monitor the position and the sleeping posture of the human body in real time, automatically follow the human body to correspondingly adjust the supporting degree of each part of the mattress supporting matrix, maintain the optimal comfort degree and realize the highest natural state.
From the above, compared with the prior art, the present invention has the following advantages: the invention innovatively provides a model building method through human engineering mechanics according to real-time pressure distribution data in the sleeping process; on the basis of Kinect Fusion, non-rigid deformation estimation is added, three-dimensional human body reconstruction is carried out on images obtained by continuous shooting of a camera through a real-time scanning splicing algorithm to obtain a model, the three-dimensional human body model is flattened to a two-dimensional plane model through conformal parameterization, and then optimal transmission is utilized to realize mapping from a conformal to a conformal curvature and a conformal area; further obtaining the contact position of the human body and the mattress, and using the contact position as a basis for judging the relative position relation between the posture of the sleeper and the mattress supporting matrix; combining the pressure distribution data with the model to construct a special sleep model of the sleeper; when the posture of the sleeper changes, the mattress control system can monitor the position and the sleeping posture of the sleeper in real time, automatically adjust the supporting degree of each part of the mattress along with the posture of the sleeper, ensure that the contact area between a human body and the mattress is maximized at any time, maintain the optimal comfort degree and realize the technical effect of 'moving the bed along with the person'.
Drawings
FIG. 1 is a schematic diagram of a Kinect Fusion module structure.
Fig. 2 is a flow chart of the ICP algorithm of the present invention.
Fig. 3 is a block diagram of the mattress system of the present invention.
Fig. 4 is a flow chart of the SDF algorithm of the system of the present invention.
Fig. 5 is a schematic diagram of the overall calculation process of the guaranteed measurement parameterization based on the optimal transmission.
The reference numbers illustrate: the system comprises a human body posture monitoring module, a 2-pressure support matrix, a 3-signal acquisition module, a 4-storage and control module, a 5-display module and a 6-power supply module.
Detailed Description
The invention and its advantageous technical effects are explained in further detail below with reference to the drawings and preferred embodiments.
Referring to fig. 1 to 5, a mattress system based on optimal transmission according to a preferred embodiment of the present invention includes a human body posture monitoring module 1, a mattress support matrix 2, a signal acquisition module 3, a storage and control module 4, a display module 5, and a power supply module 6; the mattress support matrix 2 is arranged at the bottom of the mattress, is controlled by the storage and control module 4, is used for adjusting the support degree of each part of the mattress to realize the adjustment of hardness, and consists of a plurality of support pieces which are distributed in a matrix form and can adjust the elongation;
the human body posture monitoring module 1 is used for monitoring the contact position of a human body and a mattress and is used for judging the relative position relation between the posture of a sleeper and the mattress supporting matrix 2; the human body posture monitoring module 1 comprises a camera capable of monitoring the whole mattress and a pressure sensor matrix for detecting the pressure of a human body on the mattress; the camera comprises an auxiliary fixing device which can drive the camera body to rotate around the mattress;
the input end of the signal acquisition module 3 is connected with the output end of the human body posture monitoring module 1 and is used for acquiring human body position posture data and acquiring pressure distribution data generated by a human body on the mattress from the human body posture monitoring module 1;
the input end of the storage and control module 4 is connected with the output end of the signal acquisition module 3, and the output end of the storage and control module is connected with the input end of the mattress support matrix 2, so that the storage and control module is used for storing a sleep mode, outputting a control instruction to the mattress support matrix 2 and outputting display data to the display module 5;
the power supply module 6 is used for supplying a low-voltage direct-current power supply to the operation of each module in the mattress system;
on the basis of the modules, a non-rigid body deformation estimation is added on the basis of Kinect Fusion through human engineering mechanics modeling, a real-time scanning splicing algorithm is realized to carry out three-dimensional human body reconstruction on images continuously shot and obtained by a camera to obtain a model, and then the position of the contact of a human body and a mattress is obtained and is used as the basis for judging the relative position relation between the posture of a sleeper and the mattress support matrix 2; combining the pressure distribution data with the model to construct a special sleep model of the sleeper; when the posture of the sleeper changes, the mattress control system can monitor the position and the sleeping posture of the sleeper in real time, automatically adjust the supporting degree of each part of the mattress along with the posture of the sleeper, and maintain the optimal comfortable degree; calculating an initial mapping for optimal transmission; giving a target measure of optimal transmission; calculating on the optimal transmission of a target parameter domain with constant curvature to obtain final guaranteed measurement parametric mapping;
the method is characterized by comprising the following steps:
step 1: preprocessing depth data, namely mapping the camera internal parameters and depth frames to a camera space function to convert a depth map into 3D point cloud, and then calculating a normal vector of each point; wherein the camera internal parameters comprise fx, fy, cx, cy and 3 radial distortion parameters, and the camera space function is provided by the SDK of Kinect 2.0;
step 2: performing camera tracking, namely performing ICP (inductively coupled plasma) matching on the 3D point cloud converted by the current frame and the predicted 3D point cloud generated by the existing model so as to obtain the pose of the current camera; wherein, ICP is an abbreviation of Iterative Closest Point cloud matching algorithm, and the flow of ICP algorithm is shown in FIG. 2; preferably, the ICP matching comprises the steps of: (1) For each point in the input source point cloud (usually from a dense set of entire vertices or a selection of vertex pairs for each model) matching the closest point in the reference point cloud; (2) Estimating a combination of rotation and translation by minimizing a root mean square point-to-point distance metric to optimally align each source point with the matching points obtained in step (1); (3) transforming the source point using the obtained transform; (4) reselecting the association points and iterating the process;
and step 3: depth data fusion, namely fusing the 3D point cloud of the current frame into the existing model by using a TSDF point cloud fusion algorithm according to the calculated pose of the current camera, realizing the reconstruction of a TSDF human surface model and obtaining a three-dimensional human body model; the TSDF is an abbreviation of Truncated signaled Distance Functions, and refers to a Truncated symbolic Distance function, which allows integration of multiple depth images acquired from different viewpoints; the TSDF is different from the Kinect Fusion, as shown in FIG. 4, the Dynamic Fusion introduces a productive TSDF, which is abbreviated as PSDF, and actually calculates the TSDF value under the camera space; preferably, the TSDF artifact surface model reconstruction includes the steps of: (1) Converting coordinates of the point cloud model after ICP matching from a world coordinate system to a camera coordinate system through a rotation matrix and a translation matrix; (2) Converting point cloud coordinates under a camera coordinate system into a screen image coordinate system through an internal reference matrix of the Kinect camera, and obtaining the depth H of each point; (3) The distance between the point cloud of different frames and the camera screen can be obtained by comparing the image depth H with the coordinate Z of the corresponding coordinate point; (4) Defining the distance as D, giving a weight W, and calculating and updating D and W under different frames; (5) Acquiring intersecting surface vertexes under multiple point clouds by using a preset distance truncation range; (6) Iterating the above process to obtain final model surface point cloud to form a human surface model, and completing TSDF human surface model reconstruction;
and 4, step 4: scene rendering, namely predicting the environmental point cloud observed by the current camera by using a computer graphics method of ray tracing and combining the existing model and the pose of the current camera; the obtained environmental point cloud is used for displaying and is also provided for the step 2 for ICP matching;
and 5: carrying out Ricci Flow conformal parameterization on the reconstructed three-dimensional human body model to obtain a two-dimensional plane model, and then calculating on a target parameter domain with constant curvature through optimal transmission to obtain final guaranteed measurement parameterized mapping; then, recognizing and marking different parts by using deep learning, and then remapping marked textures to a three-dimensional human body model;
step 6: the storage and control module 4 adjusts the mattress support matrix 2 step by step according to the collected pressure data of the three-dimensional human body model, so that the overlarge pressure generated by the reaction of the original mattress on a certain part of the human body can be dispersed, the original suspended part of the human body can be supported more sufficiently, and the human body is in the most comfortable state; the storage and control module 4 stores height change data of the mattress support matrix 2 in the current sleeping posture, and the height change data and the human body position posture data jointly form a sleeping mode in the posture;
and 7: when the sleeper changes the sleeping posture, repeating the steps 3-6 to correct the sleeping posture so as to maintain the optimal comfortable degree;
and 8: when the sleeper leaves the bed to do other activities, the mattress system is switched to a standby mode; when the sleeper lies on the bed again, the position and posture information of the human body is collected again, the mode data of the corresponding sleep mode is called out, and the step 1 is executed again.
Preferably, the supporting element is a hydraulic spring, the hydraulic spring can change the elongation under the drive of an electric signal, and the hardness of the corresponding position of the mattress is changed through the supporting force of the hydraulic spring; the hydraulic spring can tolerate the fluctuation of the reaction force of the mattress within the range of the set interval value, when the reaction force exceeds the range of the set interval value, the hydraulic spring can adjust the elongation amount according to the reaction force, the original elongation amount is recovered when the reaction force returns to the range of the set interval value, and a locking device for preventing the hydraulic spring from being damaged by too large external force is arranged in the hydraulic spring.
Preferably, the storage and control module 4 may record several sleep modes, including a deep sleep mode, a light sleep mode, a night sound sleep mode, a nap mode, and a massage sleep mode.
Preferably, the storage and control module 4 can store sleep mode data corresponding to a plurality of different sleeping postures; when the sleeper changes the sleeping posture, the storage and control module 4 automatically matches the corresponding comfortable sleeping posture, when the sleeping posture which the sleeper is accustomed to is not matched, the storage and control module 4 marks that the sleeper is not in a normal state, and the storage and control module 4 sends a reminding signal to the display module 5.
Preferably, the mattress is made of coconut palm, memory cotton or latex.
When the device is used, the signal acquisition module 3, the storage and control module 4, the display module 5 and the power supply module 6 are integrated into a whole to form a controller, the controller is arranged on or near a mattress and is electrically connected with the human body posture monitoring module 1 and the relevant interfaces of the mattress support matrix 2 according to specific wiring requirements, and the various algorithms or control processes are stored in the storage and control module 4 in the form of control execution programs, such as a program memory at the periphery of an MCU or an MCU. The mattress system has the following working procedures:
(1) When the mattress system starts to be started, the human body posture monitoring module 1 starts to monitor the human body position posture of a sleeper, a camera serving as a single depth sensor Kinect winds around the sleeper lying on the mattress, and the signal acquisition module acquires human body position posture data; the mattress system adopts a Dynamic Fusion three-dimensional human body model reconstruction technology in computer graphics to reconstruct a human surface model in real time, flattens the three-dimensional human body model to a two-dimensional plane model through conformal parameterization, and then realizes the mapping of the conformal to the combination of conformal rate and conformal area by utilizing optimal transmission;
(2) The mattress system adjusts the pressure distribution data acquired by the human body model and the signal acquisition module 3, and the suspended part of the human body can be fully supported by the mattress, namely the pressure generated by the protruded part of the human body on the mattress is uniformly dispersed, so that the contact area of the human body and the mattress is increased, the reaction force of the mattress on the protruded part of the human body is reduced, and the body of a sleeper is in the most comfortable posture on the mattress; the storage and control module 4 automatically stores the height change data of the support piece in the current sleeping posture, and the height change data and the human body position posture data jointly form a sleeping mode in the posture.
(3) And (3) repeating the steps 1 and 2 when the sleeper changes the sleeping posture, so that the sleeping posture is corrected to maintain the optimal comfort level.
(4) When the sleeper leaves the bed for other activities (e.g., goes to a restroom), the mattress system switches to a standby mode; when the sleeper lies on the bed again, the position and posture information of the human body is collected again, and relevant sleep mode data is called out and executed again. The mattress system can quickly adapt to the same sleeping posture of the same sleeper without repeated adjustment, and has high execution efficiency.
In the above description, the conventional algorithms, physical structures and processes used in the prior art are not described in detail for saving space. Processing the undisclosed processing technique and parts according to the conventional technology in the prior art.
The invention is not limited in any way by the above description and the specific examples, which are not limited to the specific embodiments disclosed and described above, but rather, several modifications and variations of the invention are possible within the scope of the invention as defined in the claims.
Claims (8)
1. A mattress system based on optimal transmission comprises a human body posture monitoring module (1), a mattress supporting matrix (2), a signal acquisition module (3), a storage and control module (4), a display module (5) and a power supply module (6); the mattress supporting matrix (2) is arranged at the bottom of the mattress, is controlled by the storage and control module (4), is used for adjusting the supporting degree of each part of the mattress to realize the adjustment of hardness, and consists of a plurality of supporting pieces which are distributed in a matrix manner and can adjust the elongation;
the human body posture monitoring module (1) is used for monitoring the contact position of a human body and the mattress as a relative position relation between the posture of the sleeper and the mattress supporting matrix (2); the human body posture monitoring module (1) comprises a camera capable of monitoring the whole mattress and a pressure sensor matrix for detecting the pressure of a human body on the mattress; the camera comprises an auxiliary fixing device which can drive the camera body to rotate around the mattress;
the input end of the signal acquisition module (3) is connected with the output end of the human body posture monitoring module (1) and is used for acquiring human body position posture data and acquiring pressure distribution data generated by a human body on the mattress from the human body posture monitoring module (1);
the input end of the storage and control module (4) is connected with the output end of the signal acquisition module (3), and the output end of the storage and control module is connected with the input end of the mattress support matrix (2) and is used for storing a sleep mode, outputting a control instruction to the mattress support matrix (2) and outputting display data to the display module (5);
the power supply module (6) is used for supplying a low-voltage direct-current power supply to the operation of each module in the mattress system;
on the basis of the modules, a non-rigid body deformation estimation is added on the basis of Kinect Fusion through human engineering mechanics modeling, a real-time scanning splicing algorithm is realized to carry out three-dimensional human body reconstruction on images continuously shot and obtained by a camera to obtain a model, and then the position of the contact of a human body and a mattress is obtained and is used as the basis for judging the relative position relation between the posture of a sleeper and the mattress support matrix (2); combining the pressure distribution data with the model to construct a special sleep model of the sleeper; when the posture of the sleeper changes, the mattress control system can monitor the position and the sleeping posture of the sleeper in real time, automatically adjust the supporting degree of each part of the mattress along with the posture of the sleeper, and maintain the optimal comfortable degree; calculating an initial mapping for optimal transmission; giving a target measure of optimal transmission; calculating on the optimal transmission of a target parameter domain with constant curvature to obtain final guaranteed measurement parametric mapping;
the method is characterized by comprising the following steps:
step 1: preprocessing depth data, namely mapping the camera internal parameters and depth frames to a camera space function to convert a depth map into 3D point cloud, and then calculating a normal vector of each point; wherein the camera internal parameters comprise fx, fy, cx, cy and 3 radial distortion parameters, and the camera space function is provided by the SDK of Kinect 2.0;
step 2: performing camera tracking, namely performing ICP (inductively coupled plasma) matching on the 3D point cloud converted by the current frame and the predicted 3D point cloud generated by the existing model so as to obtain the pose of the current camera;
and step 3: depth data fusion, namely fusing the 3D point cloud of the current frame into the existing model by using a TSDF point cloud fusion algorithm according to the calculated pose of the current camera, realizing the reconstruction of a TSDF human surface model and obtaining a three-dimensional human body model;
and 4, step 4: scene rendering, namely predicting the environmental point cloud observed by the current camera by using a computer graphics method of ray tracing and combining the existing model and the pose of the current camera; the obtained environmental point cloud is used for displaying and is also provided for the step 2 for ICP matching;
and 5: carrying out Ricci Flow conformal parameterization on the reconstructed three-dimensional human body model to obtain a two-dimensional plane model, and then calculating on a target parameter domain with constant curvature through optimal transmission to obtain final guaranteed measurement parameterized mapping; then, recognizing and marking different parts by using deep learning, and remapping marked textures to a three-dimensional human body model;
step 6: the storage and control module (4) combines the three-dimensional human body model with the acquired pressure data to adjust the mattress support matrix (2) step by step, so that the excessive pressure generated by the reaction of the original mattress on a certain part of the human body can be dispersed, the original suspended part of the human body can be supported more sufficiently, and the human body is in the most comfortable state; the storage and control module (4) stores height change data of the mattress support matrix (2) in the current sleeping posture, and the height change data and the human body position posture data jointly form a sleeping mode in the posture;
and 7: when the sleeper changes the sleeping posture, repeating the steps 3-6, so that the sleeping posture is corrected to maintain the optimal comfort degree;
and 8: when the sleeper leaves the bed to do other activities, the mattress system is switched to a standby mode; when the sleeper lies on the bed again, the position and posture information of the human body is collected again, the mode data of the corresponding sleep mode is called out, and the step 1 is executed again;
in the step 1, when the mattress system starts to be started, the human body posture monitoring module (1) starts to monitor the human body position posture of a sleeper, a camera serving as a single depth sensor Kinect winds around the sleeper lying on the mattress, the human body position posture data are collected through the signal collecting module (3), and the mattress system adopts a Dynamic Fusion three-dimensional human body model reconstruction technology in computer graphics to reconstruct a human body surface model in real time.
2. The mattress system based on optimal transport of claim 1, wherein: the mattress system adjusts the human body model by combining the pressure distribution data acquired by the signal acquisition module (3), the suspended part of the human body can be fully supported by the mattress originally, namely, the pressure generated by the protruded part of the human body on the mattress is uniformly dispersed, so that the body of a sleeper is in the most comfortable posture on the mattress; the storage and control module (4) automatically stores height change data of the support piece in the current sleeping posture, and the height change data and the human body position posture data jointly form a sleeping mode in the posture.
3. The mattress system based on optimal transport of claim 2, wherein: the supporting piece is a hydraulic spring, the extension amount of the hydraulic spring can be changed under the driving of an electric signal, and the hardness of the corresponding position of the mattress is changed through the supporting force of the hydraulic spring; the hydraulic spring can tolerate the fluctuation of the reaction force of the mattress within the range of the set interval value, when the reaction force exceeds the range of the set interval value, the hydraulic spring can adjust the elongation amount according to the reaction force, the original elongation amount is recovered when the reaction force returns to the range of the set interval value, and a locking device for preventing the hydraulic spring from being damaged by too large external force is arranged in the hydraulic spring.
4. The mattress system based on optimal transport of claim 3, wherein: the storage and control module (4) can record a plurality of sleep modes, wherein the sleep modes comprise a deep sleep mode, a light sleep mode, a night sound sleep mode, a rest mode in the middle of the noon and a massage sleep mode.
5. The optimal transfer based mattress system of claim 4, wherein: the storage and control module (4) can store a plurality of sleep mode data corresponding to different sleeping postures; when the sleeper changes the sleeping posture, the storage and control module (4) automatically matches the corresponding comfortable sleeping posture, when the sleeping posture which the sleeper is accustomed to is not matched, the storage and control module (4) marks that the sleeper is not in a normal state, and sends a reminding signal to the display module (5).
6. The mattress system based on optimal transport of claim 5, wherein: the mattress is made of coconut palm, memory cotton or latex.
7. An optimal transfer based mattress system according to claim 1 or 2 wherein: the ICP matching comprises the following steps: (1) Matching, for each point in the input source point cloud, a closest point in the reference point cloud; (2) Estimating a combination of rotation and translation by minimizing a root mean square point-to-point distance metric to optimally align each source point with the matching points obtained in step (1); (3) transforming the source point using the obtained transform; and (4) reselecting the association points and iterating the process.
8. An optimal transfer based mattress system according to claim 1 or 2 wherein: the TSDF surface model reconstruction method comprises the following steps: (1) Converting coordinates of the point cloud model after ICP matching from a world coordinate system to a camera coordinate system through a rotation matrix and a translation matrix; (2) Converting point cloud coordinates under a camera coordinate system into a screen image coordinate system through an internal reference matrix of the Kinect camera, and obtaining the depth H of each point; (3) The distance between the point cloud of different frames and the camera screen can be obtained by comparing the image depth H with the coordinate Z of the corresponding coordinate point; (4) Defining the distance as D, giving a weight W, and calculating and updating D and W under different frames; (5) Acquiring intersecting surface vertexes under multiple point clouds by using a preset distance truncation range; (6) And iterating the process to obtain final model surface point clouds to form a human surface model, and finishing the reconstruction of the TSDF human surface model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011210097.0A CN112315257B (en) | 2020-11-03 | 2020-11-03 | Mattress system based on optimal transmission |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011210097.0A CN112315257B (en) | 2020-11-03 | 2020-11-03 | Mattress system based on optimal transmission |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112315257A CN112315257A (en) | 2021-02-05 |
| CN112315257B true CN112315257B (en) | 2023-04-07 |
Family
ID=74323539
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011210097.0A Active CN112315257B (en) | 2020-11-03 | 2020-11-03 | Mattress system based on optimal transmission |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112315257B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115278081B (en) * | 2022-07-29 | 2023-05-26 | 慕思健康睡眠股份有限公司 | Sleeping gesture reproduction method, system and storage medium |
| CN116098417A (en) * | 2022-08-23 | 2023-05-12 | 漳州松霖智能家居有限公司 | Method for forming sleeping gesture animation and intelligent mattress |
| CN117740281B (en) * | 2024-02-19 | 2024-04-26 | 苏州市产品质量监督检验院(苏州市质量技术监督综合检验检测中心、苏州市质量认证中心) | Mattress spring compression test system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2012010098A1 (en) * | 2010-07-22 | 2012-01-26 | 东莞市鹏康电子科技有限公司 | An intelligent temperature-adjusting mattress |
| CN108897930A (en) * | 2018-06-14 | 2018-11-27 | 暨南大学 | A method of the contact stress between prediction human body and mattress |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108618465B (en) * | 2018-04-18 | 2020-07-21 | 浙江想能云软件股份有限公司 | Soft and hard adjustable mattress for realizing intelligent learning and hardness adjusting system and method thereof |
| CN108875813B (en) * | 2018-06-04 | 2021-10-08 | 北京工商大学 | Three-dimensional grid model retrieval method based on geometric image |
| CN108937325B (en) * | 2018-08-01 | 2021-04-06 | 浙江想能睡眠科技股份有限公司 | Soft and hard adjustable mattress adapting to human body sleeping posture curve and adjusting method thereof |
| CN111642933A (en) * | 2020-06-30 | 2020-09-11 | 广东珞珈睡眠科技有限公司 | Pressure distribution monitoring feedback system of intelligent mattress |
-
2020
- 2020-11-03 CN CN202011210097.0A patent/CN112315257B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2012010098A1 (en) * | 2010-07-22 | 2012-01-26 | 东莞市鹏康电子科技有限公司 | An intelligent temperature-adjusting mattress |
| CN108897930A (en) * | 2018-06-14 | 2018-11-27 | 暨南大学 | A method of the contact stress between prediction human body and mattress |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112315257A (en) | 2021-02-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112315257B (en) | Mattress system based on optimal transmission | |
| CN112199846A (en) | System for analyzing and customizing mattress based on three-dimensional human body reconstruction technology | |
| CN109377557B (en) | Real-time three-dimensional face reconstruction method based on single-frame face image | |
| CN107028427B (en) | Novel sensing method of sleep system capable of automatically and intelligently adjusting | |
| CN108618465B (en) | Soft and hard adjustable mattress for realizing intelligent learning and hardness adjusting system and method thereof | |
| KR20110070058A (en) | Method and apparatus for motion capture of dynamic object | |
| CN107689254B (en) | Digital generation method for outer surface of full-crown prosthesis | |
| CN111091624A (en) | Method for generating high-precision drivable human face three-dimensional model from single picture | |
| WO2021204449A1 (en) | Method for eye gaze tracking | |
| CN115951784A (en) | Dressing human body motion capture and generation method based on double nerve radiation fields | |
| JP2003271933A (en) | Face detector, face detecting method, and robot device | |
| CN110148209B (en) | Human body model generation method, image processing device and device with storage function | |
| Boisvert et al. | Principal spine shape deformation modes using Riemannian geometry and articulated models | |
| Cheng et al. | An integrated approach to 3D face model reconstruction from video | |
| EP1565892B1 (en) | Virtual model generation | |
| CN116152440A (en) | Dynamic human body reconstruction method based on nerve radiation field | |
| Starck et al. | Animated statues | |
| CN113362311A (en) | Deep generation network assisted functional full crown prosthesis form generation method | |
| CN113223188B (en) | Video face fat and thin editing method | |
| CN105303611A (en) | Virtual human limb modeling method based on revolution paraboloids | |
| CN116805345A (en) | Three-dimensional human body reconstruction method and system capable of shaping nerve radiation representation | |
| CN210721364U (en) | Motion recognition tracking platform | |
| Zhang et al. | Synthesis of 3D faces using region‐based morphing under intuitive control | |
| Yano et al. | A facial expression parameterization by elastic surface model | |
| CN115619967A (en) | Parameterized human body model reconstruction method based on Chinese database |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |