CN109377527A - A kind of man-machine interaction method and switching filtering system based on 4 points of trapezoidal calibration algorithms - Google Patents
A kind of man-machine interaction method and switching filtering system based on 4 points of trapezoidal calibration algorithms Download PDFInfo
- Publication number
- CN109377527A CN109377527A CN201811024223.6A CN201811024223A CN109377527A CN 109377527 A CN109377527 A CN 109377527A CN 201811024223 A CN201811024223 A CN 201811024223A CN 109377527 A CN109377527 A CN 109377527A
- Authority
- CN
- China
- Prior art keywords
- switching
- points
- calibration algorithms
- filtering system
- trapezoidal
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
- G06T7/85—Stereo camera calibration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
- G06T2207/20028—Bilateral filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
- G06T2207/20032—Median filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20076—Probabilistic image processing
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Position Input By Displaying (AREA)
Abstract
The invention discloses a kind of man-machine interaction methods and switching filtering system based on 4 points of trapezoidal calibration algorithms, and this method comprises the following steps: Step 1: establishing the control platform based on 4 points of trapezoidal calibration algorithms on computers；Step 2: in the position coordinates of the operating point at projection four angles in interaction face and being transferred in computer using Kinect sensor detection operator；Step 3: selecting four operating points detected in step 2 in control platform carries out tracking and positioning to it, and it is corrected using 4 points of trapezoidal calibration algorithms, so that the detection range of Kinect sensor is consistent with the range that projection interacts face.The switching filtering system includes a plurality of subsystems arranged side by side and the controller for connecting and selecting an output with the output end of each subsystem by switching law.The present invention is configured co-acting planar by the human-computer interaction touch-control experiment porch of the large-size screen monitors interactive projection system of independent development and demarcates with 4 points, and then reaches precise positioning.
Description
Technical field
The present invention relates to the communications fields, and in particular to a kind of man-machine interaction method based on 4 points of trapezoidal calibration algorithms and
Switch filtering system.
Background technique
Large-screen interactive formula equipment brings a kind of past any interactive mode all never bring mystery interactions to the mankind
Experience, system can build an also true also unreal virtual scene, when participant stays wherein, so that it may directly pass through
The deliberate action of voice, limb action, gesture or even eyeball, lip is directly exchanged or is operated with virtual environment, without
It to issue order by mobile terminals such as the input equipments such as mouse, keyboard or mobile phones.Large-screen interactive formula equipment is wide
It is used in the different fields such as interactive teaching and learning, commercial advertisement, Entertainment generally, spectators can initiatively obtain letter in systems
It ceases and is immersively dissolved into the environment of setting.For user experience angle, this man-machine interaction mode can fill
Divide the enthusiasm for transferring spectators, plays good medium interaction effect.
During human-computer interaction, often will appear sensor sensing detection range in depth image interaction systems with
There is deviation in projection interaction area, this will lead to, and precision is not high, and pursuit path drifts about, so that human-computer interaction is imitated
Fruit is poor.On the other hand, the current research in this field is dedicated to developing a kind of inexpensive, convenient, accurately human-computer interaction projection
System --- the human-computer interaction optical projection system based on Kinect depth image, and Optimal improvements are made to its performance.As it is low at
Originally, the convenient and fast depth camera of operation, Kinect provide brand-new broad space for the research of large-screen interactive optical projection system,
But since the depth image of Kinect acquisition is relatively rough, included machine learning RGB-D track algorithm is also not very accurate.
Therefore, how to improve the accuracy, robustness and real-time of large-size screen monitors human-computer interaction system is a key technical problem.
In order to realize that accurate positioning and automatic calibration, the interaction range of overall system control of large screen multi-point touch are set
Setting and interacting face calibration with projection is the key that realize the precondition of stability contorting and analysis and Control system precision.How
It is scanned using 3D sensing data and realizes and be accurately positioned and to establish interactive placement corresponding relationship most important.To reach expected essence
Quasi- multi-point interactive designs the human-computer interaction touch-control experiment porch based on technologies such as infrared induction, machine vision and carries out multiple spot touching
It controls the automatic calibration of interaction area and is accurately positioned, be a challenging problem always.
Summary of the invention
Goal of the invention of the invention is to provide a kind of man-machine interaction method and switching filter based on 4 points of trapezoidal calibration algorithms
Wave system system, sets co-acting planar by the human-computer interaction touch-control experiment porch of the large-size screen monitors interactive projection system of independent development
It sets and is demarcated with 4 points, and then reach precise positioning.
To achieve the above object of the invention, the technical solution adopted by the present invention is that: it is a kind of based on 4 points of trapezoidal calibration algorithms
Man-machine interaction method includes the following steps:
Step 1: establishing the control platform based on 4 points of trapezoidal calibration algorithms on computers；
Step 2: being sat using Kinect sensor detection operator in the position of the operating point at projection four angles in interaction face
It marks and is transferred in computer；
Step 3: four operating points detected in step 2 are selected in control platform carries out tracking and positioning to it, and
It is corrected using 4 points of trapezoidal calibration algorithms, so that the range that the detection range of Kinect sensor interacts face with projection is protected
It holds consistent.
In above-mentioned technical proposal, the control platform includes the Kinect sensor detection image display window positioned at left side
Mouth and the tracking and positioning display window positioned at right side.
In above-mentioned technical proposal, the Kinect sensor detection image display window and tracking and positioning display window it
Between be additionally provided with one can up and down adjustment idler wheel.
In above-mentioned technical proposal, the Kinect sensor is connected to computer through USB data line.
A kind of switching filtering system, the switching filtering system include a plurality of subsystems arranged side by side and pass through switching law
The controller of an output is connect and selected with the output end of each subsystem, introduces pattra leaves in the switching signal of the controller output
This estimation principle.
In above-mentioned technical proposal, the switching filtering system include adaptive median filter switching subsystem arranged side by side and
Joint bilateral filtering switches subsystem.
It include mean filter in the adaptive median filter switching subsystem in above-mentioned technical proposal.
In above-mentioned technical proposal, a guiding figure is introduced in the joint bilateral filtering switching subsystem and is weighed as codomain
The calculation basis of weight.
Due to the above technical solutions, the present invention has the following advantages over the prior art:
1. the present invention is flat to interaction by the human-computer interaction touch-control experiment porch of the large-size screen monitors interactive projection system of independent development
Face is configured to be demarcated with 4 points, and then reaches precise positioning；Experiment based on 3D interaction software system simultaneously with to score
Analysis, compared to the trapezoidal calibration algorithm of traditional optical, calibration algorithm proposed by the present invention is in tracking accuracy and tracks on the time
With greater advantage, precise positioning and real-time interactive control are reached；
2. the present invention proposes one kind and is based on during based on the picture system noise filtering of 3D sensor Kinect
The interaction systems of Bayesian Estimation switching law switch filtering method, the main noise (salt-pepper noise being subject to for every frame image
And Gaussian noise), the sub- filtering system for being suitble to the noise is switched to by suitable switching law, thus image filtering system
Balance is realized between accuracy and real-time；By subjective assessment, objectively evaluate and runing time three in terms of carry out it is relevant
Emulation experiment verifying and comparative analysis, have reached good filter effect.
Detailed description of the invention
Fig. 1 is the interaction control software interface schematic diagram of the embodiment of the present invention one.
Fig. 2 is the schematic diagram of 4 points of trapezoidal calibration algorithms of the embodiment of the present invention one.
Fig. 3 to 8 is the schematic diagram that position correction is carried out to trapezoidal projection face of the embodiment of the present invention one.
Fig. 9 is the switching filtering system structural schematic diagram of the embodiment of the present invention one.
Figure 10 is the Bayesian Estimation structure chart of the embodiment of the present invention one.
Figure 11 is the on-demand sampling window schematic diagram of the embodiment of the present invention one.
Figure 12 is the improved adaptive median filter algorithm structure figure of the embodiment of the present invention one.
Specific embodiment
The invention will be further described with reference to the accompanying drawings and embodiments:
Embodiment one:
A kind of man-machine interaction method based on 4 points of trapezoidal calibration algorithms, includes the following steps:
Step 1: establishing the control platform based on 4 points of trapezoidal calibration algorithms on computers；
Step 2: being sat using Kinect sensor detection operator in the position of the operating point at projection four angles in interaction face
It marks and is transferred in computer；
Step 3: four operating points detected in step 2 are selected in control platform carries out tracking and positioning to it, and
It is corrected using 4 points of trapezoidal calibration algorithms, so that the range that the detection range of Kinect sensor interacts face with projection is protected
It holds consistent.
Above, the present invention obtains the action message of operator using Kinect sensor and is transmitted by USB data line
Into computer, content uploading will be identified extremely after interaction control software (i.e. control platform) operation of independent development on computer
Application layer triggers dependent event to realize human-computer interaction.
Human-computer interaction touch-control experiment porch of the invention by projector in any metope or curtain projection by constituting
The interaction face of projection.Wherein, the fixed range of Kinect sensor be to 1~3m above interactive screen, apart from 5~6cm of metope,
Projector is placed in immediately ahead of projection interaction face, and the distance by adjusting projector distance projection interaction face is big to adjust perspective plane
It is small.Operator station in interaction area by the click of hand, brandish, rotate etc. the interaction face of projection is operated.Kinect
Sensor connects on computers, and interaction control software interface design is as shown in Figure 1, choose monitoring tab and a left side
The Available option of inferior horn determines Maximum depth and the Minimum depth at interface according to practical operation situation
(projected base and top depth), idler wheel among the picture of left and right can up and down adjustment, determine interaction operating surface apart from metope
Distance, close to the range of metope be all effective opereating specification within the green line of appearance.It is Kinect sensor on the left of interface
Detection image display window, right side are tracking and positioning display windows, and interaction face is inverted trapezoidal.
In the present embodiment, principle such as Fig. 2 of 4 points of trapezoidal calibration algorithms shows, makees to big trapezoidal ABCD shown in Fig. 2
Differential process, is classified as N block, big trapezoidal ABCD cutting is that 4 parts are small trapezoidal by N=4 in example.Point X (x, y) be for
Operating point coordinate, point X before correcting0(x0,y0) be correction after practical operating point position.
Specific implementation step is as follows:
(1) firstly, determining quadrangle A (x to tetra- jiaos of positioning of big trapezoidal ABCD1,y1), B (x2,y1), C (x3,y2), D (x4,
y2) coordinate.
(2) it is variable by N=4, N, can thereby determines that small trapezoidal each lateral midpoint O, the y value of P, Q.
(3) secondly, repositioning, according to known O, 3 points of P, Q of y value, i.e. y7, y8, y9, by software for editing same
The left and right edges in the projection interaction face that there is distortion are detected in one y value, confirm E, F, G, H, I, six extreme coordinates of J.
To obtain the trapezoidal of 4 known endpoints.
(4) it is directed to small trapezoidal ABFE:
1. ifAndThen judge trapezoidal ABFE for inverted trapezoidal.
So interface setting condition need to meet simultaneously the following conditions:
y1<y<y7 (6)
2. ifAndThen judge trapezoidal ABFE for trapezoid.
So interface setting condition need to meet simultaneously the following conditions:
Or
Or
y1<y<y7 (9)
(5) X (x, y) is set up an office as the operating point position on not calibrated interface, and place line segment meets at point R (x with AD16, y),
It sets up an office line segment where X (x, y) and BC meets at point T (x17, y),
(6) X (x, y) that sets up an office relative position is on line segment RT:
Wherein, ratio indicates ratio.
(7) point X (x, y) is mapped on trapezoidal strigula, finds out final actual operational position X0(x0,y0):
1. point X is on line segment EF if trapezoidal ABFE is inverted trapezoidal:
y0=y (16)
2. point X is on line segment AB if trapezoidal ABFE is trapezoid:
y0=y (18)
For its excess-three small trapezoidal EFHG, GHJI, IJCD, step (4)-(7) are repeated, omits do not repeat herein
It repeats.
Specifically carrying out position correction to trapezoidal projection face is, as shown in Fig. 3 to 6, operator stands before screen, manual point
An angle (by taking the upper left corner and the upper right corner as an example) in projection throwing face is hit, interface will appear the position of corresponding interaction point on the right side of software,
It separately needs an operator to be clicked with mouse and selects the position, then click " record " key again.Successively carry out determining for four angles
Perspective plane calibration is completed, as shown in Fig. 7 in position.Final operating effect is as shown in Figure 8.
By Fig. 3 to 8 as it can be seen that 4 calibration experiments of the human-computer interaction control experiment porch progress in independent development, are realized
The projection interaction area and sensor detection zone of depth image interaction systems are consistent, ensure that interaction control and with
Track control can be realized basicly stable, realize accurately single-point touch and multiparty control on various interaction platforms for operator
Provide good guarantee condition.
Shown in Figure 9, switching system is used for the filtering of image by the present invention, by switching law determine each subsystem it
Between switching condition, to obtain edge-smoothing clearly image.It makes an uproar for the generally existing salt-pepper noise of depth image and Gauss
Joint bilateral filtering and improved adaptive median filter is respectively adopted in sound.Using filtering system as the subsystem of whole system
System is equipped with appropriate switching law, constitutes switching filtering system, which includes a plurality of subsystems arranged side by side
With the controller for connecting and selecting an output with the output end of each subsystem by switching law, the switching of the controller output
Bayesian Estimation principle is introduced in signal.
The subsystem of switching system is mostly made of the continuous differential equation or discrete differential equation, is advised by the switching in system
It then determines which subsystem is each moment switching system be switched to, and guarantees only one subsystem of each moment
It works, i.e., there is only a control laws for etching system when each.N subsystem is under the action of switching law σ (t), shape
At switching sequence it is as follows:
∑={ x0；(i0,t0),(i1,t1),…,(ij,tj),…,|tj∈M,j∈N} (19)
In formula, x0Represent the original state of system, t0Represent the initial time of system.In tj≤t<tj+1Period, activation system
The i-th of systemjA subsystem, by i-thjThe path of a subsystem generation system is unfolded according to δ x (t).Within the given time, false
If arbitrary switch occurred at the momentAssuming that the state of system does not jump perseverance surely in the moment of switching, and
The switching of limited times is carried out on section [0, t], and definition has i-thjThe minimum time interval T of a subsystemmin=min { T=
tk-tk-1}.Since the time interval continuously switched is not less than Tmin, in arbitrarily fixed switching law, it is impossible to occur
Limited time escaping phenomenon.
The present invention introduces Bayesian Estimation in switching signal and is more preferably filtered with choosing to next frame image denoising effect
Subsystem.Bayesian Estimation can be indicated by lower Figure 10.
Bayesian Estimation formula can be indicated by formula (20):
Wherein, θ represents the unknown parameter to be acquired, D={ x1,x2,…,xnThe data observed are represented, p (θ) is represented
Not considering the prior probability or marginal probability prior probability of D factor, p (D) is the marginal probability or prior probability of D, p (θ | D) generation
Conditional probability after table D generation, also referred to as posterior probability, and p (D | θ) likelihood score is represented, the condition of D is general after as known θ occurs
Rate.
Denominator is unfolded according to total probability formula, obtains formula (21):
P (D)=∫θp(D|θ)p(θ)dθ (21)
By formula (21) and then obtain:
Formula (22) are substituted into formula (20), are obtained:
Bayesian Estimation value, which can be obtained, is
Switching signal meets above-mentioned Bayesian Estimation, i.e., will be modified, and completed to original judgement using existing filtering
Anticipation to next filtering subsystem selection, accuracy, the filter effect of boostfiltering system.
For two kinds of main noises of depth image --- Gaussian noise and salt-pepper noise, the present invention have selected to improve respectively
Adaptive median filter switches subsystem and joint bilateral filtering switches subsystem.If entire filtering is regarded as one
Hybrid dynamic system, then every kind of filtering algorithm can be considered as a subsystem of the system, along with arbitrary switch rule
Then, the arbitrary switch filtering system being made of two subsystems is just constituted.
The weight of two-sided filter there are it is unstable a possibility that, lead to its adjacent edges it is possible that flop phenomenon,
In order to improve the stability of weight, present invention introduces joint two-sided filter (Joint bilateral filter, abbreviations
JBF), improve the filtering performance of image by the way that guiding figure is added as the calculation basis of codomain weight.Image processed in this way
Image can also be carried out well while filtering out noise to protect side processing.Bilateral filtering with combine the most main of bilateral filtering
Distinguish the calculation basis for exactly combining bilateral filtering one guiding figure of introducing as codomain weight.Joint bilateral filtering algorithm
It can be gone out by following table:
Wherein,Indicate the guiding figure introduced, it is similar to image to be processed.By compare introduce guiding figure with it is noisy
The edge contrast of image can effectively remove the presence in cavity, promote the precision of image.
In addition, the improved adaptive median filter that uses of the present invention is from changing self-adapting window and merged mean value filter
Two aspects of wave improve.In adaptive median filter, window size is set usually as n × n, full-size nmax×
nmax.Traditional adaptive median filter uses square window and cross window.The details that smaller filter window retains
Effect is better.In order to more fully retain the details and edge of image, therefore new self-adapting window is used to filter, such as Figure 11 institute
Show and sequentially samples.
On the other hand, there is the mean filter of greater advantage in terms of using addition removal noise to reach preferably removal and make an uproar
Sound purpose.By using revised mean filter, by SxyThe pixel that gray value in field is 0 and 255 removes.Amendment
Mean filter afterwards is indicated by following formula:
Wherein, g (s, t) is M × N image polluted by noise, gr(s, t) is residual pixel, and p is gray scale in g (s, t)
The number of pixels that value is 0, q is the number of pixels that gray value is in g (s, t) 255.
Improved filtering algorithm structure chart is as shown in figure 12, and final step is as follows:
First to each pixel zxy, B layers are calculated, if B1> 0 and B2< 0, then zxyIt is constant, otherwise turn in next step；
Secondly, A layers are calculated, if A1> 0 and A2< 0 and window size≤Smax, then z is usedmedInstead of original pixel value；If
A1≤ 0 or A2>=0, and window size is Smax, then z is usedmeaInstead of original pixel (zmeaFor revised mean filter result)；
If A1≤ 0 or A2>=0, and window size<Smax, then increase window by Figure 10 sequence, go back to step 1.
The characteristics of to system noise, selection make an uproar than being compared to the peak of image after processing.
Y-PSNR is defined as follows formula:
Wherein, the length of M representative image, the width of N representative image.It is square between image after MSE is original image and handles
Error, the filtered image of the smaller explanation of MSE value and original image are closer, illustrate filter effect more so that PSNR value is bigger
It is good.Find out the Y-PSNR of the image handled by three kinds of filtering algorithms above-mentioned respectively by Matlab software.
The Y-PSNR of each image after table 1 is tested twice
It is found that passing through joint bilateral filtering, cavity is reduced but still is existed, mainly made an uproar by Gauss comparative analysis table 1
A few frame image PSNR values of sound are larger, and the image PSNR value mainly influenced by salt-pepper noise is smaller；It is filtered by adaptive intermediate value
Wave, cavity are slightly reduced, and a few frame image PSNR values by salt-pepper noise pollution are larger, the image influenced by Gaussian noise
PSNR value is smaller；And after the switching filtering based on switching system, the PSNR value of each frame image is all than other two kinds of algorithms
Greatly, substantially 34 or so, cavity is also obviously reduced.Thus furtherly single Denoising Algorithm can only remove specific noise
The case where, and the switching that the present invention uses is filtered to the treatment effect of mixed noise obviously than any single filter
Will get well.It can be obtained from experimental data, the switching filter effect proposed by the present invention based on switching system is best.
Table 2 tests algorithms of different runing time comparison (/ms) twice
In table 2, the runing time of series filtering mutual authentication algorithms of different is added, series filtering is indicated using joint
After bilateral filtering, it is further introduced into adaptive median filter and is filtered.2 data of table explanation is calculated due to joined switching filtering
Method, although the runing time for switching filtering combines bilateral filtering algorithm than single use and adaptive median filter algorithm is long,
Time compared to the series filtering for repeating filtering is short.In comparison switching filtering algorithm can be realized Riming time of algorithm
Optimization.
In conclusion detection range and throwing of the present invention for the 3D sensor Kinect in depth image interaction systems
There is deviation in shadow interaction face, propose that a kind of trapezoidal calibration of human-computer interaction system based on 3D sensor Kinect is calculated
Method carries out four point calibration to projection interaction area using the human-computer interaction control experiment porch of independent development, solves depth map
As the inconsistent problem of the projection interaction area and sensor detection zone of interaction systems.On the other hand, The present invention gives one
Kind of the switching filtering method based on projection interaction system, separately designed the bilateral switching filtering subsystem of joint and it is improved from
Adapt to intermediate value switching filtering subsystem.For the main noise that every frame image is subject to, cut by the switching law of Bayesian Estimation
It changes and chooses the sub- filtering system for being suitble to the noise, can effectively remove noise, and then the real-time of optimal interaction system and accurate
Property.
The foregoing description of the disclosed embodiments enables those skilled in the art to implement or use the present invention.
A variety of modifications of above-described embodiment will be readily apparent to those skilled in the art, it is defined herein
General Principle can realize in other embodiments without departing from the spirit or scope of the present invention.Therefore, originally
Invention is not intended to be limited to above-described embodiment shown in this article, and is to fit to special with principles disclosed herein and novelty
The consistent widest scope of point.
Claims (8)
1. a kind of man-machine interaction method based on 4 points of trapezoidal calibration algorithms, which comprises the steps of:
Step 1: establishing the control platform based on 4 points of trapezoidal calibration algorithms on computers；
Step 2: detecting operator in the position coordinates and biography of the operating point at projection four angles in interaction face using Kinect sensor
It is defeated into computer；
Step 3: selecting four operating points detected in step 2 in control platform carries out tracking and positioning to it, and utilize
4 points of trapezoidal calibration algorithms are corrected, so that the range that the detection range of Kinect sensor interacts face with projection keeps one
It causes.
2. the man-machine interaction method according to claim 1 based on 4 points of trapezoidal calibration algorithms, it is characterised in that: the control
Platform processed includes the tracking and positioning display window positioned at the Kinect sensor detection image display window in left side and positioned at right side.
3. the man-machine interaction method according to claim 2 based on 4 points of trapezoidal calibration algorithms, it is characterised in that: described
Be additionally provided between Kinect sensor detection image display window and tracking and positioning display window one can up and down adjustment idler wheel.
4. the man-machine interaction method according to claim 1 based on 4 points of trapezoidal calibration algorithms, it is characterised in that: described
Kinect sensor is connected to computer through USB data line.
5. a kind of switching filtering system, it is characterised in that: the switching filtering system includes a plurality of subsystems arranged side by side and passes through
Switching law connect with the output end of each subsystem and selects the controller of an output, draws in the switching signal of the controller output
Enter Bayesian Estimation principle.
6. switching filtering system according to claim 1, it is characterised in that: the switching filtering system include it is arranged side by side from
It adapts to median filtering switching subsystem and joint bilateral filtering switches subsystem.
7. switching filtering system according to claim 6, it is characterised in that: the adaptive median filter switches subsystem
In include mean filter.
8. switching filtering system according to claim 6, it is characterised in that: in the joint bilateral filtering switching subsystem
Introduce calculation basis of the guiding figure as codomain weight.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811024223.6A CN109377527A (en) | 2018-09-04 | 2018-09-04 | A kind of man-machine interaction method and switching filtering system based on 4 points of trapezoidal calibration algorithms |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811024223.6A CN109377527A (en) | 2018-09-04 | 2018-09-04 | A kind of man-machine interaction method and switching filtering system based on 4 points of trapezoidal calibration algorithms |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109377527A true CN109377527A (en) | 2019-02-22 |
Family
ID=65404406
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811024223.6A Pending CN109377527A (en) | 2018-09-04 | 2018-09-04 | A kind of man-machine interaction method and switching filtering system based on 4 points of trapezoidal calibration algorithms |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109377527A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109859128A (en) * | 2019-01-21 | 2019-06-07 | 苏州大学 | A kind of interaction systems switching filtering method based on Bayesian Estimation switching law |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102679984A (en) * | 2012-05-29 | 2012-09-19 | 北京理工大学 | Finite model filtering method based on vector distance minimizing criterion |
| CN206249242U (en) * | 2016-10-26 | 2017-06-13 | 苏州大学 | Arbitrary plane projection interaction system device |
| CN107119595A (en) * | 2017-05-25 | 2017-09-01 | 向美芹 | Control method based on image procossing |
-
2018
- 2018-09-04 CN CN201811024223.6A patent/CN109377527A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102679984A (en) * | 2012-05-29 | 2012-09-19 | 北京理工大学 | Finite model filtering method based on vector distance minimizing criterion |
| CN206249242U (en) * | 2016-10-26 | 2017-06-13 | 苏州大学 | Arbitrary plane projection interaction system device |
| CN107119595A (en) * | 2017-05-25 | 2017-09-01 | 向美芹 | Control method based on image procossing |
Non-Patent Citations (1)
| Title |
|---|
| 李晨雨: "虚拟场景的3D显示控制关键技术研究", 《中国硕士学位论文全文数据库 信息科技辑》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109859128A (en) * | 2019-01-21 | 2019-06-07 | 苏州大学 | A kind of interaction systems switching filtering method based on Bayesian Estimation switching law |
| CN109859128B (en) * | 2019-01-21 | 2023-06-27 | 苏州大学 | Interaction system switching filtering method based on Bayesian estimation switching rule |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20230045386A1 (en) | Interactive and shared surfaces | |
| US10146298B2 (en) | Enhanced handheld screen-sensing pointer | |
| CN103353935B (en) | A kind of 3D dynamic gesture identification method for intelligent domestic system | |
| US8217997B2 (en) | Interactive display system | |
| US9024876B2 (en) | Absolute and relative positioning sensor fusion in an interactive display system | |
| CN107255468B (en) | Method for tracking target, target following equipment and computer storage medium | |
| US9024896B2 (en) | Identification method for simultaneously identifying multiple touch points on touch screens | |
| CN108776970A (en) | Image processing method and device | |
| US20030004678A1 (en) | System and method for providing a mobile input device | |
| KR101906634B1 (en) | Apparatus and method for providing haptic which cooperates with display apparatus | |
| CN102104767A (en) | Facial pose improvement with perspective distortion correction | |
| US8982050B2 (en) | Motion compensation in an interactive display system | |
| US8482483B2 (en) | Display control device, display, display system and computer-readable medium | |
| US20150146048A1 (en) | Radial distortion parameter acquiring method and apparatus | |
| US20190335034A1 (en) | Input method, device and system | |
| US9965102B2 (en) | Virtual keyboard system and typing method thereof | |
| CN107096230B (en) | message display processing method and device | |
| CN104793784A (en) | Simulation touch operation system and operation method based on depth data | |
| CN105306819B (en) | A kind of method and device taken pictures based on gesture control | |
| CN106558039A (en) | A kind of facial image processing method and device | |
| CN101483722A (en) | Image rectification method and related device for a video device | |
| CN108200419A (en) | A kind of projecting method and electronic equipment | |
| CN105278947A (en) | Interface element configuration method and apparatus | |
| CN106445386A (en) | Handwriting display method and device | |
| CN108176044B (en) | Method and device for acquiring virtual handle rocker signal |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190222 |