CN109177175A - A kind of 3D printing spray head end real-time tracking localization method - Google Patents
A kind of 3D printing spray head end real-time tracking localization method Download PDFInfo
- Publication number
- CN109177175A CN109177175A CN201810748743.5A CN201810748743A CN109177175A CN 109177175 A CN109177175 A CN 109177175A CN 201810748743 A CN201810748743 A CN 201810748743A CN 109177175 A CN109177175 A CN 109177175A
- Authority
- CN
- China
- Prior art keywords
- printing
- spray head
- image
- edge image
- localization method
- 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
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B29—WORKING OF PLASTICS; WORKING OF SUBSTANCES IN A PLASTIC STATE IN GENERAL
- B29C—SHAPING OR JOINING OF PLASTICS; SHAPING OF MATERIAL IN A PLASTIC STATE, NOT OTHERWISE PROVIDED FOR; AFTER-TREATMENT OF THE SHAPED PRODUCTS, e.g. REPAIRING
- B29C64/00—Additive manufacturing, i.e. manufacturing of three-dimensional [3D] objects by additive deposition, additive agglomeration or additive layering, e.g. by 3D printing, stereolithography or selective laser sintering
- B29C64/30—Auxiliary operations or equipment
- B29C64/386—Data acquisition or data processing for additive manufacturing
- B29C64/393—Data acquisition or data processing for additive manufacturing for controlling or regulating additive manufacturing processes
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B33—ADDITIVE MANUFACTURING TECHNOLOGY
- B33Y—ADDITIVE MANUFACTURING, i.e. MANUFACTURING OF THREE-DIMENSIONAL [3-D] OBJECTS BY ADDITIVE DEPOSITION, ADDITIVE AGGLOMERATION OR ADDITIVE LAYERING, e.g. BY 3-D PRINTING, STEREOLITHOGRAPHY OR SELECTIVE LASER SINTERING
- B33Y50/00—Data acquisition or data processing for additive manufacturing
- B33Y50/02—Data acquisition or data processing for additive manufacturing for controlling or regulating additive manufacturing processes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/168—Segmentation; Edge detection involving transform domain methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- 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
- G06T2207/20048—Transform domain processing
- G06T2207/20061—Hough transform
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30241—Trajectory
Abstract
The present invention provides a kind of 3D printing spray head end real-time tracking localization methods, input a model to mechanical arm first；Secondly the inclined direction of detection mechanical arm, and start four magazine two cameras for being right against inclined direction for being suspended on surrounding, form binocular vision system, printing head is tracked with track algorithm, obtain the ROI area-of-interest of tracking, and classify to the image in the region with kmeans algorithm, and obtain the image of the class where spray head, by the outer profile and abscissa that obtain printing end with canny algorithm, hough line detection method is used later, straight-line detection is carried out to the edge image, and judge the straight line on spray head both sides, calculate the midpoint that straight line and edge image abscissa cross, as tracing positional.The present invention, come the tracking and positioning of real-time monitoring 3D printing spray head end, reaches feedback by the method for vision, and amendment printing in real time, improves printing precision.
Description
Technical field
The present invention relates to a kind of 3D printing techniques, and in particular to a kind of new 3D printing spray head end real-time tracking positioning side
Method.
Background technique
3D printing technique refers to be superimposed by continuous physical layer, increases material successively to generate the technology of 3D solid,
It is different from traditional removal materials processing technology, therefore also known as addition manufacture (AM, Additive Manufacturing).Make
For a kind of comprehensive application technology, 3D printing combines digital modeling techniques, Electromechanical Control technology, information technology, material science
With all various cutting edge technology knowledge such as chemistry, there is very high scientific and technological content.
3D printing technique appears in mid-term the 1990s, actually most using the technologies such as photocuring and paper layer be folded
New rapid molding device.By the exploration and development of more than ten years, 3D printing technique has significant progress, has been able at present
The fine-resolution of 600dpi is realized on the thickness in monolayer of 0.01mm.More advanced product may be implemented per hour in the world at present
The vertical speed rate of 25mm thickness, and can realize the colour print of 24 colors.Since the 1990s, domestic more colleges and universities
The independent research of 3D printing technique is carried out.Tsinghua University is in sides such as modern times molding theory, layer separated growth, FDM techniques
There is certain Superiority of Scientific Research in face；The Central China University of Science and Technology is advantageous in layer separated growth process aspect, and HRP system has been introduced
Column molding machine and moulding material；Xi'an Communications University's independent development three-dimensional printer spray head, and develop Stereolithography system
System and corresponding shaping material, formed precision reach 0.2mm；Chinese University of Science and Technology voluntarily has developed eight spray heads combination injection apparatus,
It is expected to be applied in micro manufacturing, field of photoelectric devices.But in general, domestic 3D printing technique R & D Level is compared with foreign countries
There are also larger gaps.
Currently, the method that 3D printing equipment is mainly printed using three coordinates, the model of input is considered preferably,
But due to the step of lacking real-time monitoring correction, the equipment precision of printing is all less high.
Therefore, it is urgent to provide one kind can timely feedback printing monitoring method, detects the spray head of mechanical arm printing end
Location information, to improve 3D printing precision.
Summary of the invention
The object of the present invention is to provide a kind of 3D printing spray head end real-time tracking localization methods, so that printing head end
Position can accurately and effectively detected in real time, to make up the deficiencies in the prior art.
In order to achieve the above objectives, the present invention take the specific technical proposal is:
A kind of 3D printing spray head end real-time tracking localization method inputs a model to mechanical arm first；Secondly it detects
The inclined direction of mechanical arm, and start four magazine two cameras for being right against inclined direction for being suspended on surrounding, composition
Binocular vision system tracks printing head with track algorithm, obtains the ROI area-of-interest of tracking, the area Bing Duigai
The image in domain is classified with kmeans algorithm, and obtains the image of the class where spray head, by obtaining with canny algorithm
The outer profile and abscissa of end are printed, uses hough line detection method later, straight-line detection is carried out to the edge image,
And judge the straight line on spray head both sides, by mathematical computations, the midpoint that straight line and edge image abscissa cross is calculated, as
Tracing positional.
Further, before opening camera, it is firstly received the ideal position of mechanical arm tail end, using the position as image
Active position, be set as initial ROI region, and send in the feature vector of track algorithm, and to original video frame carry out
Scaling, while initial ROI region is also zoomed in and out.
Further, the track algorithm is correlation filtering target tracking algorism, the process of this method are as follows: next frame, it should
Extracted region Hog feature of the method first to multiple surroundings of selected ROI region, then solution next frame is carried out with circular matrix
Selected ROI region, when obtaining a new selected ROI region, we are by the scaling of leading portion, the area zoom
To former ratio, and tracking box is shown in original image.
Further, when one new selected ROI region of acquisition, then with kmeans method progress classification processing.
Further, it is found through experiments that selected ROI region is divided into 3 classes and can preferably distinguish spray head end, prints disk
Face and printed matter, and the classification of printing head is exactly the second class of kmeans function category processing, i.e., only extracts the second class,
Remaining classification setting is white；Due to classification processing, it is dry that second sorted image of acquisition can preferably mask noise
It disturbs, then the method for directlying adopt canny detection can effectively obtain the edge image of printing head end.
Further, for edge image obtained above, data of the point set of edge image as hough straight-line detection
Point carries out hough straight-line detection.
Further, since edge image is rough, the item number of the straight line of detection is more, and houghlineP function is to straight line
Detection in, can accurate extraction spray head two when accumulative threshold value is set as 30, when the threshold value between two straight lines is set as 10
The straight line on side；By two straight lines of fitting, the abscissa to cross is calculated according to the minimum point abscissa of edge image respectively, it is horizontal
The midpoint of coordinate is that we want the position of tracking and positioning.
The advantages of the present invention:
The present invention is solved according to the ideal initial point that ideal model calculates in input computer, and according to the point
The region Initial R OI；To the kmeans classification processing of ROI region, the class where spray head is extracted, and extracts edge image；Hough
Straight line fitting, and judge the straight line on spray head both sides.Compared with prior art, the present invention can play good feedback effects, simultaneously
It can be reduced the interference of spray head end substance, navigate to the final word of printing head in time, play the rail of real time correction 3D printing
Mark.The present invention, come the tracking and positioning of real-time monitoring 3D printing spray head end, reaches feedback by the method for vision, and corrects in real time
Printing improves printing precision.
Detailed description of the invention
Fig. 1 is basic flow chart of the invention.
Fig. 2 is that the classification results of printing head in embodiment 1 extract figure.
Fig. 3 is the Outside contour extraction figure of printing head in embodiment 1.
Fig. 4 is the result figure of printing head end tracking in embodiment 1.
Specific embodiment
The present invention is explained further and is illustrated below by way of specific embodiment and in conjunction with attached drawing.
Embodiment 1:
A kind of 3D printing spray head end real-time tracking localization method, detailed process to mechanical arm as shown in Figure 1, input first
One model；Secondly the inclined direction of detection mechanical arm, and start and be suspended on four of surrounding and magazine be right against inclination side
To two cameras, form binocular vision system, printing head is tracked with track algorithm, obtain tracking ROI sense
Interest region, and classify to the image in the region with kmeans algorithm, and obtain the image of the class where spray head, pass through
The outer profile and abscissa of printing end are obtained with canny algorithm, hough line detection method are used later, to the edge
Image carries out straight-line detection, and judges the straight line on spray head both sides, by mathematical computations, calculates straight line and edge image abscissa is handed over
The midpoint of remittance, as tracing positional.Specifically includes the following steps:
(1) before opening camera, it is firstly received the ideal position of mechanical arm tail end, using the position as the effective of image
Position is set as initial ROI region, and sends in the feature vector of track algorithm, and zooms in and out to original video frame, together
When initial ROI region is also zoomed in and out.
(2) track algorithm is correlation filtering target tracking algorism, the process of this method are as follows: next frame, this method are first
First to the extracted region Hog feature of multiple surroundings of selected ROI region, then carried out solving what next frame was selected with circular matrix
ROI region, when obtaining a new selected ROI region, we are by the scaling of leading portion, the area zoom to original ratio
Example, and tracking box is shown in original image.
Motion target tracking is exactly to establish the positional relationship for the object of being tracked in continuous video sequence, obtain object
The complete motion profile of body.Currently, there are mainly two types of thinkings for more classical target tracking algorism: 1) knowing independent of priori
Know, moving target is measured directly from image sequence, and carry out target identification, finally determine interested moving target；
2) priori knowledge for depending on target, is first modeling target, then finds matched movement in real time in image sequence
Target.In current research, more outstanding several mainstream algorithms: Meanshift, Particle Filter and Kalman
Filter, the optical flow algorithm based on characteristic point and the method based on deep learning and correlation filtering.
(3) when one new selected ROI region of acquisition, then with kmeans method progress classification processing.
(4) it is found through experiments that selected ROI region is divided into 3 classes and can preferably distinguish spray head end, as shown in Fig. 2, beating
Disk and printed matter are printed, and the classification of printing head is exactly the second class of kmeans function category processing, i.e., only extracts the
Two classes, remaining classification setting are white；Due to classification processing, second sorted image of acquisition, which can be masked preferably, makes an uproar
Acoustic jamming, then the method for directlying adopt canny detection can effectively obtain the edge image of printing head end, as shown in Figure 3.
(5) for edge image obtained above, data point of the point set of edge image as hough straight-line detection, into
Row hough straight-line detection.
(6) since edge image is rough, the item number of the straight line of detection is more, detection of the houghlineP function to straight line
In, when accumulative threshold value is set as 30, when the threshold value between two straight lines is set as 10, energy is accurate to extract the straight of spray head both sides
Line；By two straight lines of fitting, the abscissa to cross is calculated according to the minimum point abscissa of edge image respectively, abscissa
Midpoint is that we want the position of tracking and positioning.
Claims (7)
1. a kind of 3D printing spray head end real-time tracking localization method, which is characterized in that the localization method includes: first to machinery
Arm inputs a model；Secondly the inclined direction of detection mechanical arm, and start and be suspended on four of surrounding and magazine be right against
Two cameras of inclined direction form binocular vision system, track with track algorithm to printing head, obtain tracking
ROI area-of-interest, and classify to the image in the region with kmeans algorithm, and obtain the figure of the class where spray head
Picture uses hough line detection method by obtaining the outer profile and abscissa of printing end with canny algorithm later, right
The edge image carries out straight-line detection, and judges the straight line on spray head both sides, by mathematical computations, calculates straight line and edge image is horizontal
The midpoint that coordinate crosses, as tracing positional.
2. localization method as described in claim 1, which is characterized in that before opening camera, be firstly received mechanical arm end
The ideal position at end is set as initial ROI region using the position as the active position of image, and sends track algorithm to
In feature vector, and original video frame is zoomed in and out, while initial ROI region is also zoomed in and out.
3. localization method as described in claim 1, which is characterized in that the track algorithm is correlation filtering target following calculation
Method, the process of this method are as follows: next frame, this method are special to the extracted region Hog of multiple surroundings of selected ROI region first
Sign, then carried out solving the selected ROI region of next frame with circular matrix, when obtaining a new selected ROI region, by preceding
The scaling of section the area zoom to former ratio, and shows tracking box in original image.
4. localization method as claimed in claim 3, which is characterized in that when one new selected ROI region of acquisition, then use
Kmeans method carries out classification processing.
5. localization method as claimed in claim 4, which is characterized in that be found through experiments that selected ROI region is divided into 3 class energy
Preferable difference spray head end, prints disk and printed matter, and the classification of printing head is exactly at kmeans function category
Second class of reason, i.e., only extract the second class, remaining classification setting is white；Due to classification processing, after obtaining second classification
Image can preferably mask noise jamming, then directly adopt canny detection method can effectively obtain printing head end
The edge image at end.
6. localization method as claimed in claim 5, which is characterized in that for edge image obtained above, edge image
Data point of the point set as hough straight-line detection carries out hough straight-line detection.
7. localization method as claimed in claim 5, which is characterized in that since edge image is rough, the item of the straight line of detection
Number is more, and houghlineP function is in the detection of straight line, and when accumulative threshold value is set as 30, the threshold value between two straight lines is arranged
When being 10, the straight line on spray head both sides can accurately be extracted；By two straight lines of fitting, respectively according to the minimum point of edge image
Abscissa calculates the abscissa to cross, and the midpoint of abscissa is that we want the position of tracking and positioning.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810748743.5A CN109177175A (en) | 2018-07-10 | 2018-07-10 | A kind of 3D printing spray head end real-time tracking localization method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810748743.5A CN109177175A (en) | 2018-07-10 | 2018-07-10 | A kind of 3D printing spray head end real-time tracking localization method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109177175A true CN109177175A (en) | 2019-01-11 |
Family
ID=64936345
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810748743.5A Pending CN109177175A (en) | 2018-07-10 | 2018-07-10 | A kind of 3D printing spray head end real-time tracking localization method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109177175A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109080146A (en) * | 2018-07-28 | 2018-12-25 | 中国科学院福建物质结构研究所 | A kind of 3D printing spray head end profile real time extracting method based on classification |
| CN110977960A (en) * | 2019-11-01 | 2020-04-10 | 北京工业大学 | Visual following system of three drive arms |
| CN113642406A (en) * | 2021-07-14 | 2021-11-12 | 广州市玄武无线科技股份有限公司 | System, method, device, equipment and storage medium for counting densely hung paper sheets |
| WO2021228181A1 (en) * | 2020-05-13 | 2021-11-18 | 中国科学院福建物质结构研究所 | 3d printing method and device |
| WO2021226891A1 (en) * | 2020-05-13 | 2021-11-18 | 中国科学院福建物质结构研究所 | 3d printing device and method based on multi-axis linkage control and machine visual feedback measurement |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100319100A1 (en) * | 2008-01-28 | 2010-12-23 | Jian Chen | Simple techniques for three-dimensional modeling |
| CN107452022A (en) * | 2017-07-20 | 2017-12-08 | 西安电子科技大学 | A kind of video target tracking method |
| CN108161930A (en) * | 2016-12-07 | 2018-06-15 | 广州映博智能科技有限公司 | A kind of robot positioning system of view-based access control model and method |
-
2018
- 2018-07-10 CN CN201810748743.5A patent/CN109177175A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100319100A1 (en) * | 2008-01-28 | 2010-12-23 | Jian Chen | Simple techniques for three-dimensional modeling |
| CN108161930A (en) * | 2016-12-07 | 2018-06-15 | 广州映博智能科技有限公司 | A kind of robot positioning system of view-based access control model and method |
| CN107452022A (en) * | 2017-07-20 | 2017-12-08 | 西安电子科技大学 | A kind of video target tracking method |
Non-Patent Citations (1)
| Title |
|---|
| 陈国清等: "《选择性激光熔化3D打印技术》", 30 September 2016, 西安电子科技大学出版社 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109080146A (en) * | 2018-07-28 | 2018-12-25 | 中国科学院福建物质结构研究所 | A kind of 3D printing spray head end profile real time extracting method based on classification |
| CN110977960A (en) * | 2019-11-01 | 2020-04-10 | 北京工业大学 | Visual following system of three drive arms |
| WO2021228181A1 (en) * | 2020-05-13 | 2021-11-18 | 中国科学院福建物质结构研究所 | 3d printing method and device |
| WO2021226891A1 (en) * | 2020-05-13 | 2021-11-18 | 中国科学院福建物质结构研究所 | 3d printing device and method based on multi-axis linkage control and machine visual feedback measurement |
| CN113674299A (en) * | 2020-05-13 | 2021-11-19 | 中国科学院福建物质结构研究所 | 3D printing method and device |
| CN113642406A (en) * | 2021-07-14 | 2021-11-12 | 广州市玄武无线科技股份有限公司 | System, method, device, equipment and storage medium for counting densely hung paper sheets |
| CN113642406B (en) * | 2021-07-14 | 2023-01-31 | 广州市玄武无线科技股份有限公司 | System, method, device, equipment and storage medium for counting densely-suspended paper sheets |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109177175A (en) | A kind of 3D printing spray head end real-time tracking localization method | |
| CN109080144A (en) | 3D printing spray head end real-time tracking localization method based on central point judgement | |
| CN109080146A (en) | A kind of 3D printing spray head end profile real time extracting method based on classification | |
| US8249334B2 (en) | Modeling of humanoid forms from depth maps | |
| CN105957105B (en) | The multi-object tracking method and system of Behavior-based control study | |
| CN102880866A (en) | Method for extracting face features | |
| CN103886619B (en) | A kind of method for tracking target merging multiple dimensioned super-pixel | |
| Lee et al. | Body part detection for human pose estimation and tracking | |
| CN105809144A (en) | Gesture recognition system and method adopting action segmentation | |
| CN107315994B (en) | Clustering method based on Spectral Clustering space trajectory | |
| CN102479388A (en) | Expression interaction method based on face tracking and analysis | |
| CN103745218B (en) | Gesture identification method and device in depth image | |
| CN103500010A (en) | Method for locating fingertips of person through video | |
| CN111178170B (en) | Gesture recognition method and electronic equipment | |
| CN110321937A (en) | A kind of moving human hand tracking method of Faster-RCNN combination Kalman filtering | |
| CN104700088A (en) | Gesture track recognition method based on monocular vision motion shooting | |
| CN108596947A (en) | A kind of fast-moving target tracking method suitable for RGB-D cameras | |
| CN106599873A (en) | Figure identity identification method based on three-dimensional attitude information | |
| Zahedi et al. | Appearance-based recognition of words in american sign language | |
| CN103942814A (en) | Quick moving object track estimation method for asynchronous binocular vision system | |
| CN109130167A (en) | A kind of 3D printing spray head end tracking based on correlation filtering | |
| CN109016517A (en) | A kind of 3D printing spray head end profile real time extracting method | |
| CN105590100B (en) | Surpass the human motion recognition method of voxel based on identification | |
| CN107045725A (en) | A kind of human body motion capture method | |
| Ma et al. | WeldNet: A deep learning based method for weld seam type identification and initial point guidance |
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: 20190111 |