Nothing Special   »   [go: up one dir, main page]

CN106863355A - A kind of object identification method and robot for robot - Google Patents

A kind of object identification method and robot for robot Download PDF

Info

Publication number
CN106863355A
CN106863355A CN201611222770.6A CN201611222770A CN106863355A CN 106863355 A CN106863355 A CN 106863355A CN 201611222770 A CN201611222770 A CN 201611222770A CN 106863355 A CN106863355 A CN 106863355A
Authority
CN
China
Prior art keywords
sample
sample object
images
image
robot
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
Application number
CN201611222770.6A
Other languages
Chinese (zh)
Inventor
陆羽皓
赵宁宁
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Guangnian Wuxian Technology Co Ltd
Original Assignee
Beijing Guangnian Wuxian Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Beijing Guangnian Wuxian Technology Co Ltd filed Critical Beijing Guangnian Wuxian Technology Co Ltd
Priority to CN201611222770.6A priority Critical patent/CN106863355A/en
Publication of CN106863355A publication Critical patent/CN106863355A/en
Pending legal-status Critical Current

Links

Classifications

    • BPERFORMING OPERATIONS; TRANSPORTING
    • B25HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
    • B25JMANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
    • B25J19/00Accessories fitted to manipulators, e.g. for monitoring, for viewing; Safety devices combined with or specially adapted for use in connection with manipulators
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50Information retrieval; Database structures therefor; File system structures therefor of still image data
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Artificial Intelligence (AREA)
  • Robotics (AREA)
  • Mechanical Engineering (AREA)
  • Databases & Information Systems (AREA)
  • Manipulator (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses a kind of object identification method for robot and a kind of robot.Methods described includes:IMAQ is carried out to same sample object from multiple different visual angles respectively to obtain multiple different sample images;The different sample images of same sample object will be subordinated to and associate preservation with the identification label of the sample object;Images to be recognized is gathered for object to be identified, the images to be recognized is view data of the object to be identified under current visual angle;The sample image of the images to be recognized matching is searched out from all sample images for having preserved;The identification label associated by sample image that extraction is searched out is to complete the identification to the object to be identified.Compared with prior art, the method for the present invention reduces influence of the visual angle change to identification process during image recognition object, greatly expands the recognizable set and recognition efficiency of robot, improves recognition correct rate.

Description

A kind of object identification method and robot for robot
Technical field
The present invention relates to robot field, and in particular to a kind of object identification method and robot for robot.
Background technology
With continuing to develop for robot technology, intelligent robot is more and more employed the family life with the mankind In.
Currently, most intelligent robot possesses vision collecting function, view-based access control model acquisition function, and it is right that robot can be realized The image recognition of current object to be identified.Common image recognition processes are to pre-save sample image data, are carrying out image The view data of object to be identified and all sample image datas for having preserved are done into matching during identification and searches for corresponding so as to obtain Sample image data, by the use of the corresponding iamge description of sample image data as recognition result.
In above-mentioned image recognition processes, key point be sample image data with object image data to be identified Match somebody with somebody.But because the view data of sample image data and object to be identified is all planar image data, sample image data Only the image of certain angle of object is recorded, when IMAQ is carried out to object from another angle, even same object, The view data of acquisition is also different from the final sample image data for obtaining.
Therefore, even if robot was beforehand with the collection of sample image data to current object, in the interference of visual angle change Under, robot can not smoothly recognize the object after.
The content of the invention
The invention provides a kind of object identification method for robot, methods described includes:
IMAQ is carried out to same sample object from multiple different visual angles respectively to obtain multiple different sample images;
The different sample images of same sample object will be subordinated to and associate preservation with the identification label of the sample object;
Images to be recognized is gathered for object to be identified, the images to be recognized is the object to be identified in current visual angle Under view data;
The sample image of the images to be recognized matching is searched out from all sample images for having preserved;
The identification label associated by sample image that extraction is searched out is to complete the identification to the object to be identified.
In one embodiment, carry out IMAQ to same sample object to obtain multiple not from multiple different visual angles respectively Same sample image, wherein:
Still image gathers position;
Rotate the sample object;
Multi-view image collection is carried out to the sample object during the sample object is rotated.
In one embodiment, carry out IMAQ to same sample object to obtain multiple not from multiple different visual angles respectively Same sample image, wherein:
Keep the sample object motionless;
Around sample object rotation;
Multi-view image collection is carried out to the sample object during being rotated around the sample object.
In one embodiment, carry out IMAQ to same sample object to obtain multiple not from multiple different visual angles respectively Same sample image, wherein:
Judge whether the sample object can be rotated;
When the sample object can be rotated, still image gathers position and rotates the sample object, in rotation institute Multi-view image collection is carried out to the sample object during stating sample object;
Rotated around the sample object when the sample object cannot be rotated, around sample object rotation Multi-view image collection is carried out to the sample object during turning.
In one embodiment, carry out IMAQ to same sample object to obtain multiple not from multiple different visual angles respectively Same sample image, wherein:
360 degree of view transformations are carried out to the sample object based on horizontal plane;
IMAQ is carried out to the sample object with default angle interval during view transformation.
The invention allows for a kind of intelligent robot, the robot includes:
Collection view transformation module, it is configured to convert visual angle of the robot to same sample object;
Contrast images acquisition module, it is configured to during robot is converted to the visual angle of the sample object IMAQ is carried out to the sample object from multiple different visual angles respectively to obtain multiple different sample images;
Contrast images memory module, its different sample image and the sample for being configured to be subordinated to same sample object The identification label association of object is preserved;
Image recognition acquisition module, it is configured to gather images to be recognized, the images to be recognized for object to be identified It is view data of the object to be identified under current visual angle;
Image recognition matching module, it is configured to search out the images to be recognized from all sample images for having preserved The sample image of matching;
Identification tag extraction module, the identification label that it is configured to extract associated by the sample image for searching out is right to complete The identification of the object to be identified.
In one embodiment:
The collection view transformation module includes sample object whirligig, and it is configured to still image collection position and revolves Turn the sample object;
The contrast images acquisition module is configured to enter the sample object during the sample object is rotated Row multi-view image collection.
In one embodiment:
The collection view transformation module includes gearshift, and it is configured to keep the sample object motionless and drives institute Robot is stated to be rotated around the sample object;
The contrast images acquisition module is configured to right during the robot rotates around the sample object The sample object carries out multi-view image collection.
In one embodiment, the collection view transformation module is included:
Sample object parsing module, it is configured to judge whether the sample object can be rotated;
Sample object whirligig, it is configured to the still image when the sample object can be rotated and gathers position simultaneously Rotate the sample object;
Gearshift, it is configured to drive the robot around the sample when the sample object cannot be rotated This object rotates;
The contrast images acquisition module is configured to rotating the sample object or the robot around the sample Object carries out multi-view image collection during rotating to the sample object.
In one embodiment:
The collection view transformation module is configured to horizontal plane carries out 360 degree of view transformations to the sample object;
The contrast images acquisition module is configured to during view transformation with default angle interval to the sample This object carries out IMAQ.
Compared with prior art, visual angle change is to identification process during the method for the present invention reduces image recognition object Influence, greatly expand the recognizable set and recognition efficiency of robot, improve recognition correct rate.
Further feature of the invention or advantage will be illustrated in the following description.Also, Partial Feature of the invention or Advantage will be become apparent by specification, or be appreciated that by implementing the present invention.The purpose of the present invention and part Advantage can be realized or obtained by specifically noted step in specification, claims and accompanying drawing.
Brief description of the drawings
Accompanying drawing is used for providing a further understanding of the present invention, and constitutes a part for specification, with reality of the invention Apply example to be provided commonly for explaining the present invention, be not construed as limiting the invention.In the accompanying drawings:
Fig. 1 is method flow diagram according to an embodiment of the invention;
Fig. 2~Fig. 4 is the partial process view of method according to embodiments of the present invention;
Fig. 5 is robot system architecture's sketch according to an embodiment of the invention;
Fig. 6~Fig. 8 is robot system part-structure sketch according to embodiments of the present invention.
Specific embodiment
Describe embodiments of the present invention in detail below with reference to drawings and Examples, implementation personnel of the invention whereby Can fully understand how application technology means solve technical problem for the present invention, and reach technique effect implementation process and according to The present invention is embodied according to above-mentioned implementation process.If it should be noted that do not constitute conflict, each embodiment in the present invention And each feature in each embodiment can be combined with each other, the technical scheme for being formed protection scope of the present invention it It is interior.
With continuing to develop for robot technology, intelligent robot is more and more applied to the family life of the mankind In.
Currently, most intelligent robot possesses vision collecting function, view-based access control model acquisition function, and it is right that robot can be realized The image recognition of current object to be identified.Common image recognition processes are to pre-save sample image data, are carrying out image The view data of object to be identified and all sample image datas for having preserved are done into matching during identification and searches for corresponding so as to obtain Sample image data, by the use of the corresponding iamge description of sample image data as recognition result.
In above-mentioned image recognition processes, key point be sample image data with object image data to be identified Match somebody with somebody.But because the view data of sample image data and object to be identified is all planar image data, sample image data Only the image of certain angle of object is recorded, when IMAQ is carried out to object from another angle, even same object, The view data of acquisition is also different from the final sample image data for obtaining.
Therefore, even if robot was beforehand with the collection of sample image data to current object, in the interference of visual angle change Under, robot can not smoothly recognize the object after.For this problem, the present invention proposes a kind of for robot Object identification method.
In an embodiment of the present invention, the method for contrast sample's figure is employed to realize the identification to object.Specifically, When identification object is needed, the images to be recognized of object is gathered, then by the images to be recognized and the sample for having preserved This image collection does search matching, the sample image matched with images to be recognized in sample image set is obtained, so as to obtain sample Object identification tag (such as object names) corresponding to this image, and then realize the identification to object.
In above-mentioned steps, difference with the prior art is, in the sample image set for having preserved, as certain This object, sample image is not the image sampling for some visual angle of sample object, and is directed to multiple different visual angles Multiple images sampling set.
By taking a specific applied environment as an example.Assuming that the front view A of object A1, rear view A2, left side view A3, it is right Side view A4And top view A5Differ (there is larger difference).
If in sample image set, only preserving the front view A of object A1As sample image.So in identification In scene, when only robot faces object A fronts, its images to be recognized for getting is possible to matching and searches front to regard Figure A1, so as to obtain correspondence front view A1Object A identification label with complete identification.When robot in face of the object A back sides, During side or above, its images to be recognized for getting cannot also be matched even with more wide in range Similarity matching and searched Rope is to front view A1, cannot also realize the identification to object A.
In an embodiment of the present invention, for object A, the front view A of object A is preserved in sample set1, the back side regards Figure A2, left side view A3, right side view A4And top view A5As sample image.So, no matter robot is in face of object A's Front, the back side, side or top, the images to be recognized that it gets can search associate A by matching search Identification label sample image (front view A1, rear view A2, left side view A3, right side view A4And top view A5), So as to obtain the identification label of object A to complete object identification.
The method according to the invention, reduces influence of the visual angle change to identification process during image recognition object, greatly The big recognizable set and recognition efficiency for extending robot, improves recognition correct rate.
Next the detailed process of method according to embodiments of the present invention is described in detail based on accompanying drawing, in the flow chart of accompanying drawing The step of showing can perform in comprising the such as one group computer system of computer executable instructions.Although in flow charts The logical order of each step is shown, but in some cases, can be performing shown different from order herein or retouch The step of stating.
In one embodiment, gather first future object identification during required identification object sample image. As shown in figure 1, various visual angles capturing sample image (step S100) first, specifically, in the step s 100, respectively from multiple different Visual angle same sample object (sample of the object of required identification during following object identification) is carried out IMAQ with Obtain multiple different sample images.
Next the different sample images of same sample object will be subordinated to associates guarantor with the identification label of the sample object Deposit (step S110), thus got the sample image set for recognizing jobbie.For different sample objects It is repeated continuously step S100 and step S110, it is possible to obtain the sample image set for recognizing multiple different objects.
After sample image set is ready, it is possible to based on sample image set to included in sample image set The corresponding object of sample image is identified.Specifically, when being identified, images to be recognized (step S120) is gathered first, Specifically, images to be recognized is view data of the object to be identified under current visual angle.Then (protected from sample image set In all sample images deposited) search out the sample image (step S130) that images to be recognized is matched.
There are two kinds of results in step S130, one is the presence of the sample graph matched with images to be recognized in sample image set Picture;Two is in the absence of the sample image matched with images to be recognized in sample image set.Therefore, perform after step s 130 Step S140, with the presence or absence of the sample image matched with images to be recognized (to matching Search Results in judgement sample image collection Judged).
When there is sample image match with images to be recognized in sample image set, extraction with match the sample that searches out Identification label associated by this image is completing to the identification of object to be identified (using the identification label as recognition result) (step S150)。
When in sample image set in the absence of the sample image matched with images to be recognized, object to be identified is labeled as None- identified (step S160).
Further, after step S150 and S160 according to recognition result (specific identification label or None- identified Mark) perform next step operation.
In the step shown in Fig. 1, one of key point is the sample for collecting enough view numbers in the step s 100 Image.Because sample image is the depth of field relation and sample for recognizing a specific object, object and surrounding enviroment Image capture position can't produce influence substantially with the spacing of sample object to the object identification based on sample image;Cause This, in one embodiment, with the focus of single sample object, position and the direction of image collecting device is not changed, by changing The direction (changing the face of the image collecting device that sample object is faced) for becoming sample object is come respectively from multiple different visual angles pair Same sample object carries out IMAQ.
Specifically, as shown in Fig. 2 in one embodiment, still image gathers position (stationary machines people's IMAQ first The position of device and direction) (step S200);Then rotated sample object (is kept between sample object and image capture position Away from continuous to change face of the sample object towards image collecting device) (step S210);Carried out in the implementation procedure of step S210 IMAQ is so as to complete the multi-view image collection (step S220) to sample object.
Further, it is contemplated that under some application scenarios, it is impossible to (such as sample object mistake is rotated to sample object Greatly).Therefore, in one embodiment, employ with the focus of single sample object, position and the direction of sample object do not changed, IMAQ is carried out to same sample object from multiple different visual angles by the position of image collecting device respectively.
Specifically, as shown in figure 3, in one embodiment, fixed sample object space (keep sample object motionless) first (step S300);Then robot (makes image collecting device remain in face of sample object around sample object movement In the case of around sample object rotation) (step S310);IMAQ is carried out in the implementation procedure of step S310 so as to complete To the multi-view image collection (step S320) of sample object.
Further, in one embodiment, the method with reference to shown in Fig. 2 and Fig. 3 carries out multi-view image collection.As schemed Shown in 4, in one embodiment, first determine whether whether sample object can be rotated (step S400);When sample object can be by Still image collection position (step S411) and rotated sample object (step S412) during rotation, in the process of rotated sample object In multi-view image collection (step S413) is carried out to sample object;The fixed sample object when sample object cannot be rotated Position (not carrying out displacement operation to sample object) (step S421) rotates (step S422) around sample object, around described Sample object carries out multi-view image collection (step S423) during rotating to sample object.
Further, it is contemplated that in overwhelming majority identification application scenarios, robot is in same with object to be identified Individual planar rings are domestic (situation of object identification is carried out when seldom occurring that robot is located at object surface or underface).Cause This, in order to reduce the data acquisition amount of sample image, improves the matching search speed in identification process, in one embodiment, right Same sample object only carries out 360 degree of multi-view image collection in the horizontal plane.Specifically, carrying out many of sample image When multi-view image is gathered, 360 degree of view transformations are carried out to sample object based on horizontal plane;With default during view transformation Angle interval IMAQ is carried out to sample object.
By taking step shown in Fig. 2 as an example, in one embodiment, the image capture position of stationary machines people in step s 200, Image collecting device and sample object is made to be in a horizontal plane, the camera of image collecting device is towards sample object. Sample object does 360 degree of rotations (rotation) around the central shaft perpendicular to horizontal plane in step S210.In step S220, whenever Sample object rotates through a sample image of an angle (such as 10 degree of the rotation) image collecting device with regard to collecting sample object.
By taking step shown in Fig. 3 as an example, in one embodiment, the position of fixed sample object is not (to sample in step S300 This object carries out the operation of any its position of interference), make image collecting device be in a horizontal plane with sample object, image The camera of harvester is towards sample object.Holding image collecting device is same with sample object in step S210 is in a water Plane is interior and maintains the spacing between sample object and image collecting device constant, makes image collecting device around perpendicular to horizontal plane Central shaft do 360 degree of rotations (rotation).In step S220, whenever sample object rotates through (such as rotation 10 of an angle Degree) image collecting device with regard to collecting sample object a sample image.
Based on the method for the present invention, the invention allows for a kind of intelligent robot.As shown in figure 5, in one embodiment, Robot includes:
Collection view transformation module 500, it is configured to convert visual angle of the robot to same sample object;
Contrast images acquisition module 510, it is configured to during robot is converted to the visual angle of sample object IMAQ is carried out to sample object from multiple different visual angles respectively to obtain multiple different sample images;
Contrast images memory module 530, its different sample image and the sample for being configured to be subordinated to same sample object The identification label association of this object is preserved;
Image recognition acquisition module 520, it is configured to gather images to be recognized for object to be identified, wherein, it is to be identified Image is view data of the object to be identified under current visual angle;
Image recognition matching module 540, it is configured to be searched in the sample image preserved from contrast images memory module 530 Rope goes out the sample image matched with images to be recognized;
Identification tag extraction module 550, it is configured to obtain the sample image pair that image recognition matching module 540 is searched out The identification label answered is to complete the identification to object to be identified.
Further, in one embodiment, contrast images acquisition module 510 is based on same with image recognition acquisition module 520 A set of image collecting device construction.
Further, in one embodiment, collection view transformation module 500 carries out 360 based on horizontal plane to sample object Degree view transformation;
Contrast images acquisition module 510 is configured to during view transformation with default angle interval to the sample Object carries out IMAQ.
Further, as shown in fig. 6, in one embodiment, collection view transformation module 600 includes sample object rotating dress 611 are put, sample object whirligig 611 is configured to still image collection position and rotated sample object;Contrast images gather mould Block 610 is configured to carry out multi-view image to sample object during the rotated sample object of sample object whirligig 611 Gather and collection result is saved into contrast images memory module 630.Further, in one embodiment, using robot Manipulator construction sample object whirligig 611, with manipulator rotated sample object.
Further, as shown in fig. 7, in one embodiment, collection view transformation module 700 includes gearshift 711.Position Moving device 711 is configured to keep sample object motionless and driven machine people rotates around sample object;Contrast images acquisition module 710 are configured to carry out sample object during robot rotates around sample object multi-view image collection and will gather Result is saved into contrast images memory module 630.Further, in one embodiment, using the displacement mechanism of robot itself To construct gearshift 711.
Further, as shown in figure 8, in one embodiment, collection view transformation module 800 parses mould comprising sample object Block 813, sample object whirligig 812 and gearshift 811.Wherein:
Sample object parsing module 813 is configured to whether judgement sample object can be rotated;
Sample object whirligig 812 is configured to the still image when sample object can be rotated and gathers position and rotate Sample object;
Gearshift 811 is configured to the driven machine people when sample object cannot be rotated and is rotated around sample object.
Contrast images acquisition module 810 be configured to during rotated sample object or in robot around sample contents Body carries out multi-view image collection to sample object and collection result is saved into contrast images memory module during rotating 830。
While it is disclosed that implementation method as above, but described content is only to facilitate understanding the present invention and adopting Implementation method, is not limited to the present invention.Method of the present invention can also have other various embodiments.Without departing substantially from In the case of essence of the present invention, those of ordinary skill in the art work as can make various corresponding changes or change according to the present invention Shape, but these corresponding changes or deformation should all belong to scope of the claims of the invention.

Claims (10)

1. a kind of object identification method for robot, it is characterised in that methods described includes:
IMAQ is carried out to same sample object from multiple different visual angles respectively to obtain multiple different sample images;
The different sample images of same sample object will be subordinated to and associate preservation with the identification label of the sample object;
Images to be recognized is gathered for object to be identified, the images to be recognized is the object to be identified under current visual angle View data;
The sample image of the images to be recognized matching is searched out from all sample images for having preserved;
The identification label associated by sample image that extraction is searched out is to complete the identification to the object to be identified.
2. method according to claim 1, it is characterised in that carried out to same sample object from multiple different visual angles respectively The IMAQ sample images different to obtain multiple, wherein:
Still image gathers position;
Rotate the sample object;
Multi-view image collection is carried out to the sample object during the sample object is rotated.
3. method according to claim 1, it is characterised in that carried out to same sample object from multiple different visual angles respectively The IMAQ sample images different to obtain multiple, wherein:
Keep the sample object motionless;
Around sample object rotation;
Multi-view image collection is carried out to the sample object during being rotated around the sample object.
4. method according to claim 1, it is characterised in that carried out to same sample object from multiple different visual angles respectively The IMAQ sample images different to obtain multiple, wherein:
Judge whether the sample object can be rotated;
When the sample object can be rotated, still image gathers position and rotates the sample object, is rotating the sample Multi-view image collection is carried out to the sample object during this object;
Rotated around the sample object when the sample object cannot be rotated, what is rotated around the sample object During multi-view image collection is carried out to the sample object.
5. the method according to any one of claim 1-4, it is characterised in that respectively from multiple different visual angles to This object carries out IMAQ to obtain multiple different sample images, wherein:
360 degree of view transformations are carried out to the sample object based on horizontal plane;
IMAQ is carried out to the sample object with default angle interval during view transformation.
6. a kind of intelligent robot, it is characterised in that the robot includes:
Collection view transformation module, it is configured to convert visual angle of the robot to same sample object;
Contrast images acquisition module, it is configured to during robot is converted to the visual angle of the sample object respectively IMAQ is carried out to the sample object from multiple different visual angles to obtain multiple different sample images;
Contrast images memory module, its different sample image and the sample object for being configured to be subordinated to same sample object Identification label association preserve;
Image recognition acquisition module, it is configured to gather images to be recognized for object to be identified, and the images to be recognized is institute State view data of the object to be identified under current visual angle;
Image recognition matching module, it is configured to search out the images to be recognized matching from all sample images for having preserved Sample image;
Identification tag extraction module, it is configured to extract the identification label associated by the sample image for searching out to complete to described The identification of object to be identified.
7. robot according to claim 6, it is characterised in that:
The collection view transformation module includes sample object whirligig, and it is configured to still image collection position and rotates institute State sample object;
The contrast images acquisition module is configured to carry out the sample object during the sample object is rotated many Multi-view image is gathered.
8. robot according to claim 6, it is characterised in that:
The collection view transformation module includes gearshift, and it is configured to keep the sample object motionless and drives the machine Device people rotates around the sample object;
The contrast images acquisition module is configured to during the robot rotates around the sample object to described Sample object carries out multi-view image collection.
9. robot according to claim 6, it is characterised in that the collection view transformation module is included:
Sample object parsing module, it is configured to judge whether the sample object can be rotated;
Sample object whirligig, it is configured to the still image when the sample object can be rotated and gathers position and rotate The sample object;
Gearshift, it is configured to drive the robot around the sample contents when the sample object cannot be rotated Body rotates;
The contrast images acquisition module is configured to rotating the sample object or the robot around the sample object Multi-view image collection is carried out to the sample object during rotation.
10. the method according to any one of claim 6-9, it is characterised in that:
The collection view transformation module is configured to horizontal plane carries out 360 degree of view transformations to the sample object;
The contrast images acquisition module is configured to during view transformation with default angle interval to the sample contents Body carries out IMAQ.
CN201611222770.6A 2016-12-27 2016-12-27 A kind of object identification method and robot for robot Pending CN106863355A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201611222770.6A CN106863355A (en) 2016-12-27 2016-12-27 A kind of object identification method and robot for robot

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201611222770.6A CN106863355A (en) 2016-12-27 2016-12-27 A kind of object identification method and robot for robot

Publications (1)

Publication Number Publication Date
CN106863355A true CN106863355A (en) 2017-06-20

Family

ID=59165043

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201611222770.6A Pending CN106863355A (en) 2016-12-27 2016-12-27 A kind of object identification method and robot for robot

Country Status (1)

Country Link
CN (1) CN106863355A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108555909A (en) * 2018-04-17 2018-09-21 子歌教育机器人(深圳)有限公司 A kind of target seeking method, AI robots and computer readable storage medium
CN109558868A (en) * 2017-09-27 2019-04-02 缤果可为(北京)科技有限公司 Image automatic collection and tagging equipment and method
CN114769021A (en) * 2022-04-24 2022-07-22 广东天太机器人有限公司 Robot spraying system and method based on full-angle template recognition

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE102006005990A1 (en) * 2006-02-08 2007-08-09 Vision Tools Hard- Und Software Entwicklungs-Gmbh Workpiece characteristics position measuring method involves measuring position by camera, which is movably fixed to robot, for multiple picture recordings
CN101859371A (en) * 2009-04-10 2010-10-13 鸿富锦精密工业(深圳)有限公司 Pick-up device and object identification method thereof
CN102254169A (en) * 2011-08-23 2011-11-23 东北大学秦皇岛分校 Multi-camera-based face recognition method and multi-camera-based face recognition system
US20130136300A1 (en) * 2011-11-29 2013-05-30 Qualcomm Incorporated Tracking Three-Dimensional Objects
CN103473529A (en) * 2013-08-26 2013-12-25 昆明学院 Method and device for recognizing faces through multi-angle imaging
CN103984942A (en) * 2014-05-28 2014-08-13 深圳市中兴移动通信有限公司 Object recognition method and mobile terminal

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE102006005990A1 (en) * 2006-02-08 2007-08-09 Vision Tools Hard- Und Software Entwicklungs-Gmbh Workpiece characteristics position measuring method involves measuring position by camera, which is movably fixed to robot, for multiple picture recordings
CN101859371A (en) * 2009-04-10 2010-10-13 鸿富锦精密工业(深圳)有限公司 Pick-up device and object identification method thereof
CN102254169A (en) * 2011-08-23 2011-11-23 东北大学秦皇岛分校 Multi-camera-based face recognition method and multi-camera-based face recognition system
US20130136300A1 (en) * 2011-11-29 2013-05-30 Qualcomm Incorporated Tracking Three-Dimensional Objects
CN103473529A (en) * 2013-08-26 2013-12-25 昆明学院 Method and device for recognizing faces through multi-angle imaging
CN103984942A (en) * 2014-05-28 2014-08-13 深圳市中兴移动通信有限公司 Object recognition method and mobile terminal

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109558868A (en) * 2017-09-27 2019-04-02 缤果可为(北京)科技有限公司 Image automatic collection and tagging equipment and method
CN108555909A (en) * 2018-04-17 2018-09-21 子歌教育机器人(深圳)有限公司 A kind of target seeking method, AI robots and computer readable storage medium
CN114769021A (en) * 2022-04-24 2022-07-22 广东天太机器人有限公司 Robot spraying system and method based on full-angle template recognition

Similar Documents

Publication Publication Date Title
CN104793620B (en) The avoidance robot of view-based access control model feature binding and intensified learning theory
US8768071B2 (en) Object category recognition methods and robots utilizing the same
CN102999918B (en) Multi-target object tracking system of panorama video sequence image
WO2016009811A1 (en) Method for calibrating one or more cameras
US20190050668A1 (en) Method and System for Concurrent Reconstruction of Dynamic and Static Objects
Zhang et al. Robust visual odometry in underwater environment
CN106863355A (en) A kind of object identification method and robot for robot
JPH08327296A (en) Method and equipment for detecting target motion and applying method of said method
EP3035242B1 (en) Method and electronic device for object tracking in a light-field capture
EP3107007B1 (en) Method and apparatus for data retrieval in a lightfield database
Melekhov et al. Digging into self-supervised learning of feature descriptors
Gu et al. OSSID: online self-supervised instance detection by (and for) pose estimation
CN110097504A (en) A kind of image vision acquisition system for tunnel crusing robot
CN113505629A (en) Intelligent storage article recognition device based on light weight network
CN110889460A (en) Mechanical arm specified object grabbing method based on cooperative attention mechanism
CN115018886B (en) Motion trajectory identification method, device, equipment and medium
US8538142B2 (en) Face-detection processing methods, image processing devices, and articles of manufacture
CN115546021A (en) Multi-camera image splicing method applied to cold bed shunting scene detection
CN109359649A (en) A kind of recognition methods of access object, storage medium and the article-storage device of article-storage device
Li et al. A novel automatic image stitching algorithm for ceramic microscopic images
Sadeghi-Tehran et al. ATDT: autonomous template-based detection and tracking of objects from airborne camera
Kokate et al. Survey: image mosaicing based on feature extraction
CN107273850B (en) Autonomous following method based on mobile robot
Liu et al. Multi-objects tracking and online identification based on SIFT
Singh et al. Instance-specific 6-dof object pose estimation from minimal annotations

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: 20170620