CN103186896B - For the space invariance amount secondary detection method and apparatus of space map structuring - Google Patents
For the space invariance amount secondary detection method and apparatus of space map structuring Download PDFInfo
- Publication number
- CN103186896B CN103186896B CN201110444217.8A CN201110444217A CN103186896B CN 103186896 B CN103186896 B CN 103186896B CN 201110444217 A CN201110444217 A CN 201110444217A CN 103186896 B CN103186896 B CN 103186896B
- Authority
- CN
- China
- Prior art keywords
- amount
- space
- space invariance
- invariance
- dij
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
Disclose the space invariance amount secondary detection method and apparatus for space map structuring.The described method includes：In one group of observed image obtained from successively to unknown space observe twice with the first pose and the second pose by using an observation main body, the n space invariance amount that Preliminary detection arrives is obtained；The distance of additional space invariant in observation twice is calculated respectively, with the distance of j-th of space invariance amount is dij in being observed in first time wherein for i-th of space invariance amount, with the distance of j-th of space invariance amount it is dij ' in being observed second, wherein i, j, which are respectively less than, is equal to n, and j ≠ i；Based on predetermined temporal voting strategy, vote certain amount of space invariance amount；And correct N number of space invariance amount is judged among n space invariance amount according to voting results and predetermined decision plan, so as to filter out incorrect matched space invariance amount, wherein N is natural number and is less than or equal to n.
Description
Technical field
The present invention relates to fast for the space invariance amount secondary detection method and apparatus and robot of space map structuring
Fast location estimation method and equipment.More specifically to can reduce calculation amount and improve the space invariance amount of the speed of service
Secondary detection method and apparatus and the robot rapid location estimation using the space invariance amount secondary detection method and apparatus
Method and apparatus.
Background technology
Three-dimensional localization and map structuring are that mobile robot carries out movement navigation, path planning, object in circumstances not known
One of critical support technology of identification and semantic understanding.Robot is during continuous moving, by matching in different observation bits
The three-dimensional observation data put can calculate the position of robot.Matching observation data and the core of calculating location information are accurate
Really find the space invariance amount in observed result twice, and then point calculates observation point twice on the basis of space invariance amount
Location variation completes the location estimation of robot.
Figure 1A, 1B are shown based on the schematic diagram for observing data progress position and attitude variation estimation twice.In figure 1A, show
Observation data of the Chu Liao robots under (0,0,0,0,0,0) position and attitude.In fig. ib, show robot be moved to (x,
Y, z, roll, pitch, yaw) observation data under position and attitude.Wherein, 6 boxes expression in Figure 1A, 1B is observed from twice
The 6 groups of space invariance amounts (6 groups of three-dimensional points to) detected in data.Explanation is needed exist for, so-called space invariance amount refers to
Identical spatial point in the observation twice carried out in observation main body (robot i.e., here) with different positions and pose.Due to robot
Movement, twice observe in, the position of with respect to 6 space invariance points of robot changes, based on change can calculate (x,
Y, z, roll, pitch, yaw) concrete numerical value, so as to complete the position and attitude of robot estimation.Although it is pointed out that
It is described in Figure 1A, 1B using 6 groups of space invariance amounts as example, but in theory, it is only necessary to by 3 groups in space not
Conllinear space invariance point, so that it may acquire the concrete numerical value of (x, y, z, roll, pitch, yaw).
It can be seen that it can detect that the space invariance amount observed twice in data (alternatively referred to as characteristic point) is exactly
The key of problem.There are many achievements in research can be more efficiently from 2-D gray image or three-dimensional point cloud observation number at present
Space invariance amount (for example, by the matched mode of gray value) is detected in, but there are still large errors, it may occur that mistake is examined
The situation of survey.
Fig. 2A, 2B show that incorrect space invariance amount triggers the schematic diagram of position and attitude erroneous estimation.In Fig. 2A, 2B
In, due to the 4th group and the incorrect matching of the 6th group of space invariance amount, the erroneous estimation of robot location's posture will be caused.Cause
This, existing solution generally post-processes the space invariance amount that Preliminary detection goes out, and carries out the two of space invariance amount
Secondary detection, filters out wrong data.
Existing space invariance amount secondary detection uses RANSAC (RANdom SAmpling Consensus) method.It should
Method is by multiple random experiments, to select space invariance amount.In each experiment, wherein three groups of space invariance amounts are randomly selected
(A1, A2, A3), this three groups of invariants calculate position and attitude information, and using the position and attitude information as standard, judge others
Whether space invariance amount is correct, it will finds all accurate space invariance amounts that space invariance property is kept with (A1, A2, A3)
M1=(A1, A2, A3, A4 ... An), and calculate the error of this group of space invariance amount M1 so obtained.By repeatedly random real
It tests, multigroup space invariance amount (M1, M2, M3 ..., Mn) selected will be obtained.Choose wherein error minimum one group is used as most
The whole space invariance amount for filtering out wrong data.
This method needs the position and attitude variable quantity of multiple calculating robot in each random experiments, computationally intensive, fortune
Scanning frequency degree is slow.In addition, when experiment number is less, it is impossible to find optimal space invariance amount.
The content of the invention
In view of situation above, the present invention proposes the new space invariance amount two based on space invariance amount election inspection policies
What the robot rapid location of secondary detection method and equipment and use the space invariance amount secondary detection method and apparatus was estimated
The position and attitude variable quantity of multiple calculating robot is not required in method and apparatus, and calculation amount is small, and the speed of service is fast, by surveying
Examination can improve the 1-2 order of magnitude compared with RANSAC methods.
It is secondary to provide a kind of space invariance amount for space map structuring to one side according to embodiments of the present invention
Detection method, the described method includes：Main body is being observed successively with the first pose and the second pose to unknown space by using one
It carries out in one group of observed image obtained from observing twice, the n space invariance amount that Preliminary detection arrives is obtained, wherein described
Observation main body at least possesses an image acquisition units to obtain observed image, and n is natural number；Phase in observation twice is calculated respectively
Answer the distance of space invariance amount, wherein for i-th of space invariance amount, in being observed in first time with j-th space invariance amount
Distance with the distance of j-th of space invariance amount is dij ' in being observed at second, wherein i, j, which are respectively less than, is equal to n, and j for dij
≠i；It based on predetermined temporal voting strategy, votes certain amount of space invariance amount, wherein for some space invariance amount
Ballot is carried out to refer to be voted for some described space invariance amount by every other space invariance amount；And according to ballot
As a result and predetermined decision plan judges correct N number of space invariance amount among n space invariance amount, so as to filter out not just
True matched space invariance amount, wherein N are natural number and are less than or equal to n.
Preferably, according to the method for the embodiment of the present invention, the space invariance amount refers in observation main body with difference
Identical spatial point in the observation twice that pose carries out.
Preferably, according to the method for the embodiment of the present invention, the predetermined temporal voting strategy is：
Judge | dij-dij ' |/max (| dij |, | dij ' |) whether it is less than predetermined threshold, if | dij-dij ' |/max (|
Dij |, | dij ' |) be less than the predetermined threshold, then i-th of space variable and j-th of space variable logrolling, otherwise mutually
It votes against.
Preferably, according to the method for the embodiment of the present invention, the observation main body is robot.
Preferably, according to the method for the embodiment of the present invention, vote all space invariance amounts, and it is described pre-
Determining decision plan is：Judge that it is correct space invariance amount to obtain the most N number of space invariance amount of affirmative vote.
Preferably, according to the method for the embodiment of the present invention, the numerical value of N is determined according to design accuracy.
Preferably, according to the method for the embodiment of the present invention, arbitrarily a space invariance amount is selected to vote, and
The predetermined decision plan is：If selected space invariance measures the affirmative vote more than n/2, determine that it is correct
Space invariance amount, and the space invariance amount voted for for it is determined as correct space invariance amount, while delete as it
The space invariance amount voted against；And if selected space invariance amount does not obtain the affirmative vote more than n/2, it determines that it is
The space invariance amount of mistake, and arbitrarily choose one of remaining space invariance amount, then repeat more than handle, until being judged
Until correct space invariance amount.
Other side according to embodiments of the present invention provides a kind of robot rapid location method of estimation, including such as
Lower step：Robot observes unknown space with initial pose and current pose to obtain one group of observation chart respectively twice
Picture；The space invariance amount of same space point is represented in this group of observed image of Preliminary detection；According to the above method to space invariance amount
Secondary detection is carried out, to filter out incorrect matched space invariance amount；And it is obtained based on correct matched space invariance amount
Obtain the current pose of robot.
Another aspect according to embodiments of the present invention provides a kind of space invariance amount two for space map structuring
Secondary detection device, the equipment include：Acquiring unit, for observing main body successively with the first pose and second by using one
Pose carries out in one group of observed image obtained from observing twice unknown space, obtains the n space that Preliminary detection arrives
Invariant, wherein the observation main body at least possesses an image acquisition units to obtain observed image, and n is natural number；Distance
Computing unit, for calculating the distance of additional space invariant in observation twice respectively, wherein for i-th of space invariance amount,
With the distance of j-th of space invariance amount be dij in being observed in first time, in being observed at second with j-th space invariance amount
Distance is dij ', and wherein i, j, which are respectively less than, is equal to n, and j ≠ i；Ballot unit, for being based on predetermined temporal voting strategy, to specific quantity
Space invariance amount vote, wherein for some space invariance amount carry out ballot refer to by every other space invariance amount
It votes for some described space invariance amount；And identifying unit, for according to voting results and predetermined judgement plan
Correctly N number of space invariance amount is judged slightly among n space invariance amount, so as to filter out incorrect matched space invariance
Amount, wherein N are natural number and are less than or equal to n.
Preferably, in equipment according to embodiments of the present invention, the space invariance amount refers in observation main body with difference
Identical spatial point in the observation twice that pose carries out.
Preferably, in equipment according to embodiments of the present invention, the predetermined temporal voting strategy is：Judge | djj-dij ' |/
Max (| dij |, | dij ' |) whether it is less than predetermined threshold, if | dij-dij ' |/max (| dij |, | dij ' |) less than described pre-
Determine threshold value, then i-th of space variable and j-th of space variable logrolling, otherwise mutually vote against.
Preferably, in equipment according to embodiments of the present invention, the observation main body is robot.
Preferably, in equipment according to embodiments of the present invention, the ballot unit throws all space invariance amounts
Ticket, and the predetermined decision plan is：Judge that it is correct space invariance to obtain the most N number of space invariance amount of affirmative vote
Amount.
Preferably, in equipment according to embodiments of the present invention, the numerical value of N is determined according to design accuracy.
Preferably, in equipment according to embodiments of the present invention, the ballot unit arbitrarily one space invariance amount of selection
It votes, and the predetermined decision plan is：If selected space invariance measures the affirmative vote more than n/2,
Correct space invariance amount is determined that it is, and the space invariance amount voted for for it is determined as correct space invariance
Amount, while delete the space invariance amount voted against for it；And if selected space invariance amount does not obtain praising more than n/2
Into ticket, then determine that it is the space invariance amount of mistake, and arbitrarily choose one of remaining space invariance amount, then repeat more than
Reason, until obtaining being determined as correct space invariance amount.
Another aspect according to embodiments of the present invention provides a kind of robot rapid location estimation equipment, including：Figure
As collecting unit, unknown space is observed twice with initial pose and current pose for respectively to obtain one group of observation
Image；Space invariance amount Preliminary detection unit, Preliminary detection image acquisition units obtain this group of observed image in represent identical
The space invariance amount of spatial point；Space invariance amount secondary detection equipment as described above, for space invariance amount Preliminary detection
The space invariance amount that unit detects carries out secondary detection, to filter out incorrect matched space invariance amount；And pose
Computing unit, for obtaining the current pose of robot based on correct matched space invariance amount.
By space invariance amount secondary detection method and apparatus according to embodiments of the present invention, calculating can be greatly reduced
It measures and promotes processing speed, while improve the matching precision of space invariance amount.In addition, using above-mentioned space invariance amount secondary detection
The robot rapid location method of estimation and equipment of method and apparatus can reduce calculation amount and promote the same of processing speed
When, preferably complete pose estimation.
Description of the drawings
Figure 1A -1B are shown based on the schematic diagram for observing data progress position and attitude variation estimation twice；
Fig. 2A -2B show that incorrect space invariance amount triggers the schematic diagram of position and attitude erroneous estimation；
Fig. 3 A-3B respectively illustrate the relative distance between each space invariance amount in observation twice；
Fig. 4 is to show the space invariance amount secondary detection method for space map structuring according to embodiments of the present invention
Flow flow chart；
Fig. 5 is the flow chart for the flow for showing robot rapid location method of estimation according to embodiments of the present invention；
Fig. 6 is to show the space invariance amount secondary detection equipment for space map structuring according to embodiments of the present invention
Configuration block diagram；And
Fig. 7 is the block diagram for the configuration for showing robot rapid location estimation equipment according to embodiments of the present invention.
Specific embodiment
Each preferred embodiment of the present invention is described below with reference to accompanying drawings.It provides referring to the drawings
Description, to help the understanding of the example embodiment of the present invention to being limited by appended claims and their equivalents.It includes side
The various details of assistant's solution, but they can only be counted as illustratively.Therefore, it would be recognized by those skilled in the art that
Embodiment described herein can be made various changes and modifications, without departing from scope and spirit of the present invention.Moreover, in order to
Make specification more clear succinct, will omit pair it is well known that function and the detailed description of construction.
According to the method for the embodiment of the present invention in equipment, relative distance between each space invariance amount as judge according to
According to, to each invariant carry out vote by ballot, quickly filter out correct space invariance amount.Fig. 3 A-3B are respectively illustrated to be seen twice
Relative distance in survey between each space invariance amount.As shown in Fig. 3 A, 3B, for correct space invariance amount (1,2,3,4,5)
For, the relative distance in observing twice between them remains unchanged.
Fig. 4 is to show the space invariance amount secondary detection method for space map structuring according to embodiments of the present invention
Flow flow chart.
As shown in figure 4, described method includes following steps：
First, in step S401, main body is being observed successively with the first pose and the second pose to unknown sky by using one
Between observed twice obtained from one group of observed image, obtain the n space invariance amount that Preliminary detection arrives, wherein institute
It states observation main body and at least possesses an image acquisition units to obtain observed image, and n is natural number.For example, sight described here
It can be robot to survey main body, and the first pose can be (0,0,0,0,0,0), the second pose can be (x, y, z, roll,
Pitch, yaw).In addition, it is necessary to, it is noted that the space invariance amount refers to carry out twice with different positions and pose in observation main body
Identical spatial point in observation.
Then, in step S402, the distance of additional space invariant in observation twice is calculated respectively, wherein for i-th
Space invariance amount with the distance of j-th of space invariance amount is dij in being observed in first time, empty with j-th in being observed at second
Between invariant distance for dij ', wherein i, j, which are respectively less than, is equal to n, and j ≠ i.
Then, in step S403, based on predetermined temporal voting strategy, vote certain amount of space invariance amount.Here,
According to the difference of the quantity for the space invariance amount voted, two kinds of embodiments of full ballot and part ballot can be divided into.It needs
Refer to, it is noted that carrying out ballot for some space invariance amount by every other space invariance amount for some described sky
Between invariant vote.
For example, the predetermined temporal voting strategy is：Judge | dij-dij ' |/max (| dij |, | dij ' |) whether less than predetermined
Threshold value, if | dij-dij ' |/max (| dij |, | dij ' |) be less than the predetermined threshold, then i-th of space variable and j-th
Space variable logrolling, otherwise mutually votes against.
Finally, in step S404, judged just among n space invariance amount according to voting results and predetermined decision plan
True N number of space invariance amount, so as to filter out incorrect matched space invariance amount, wherein N is natural number and is less than or equal to
n。
In the following, full ballot embodiment is described first.In this embodiment, vote all space invariance amounts.At this
In the case of kind, the predetermined decision plan in above-mentioned steps S404 is：Judge to obtain the most N number of space invariance amount of affirmative vote
For correct space invariance amount.
It is pointed out that N here can be adjusted according to design accuracy.For example, design accuracy is higher, then it is above-mentioned to sentence
The quantity for the space invariance amount chosen in fixed strategy is bigger, i.e. N values are bigger.Conversely, then N values are smaller.In general, the value of N
Scope is 50~300.
As described above, the accuracy of existing RANSAC methods is limited be subject to experiment number, it is only capable of finding accurately
Space invariance amount.And by the full ballot embodiment according to the present invention, space invariance amount the most accurate can be found.And
And due to the position and attitude variable quantity that multiple calculating observation main body (that is, robot) is not required, all calculating pertain only to simple
Distance calculates and size judges, therefore calculation amount is greatly reduced, and processing speed gets a promotion.After tested, performance can
Promote 1~2 order of magnitude.
Next, description section is voted embodiment.The embodiment it is different from the full ballot embodiment that describes before it
It is in not voting whole space invariance amounts.Alternatively, arbitrarily one space invariance amount of selection is thrown first
Ticket.In this case, the predetermined decision plan in above-mentioned steps S404 is：If selected space invariance measures
Affirmative vote more than n/2 then determines that it is correct space invariance amount, and the space invariance amount voted for for it is judged
For correct space invariance amount, while delete the space invariance amount voted against for it；If selected space invariance amount is not
Obtain the affirmative vote more than n/2, then determine that it is the space invariance amount of mistake, and arbitrarily choose remaining space invariance amount it
One, then repeat more than handle, until obtaining being determined as correct space invariance amount.
According to above description, part ballot embodiment reduces further fortune on the basis of ballot embodiment entirely
Calculation amount simultaneously improves processing speed.But part ballot embodiment, compared with ballot embodiment entirely, shortcoming is that precision is not high.
In other words, it is impossible to space invariance amount the most accurate can be found embodiment as voting entirely.Therefore, according to specific design
It is required that it can suitably choose full ballot or part ballot embodiment.
The foregoing describe space invariance amount secondary detection methods according to embodiments of the present invention.It should be evident that this method
It can be applied to robot rapid location estimation of the prior art.Since space invariance amount according to embodiments of the present invention is secondary
Detection method can find correct matched space invariance amount while reducing calculation amount and promoting processing speed, so as to
Preferably complete the position and attitude estimation of robot.
The robot of space invariance amount secondary detection method according to embodiments of the present invention is applied referring to Fig. 5 descriptions
The flow of rapid location method of estimation.
As shown in figure 5, robot rapid location method of estimation includes the following steps：
First, in step S501, unknown space is respectively observed with initial pose and current pose twice by robot
So as to obtain one group of observed image.
Then, the space invariance amount of same space point is represented in step S502, Preliminary detection group observation data.Example
Such as, as previously mentioned, in the prior art, had at present many achievements in research can more efficiently from 2-D gray image or
Space invariance amount (for example, by the matched mode of gray value) is detected in three-dimensional point cloud observation data.This is space invariance
The preliminary matches of amount.
Next, in step S503, by according to above with reference to the space invariance amount secondary detection method described in Fig. 4 to sky
Between invariant carry out secondary detection, to filter out incorrect matched space invariance amount.
Finally, in step S504, the current pose based on correct matched space invariance amount acquisition robot.
It is secondary next, with reference to Fig. 6 descriptions space invariance amount for space map structuring according to embodiments of the present invention
Detection device.As shown in fig. 6, the equipment 600 includes acquiring unit 601, metrics calculation unit 602, ballot unit 603 and sentences
Order member 604.
Acquiring unit 601 successively carries out unknown space with the first pose and the second pose by using an observation main body
In one group of observed image obtained from observing twice, the n space invariance amount that Preliminary detection arrives is obtained.Wherein, the sight
It surveys main body and at least possesses an image acquisition units to obtain observed image, and n is natural number.
In the n space invariance amount that metrics calculation unit 602 is obtained in acquiring unit 601, calculate respectively in observation twice
The distance of additional space invariant, wherein for i-th of space invariance amount, in being observed in first time with j-th of space invariance amount
Distance for dij, with the distance of j-th of space invariance amount be dij ' in being observed at second, wherein i, j, which are respectively less than, is equal to n, and
j≠i。
Relative distance between each space invariance amount that ballot unit 603 is calculated according to metrics calculation unit 602, is based on
Predetermined temporal voting strategy votes to certain amount of space invariance amount.
Here, according to the difference of the quantity for the space invariance amount voted, full ballot and part ballot two can be divided into
Kind embodiment, and carry out ballot for some space invariance amount and refer to by every other space invariance amount for described a certain
A space invariance amount is voted.
The voting results and predetermined decision plan that identifying unit 604 is obtained according to ballot unit 603 are in n space invariance
Correctly N number of space invariance amount is judged among amount, so as to filter out incorrect matched space invariance amount, wherein N is natural number
And less than or equal to n.
Due in space invariance amount secondary detection equipment according to embodiments of the present invention, related " predetermined temporal voting strategy ",
" embodiment of voting entirely ", " part ballot embodiment " and the content of " predetermined decision plan " and aforesaid space invariant secondary detection
Corresponding content in method is identical, therefore for brevity, details repeats no more.
Similarly, above-mentioned space invariance amount secondary detection equipment 600 according to embodiments of the present invention can also be applied to existing
There is the robot rapid location estimation equipment in technology.Sky according to embodiments of the present invention is applied below with reference to Fig. 7 descriptions
Between invariant secondary detection equipment robot rapid location estimation equipment 700.
As shown in fig. 7, robot rapid location estimation equipment 700 includes：At the beginning of image acquisition units 701, space invariance amount
Walk detection unit 702, space invariance amount secondary detection equipment 600 as described above and pose computing unit 703.
Image acquisition units 701 are observed obtaining twice with initial pose and current pose to unknown space respectively
One group of observed image.
It is preliminary to examine in this group of observed image that space invariance amount Preliminary detection unit 702 is obtained in image acquisition units 701
Survey the space invariance amount for representing same space point；
The space invariance that space invariance amount secondary detection equipment 600 detects space invariance amount Preliminary detection unit 702
Amount carries out secondary detection, to filter out incorrect matched space invariance amount；
Pose computing unit 703 is based on the correct matched space after the filtering of space invariance amount secondary detection equipment 600
Invariant obtains the current pose of robot.
Above by reference to attached drawing, describe space invariance amount secondary detection method and apparatus according to embodiments of the present invention and
Apply its robot rapid location method of estimation and equipment.Pass through the secondary inspection of space invariance amount according to embodiments of the present invention
Method and apparatus is surveyed, calculation amount can be greatly reduced and promote processing speed, while improve the matching precision of space invariance amount.
In addition, using the robot rapid location method of estimation and equipment of above-mentioned space invariance amount secondary detection method and apparatus, it can
While reducing calculation amount and promoting processing speed, pose estimation is preferably completed.
It should be noted that in the present specification, term " comprising ", "comprising" or its any other variant are intended to
Non-exclusive inclusion, so that process, method, article or equipment including a series of elements not only will including those
Element, but also including other elements that are not explicitly listed or further include as this process, method, article or equipment
Intrinsic element.In the absence of more restrictions, the element limited by sentence " including ... ", it is not excluded that wrapping
Include in process, method, article or the equipment of the element that also there are other identical elements.
Finally, it is to be noted that, a series of above-mentioned processing are not only included with order described here in temporal sequence
The processing of execution, and the processing including performing parallel or respectively rather than in chronological order.
Through the above description of the embodiments, those skilled in the art can be understood that the present invention can be by
Software adds the mode of required hardware platform to realize, naturally it is also possible to all be implemented by software.Based on such understanding,
Technical scheme can be embodied in the form of software product in whole or in part to what background technology contributed,
The computer software product can be stored in storage medium, such as ROM/RAM, magnetic disc, CD, including some instructions making
Computer equipment (can be personal computer, server or the network equipment etc.) performs each embodiment of the present invention
Or the method described in some parts of embodiment.
The present invention is described in detail above, specific case used herein is to the principle of the present invention and embodiment party
Formula is set forth, and the explanation of above example is only intended to help to understand method and its core concept of the invention；It is meanwhile right
In those of ordinary skill in the art, thought according to the invention has change in specific embodiments and applications
Part, in conclusion this specification content should not be construed as limiting the invention.
Claims (14)
1. a kind of space invariance amount secondary detection method for space map structuring, the described method includes：
It is obtained successively being observed twice unknown space with the first pose and the second pose by using an observation main body
One group of observed image in, n space invariance amount that Preliminary detection arrives is obtained, wherein the main body of observing at least possesses one
Image acquisition units are to obtain observed image, and n is natural number；
The distance of additional space invariant in observation twice is calculated respectively, wherein for i-th of space invariance amount, is seen in first time
With the distance of j-th of space invariance amount it is dij in survey, with the distance of j-th of space invariance amount is dij ' in being observed second,
Wherein i, j, which are respectively less than, is equal to n, and j ≠ i；
It based on predetermined temporal voting strategy, votes certain amount of space invariance amount, wherein for some space invariance amount
Ballot is carried out to refer to be voted for some described space invariance amount by every other space invariance amount；And
Correct N number of space invariance amount is judged among n space invariance amount according to voting results and predetermined decision plan, from
And incorrect matched space invariance amount is filtered out, wherein N is natural number and is less than or equal to n；
Wherein described predetermined temporal voting strategy is：
Judge | dij-dij ' |/max (| dij |, | dij ' |) whether it is less than predetermined threshold, if | dij-dij ' |/max (| dij
|, | dij ' |) be less than the predetermined threshold, then otherwise i-th of space variable and j-th of space variable logrolling is mutually thrown anti-
To ticket.
2. according to the method described in claim 1, wherein described space invariance amount refers to carry out with different positions and pose in observation main body
Observe twice in identical spatial point.
3. according to the method described in claim 1, wherein described observation main body is robot.
4. according to the method described in claim 1, wherein
It votes all space invariance amounts, and
The predetermined decision plan is：Judge that it is correct space invariance amount to obtain the most N number of space invariance amount of affirmative vote.
5. according to the method described in claim 1, wherein
Arbitrarily one space invariance amount of selection is voted, and
The predetermined decision plan is：If selected space invariance measures the affirmative vote more than n/2, determine that it is just
True space invariance amount, and the space invariance amount voted for for it is determined as correct space invariance amount, it deletes simultaneously
The space invariance amount voted against for it；
If selected space invariance amount does not obtain the affirmative vote more than n/2, the space invariance amount of mistake is determined that it is, and
Arbitrarily choose one of remaining space invariance amount, then repeat more than handle, until obtaining being determined as correct space invariance amount
Until.
6. according to the method described in claim 4, the numerical value of N is wherein determined according to design accuracy.
7. a kind of robot rapid location method of estimation, includes the following steps：
Robot observes unknown space with initial pose and current pose to obtain one group of observed image respectively twice；
The space invariance amount of same space point is represented in this group of observed image of Preliminary detection；
Secondary detection is carried out to space invariance amount by the method according to claim 11, to filter out incorrect matching
Space invariance amount；And
The current pose of robot is obtained based on correct matched space invariance amount.
8. a kind of space invariance amount secondary detection equipment for space map structuring, the equipment includes：
Acquiring unit, for successively carrying out two to unknown space with the first pose and the second pose by using an observation main body
In one group of observed image obtained from secondary observation, the n space invariance amount that Preliminary detection arrives is obtained, wherein the observation master
Body at least possesses an image acquisition units to obtain observed image, and n is natural number；
Metrics calculation unit, for calculating the distance of additional space invariant in observation twice respectively, wherein for i-th of space
Invariant with the distance of j-th of space invariance amount is dij in being observed in first time, in being observed at second with j-th of space not
The distance of variable is dij ', and wherein i, j, which are respectively less than, is equal to n, and j ≠ i；
Ballot unit, for being based on predetermined temporal voting strategy, votes to certain amount of space invariance amount, wherein for a certain
A space invariance amount carries out ballot and refers to be voted for some described space invariance amount by every other space invariance amount；
And
Identifying unit, it is correctly N number of for being judged according to voting results and predetermined decision plan among n space invariance amount
Space invariance amount, so as to filter out incorrect matched space invariance amount, wherein N is natural number and is less than or equal to n；
Wherein described predetermined temporal voting strategy is：
Judge | dij-dij ' |/max (| dij |, | dij ' |) whether it is less than predetermined threshold, if | dij-dij ' |/max (| dij
|, | dij ' |) be less than the predetermined threshold, then otherwise i-th of space variable and j-th of space variable logrolling is mutually thrown anti-
To ticket.
9. equipment according to claim 8, wherein the space invariance amount refers to carry out with different positions and pose in observation main body
Observe twice in identical spatial point.
10. equipment according to claim 8, wherein the observation main body is robot.
11. equipment according to claim 8, wherein
The ballot unit votes to all space invariance amounts, and
The predetermined decision plan is：Judge that it is correct space invariance amount to obtain the most N number of space invariance amount of affirmative vote.
12. equipment according to claim 8, wherein
Ballot unit arbitrarily one space invariance amount of selection is voted, and
The predetermined decision plan is：If selected space invariance measures the affirmative vote more than n/2, determine that it is just
True space invariance amount, and the space invariance amount voted for for it is determined as correct space invariance amount, it deletes simultaneously
The space invariance amount voted against for it；
If selected space invariance amount does not obtain the affirmative vote more than n/2, the space invariance amount of mistake is determined that it is, and
Arbitrarily choose one of remaining space invariance amount, then repeat more than handle, until obtaining being determined as correct space invariance amount
Until.
13. equipment according to claim 11, wherein determining the numerical value of N according to design accuracy.
14. a kind of robot rapid location estimates equipment, including：
Image acquisition units observe unknown space with initial pose and current pose for respectively to obtain one twice
Group observed image；
Space invariance amount Preliminary detection unit, Preliminary detection image acquisition units obtain this group of observed image in represent identical sky
Between the space invariance amount put；
Space invariance amount secondary detection equipment according to claim 8, for being examined to space invariance amount Preliminary detection unit
The space invariance amount measured carries out secondary detection, to filter out incorrect matched space invariance amount；And
Pose computing unit, for obtaining the current pose of robot based on correct matched space invariance amount.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110444217.8A CN103186896B (en) | 2011-12-27 | 2011-12-27 | For the space invariance amount secondary detection method and apparatus of space map structuring |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110444217.8A CN103186896B (en) | 2011-12-27 | 2011-12-27 | For the space invariance amount secondary detection method and apparatus of space map structuring |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103186896A CN103186896A (en) | 2013-07-03 |
| CN103186896B true CN103186896B (en) | 2018-06-01 |
Family
ID=48678054
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201110444217.8A Active CN103186896B (en) | 2011-12-27 | 2011-12-27 | For the space invariance amount secondary detection method and apparatus of space map structuring |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103186896B (en) |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4492036B2 (en) * | 2003-04-28 | 2010-06-30 | ソニー株式会社 | Image recognition apparatus and method, and robot apparatus |
| US7844105B2 (en) * | 2007-04-23 | 2010-11-30 | Mitsubishi Electric Research Laboratories, Inc. | Method and system for determining objects poses from range images |
| US20100312386A1 (en) * | 2009-06-04 | 2010-12-09 | Microsoft Corporation | Topological-based localization and navigation |
-
2011
- 2011-12-27 CN CN201110444217.8A patent/CN103186896B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN103186896A (en) | 2013-07-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9734414B2 (en) | Unified framework for precise vision-aided navigation | |
| CN103502833B (en) | Localization method | |
| JP5759161B2 (en) | Object recognition device, object recognition method, learning device, learning method, program, and information processing system | |
| Zhang et al. | Hierarchical topic model based object association for semantic SLAM | |
| JP5800494B2 (en) | Specific area selection device, specific area selection method, and program | |
| CN103889049A (en) | Wireless signal indoor positioning system and method based on inertia measurement element assistance | |
| US9651388B1 (en) | System and method for improved simultaneous localization and mapping | |
| CN106649709B (en) | Vehicle stopping behavior pattern prediction and evaluation method based on data mining | |
| McBrearty et al. | Earthquake arrival association with backprojection and graph theory | |
| CN109916488B (en) | Dynamic vehicle weighing method and device | |
| US20150301216A1 (en) | Method for automatic recognition of a mobile magnetic object | |
| CN107656265A (en) | Particle filter fusion method for tracking short flight path before multi frame detection | |
| CN110352389A (en) | Information processing unit and information processing method | |
| CN106441279A (en) | Robot positioning method based on self-positioning and edge searching and system | |
| CN110095769A (en) | A kind of method for detecting parking stalls, device and electronic equipment | |
| CN105444763A (en) | IMU indoor positioning method | |
| CN107274679A (en) | Vehicle identification method, device, equipment and computer-readable recording medium | |
| CN103885016A (en) | Determination of a phase difference map | |
| CN105387844B (en) | Pavement behavior measurement system and pavement behavior assay method | |
| WO2011076767A1 (en) | Method and device for identifying a subset of measurements, method and system for locating an object, recording medium for these methods | |
| CN108663675A (en) | The method positioned simultaneously for life detection radar array multiple target | |
| CN105021199B (en) | Multi-model self-adapting method for estimating state and system based on LS | |
| CN102981160B (en) | Method and device for ascertaining aerial target track | |
| JP2017010393A (en) | Map generation device and program | |
| CN104361600B (en) | motion recognition method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |