CN106840161A

CN106840161A - Air navigation aid and device

Info

Publication number: CN106840161A
Application number: CN201611170559.4A
Authority: CN
Inventors: 霍光磊; 王瑾琨; 常元章; 严洁; 易梅
Original assignee: Hna Ecology Technology Group Co Ltd
Current assignee: Hna Ecology Technology Group Co Ltd
Priority date: 2016-12-16
Filing date: 2016-12-16
Publication date: 2017-06-13

Abstract

The invention discloses a kind of air navigation aid and device.Wherein, the method includes：Obtain starting point semantic region and target semantic region in predetermined space, wherein, the predetermined space is divided into multiple semantic regions, the multiple semantic region includes the starting point semantic region and the target semantic region, and the feature for identifying the semantic region from the multiple semantic region is included in each semantic region；Based on semantic region planning from the starting point semantic region to the path of the target semantic region；Equipment is set to move to the target semantic region from the starting point semantic region according to the path.The present invention solves the too wide in range technical problem of navigational semantic scope.

Description

Air navigation aid and device

Technical field

The present invention relates to navigation field, in particular to a kind of air navigation aid and device.

Background technology

2005, Galindo etc. built the double-deck map of spatial information layer and semantic information layer, and every layer of map passes through " anchor " Association.Semantic information is absorbed in their research, the data (such as view data or grid information) gathered by sensor, makes The information for obtaining space Information Level becomes meaningful.Indoor material object is the source of semantic information, in the map that they design, language Adopted information is artificial given, though not having flexibility, is realized simple.

Vasudevan etc. devises a kind of full probability semanteme map, and semantic information is realized by object identification.Indoor environment Comprising object be converted to space with semanteme it is abstract.Spatial abstraction is converted to concept by robot, is recognized with semantic abstraction Object.Object collection obtains concept by training.Test phase divides space to object detection.The information of object in indoor environment It is used for classifying space, but classification information cannot be used for object in reasoning room.

Meger etc. realizes the automatic detection of indoor object, forms the geometry map with object information.They realize The decorum it is more advanced, and realize a kind of vision subsystem.Map structuring realizes that the model of object is relied on by FastSLAM View data on Internet, by training object in divided chamber.The advantage of the system is object in automatic detection room, clothes Business robot can automatically obtain the semantic information in room.Viswanathan etc. proposes a kind of robustness with automation energy The more preferable system of power, their system extracts identification Spatial Semantics using the semantic label of entity information.Their system data It is same to be obtained from Internet, Bayesian model is built according to the frequency for detecting entity, this method test result indicate that, The system can stably search for typical entity, and reasoning typical space is semantic.The method is relative to method mentioned above More stablize.

Hertzberg etc. proposes the Spatial Semantics map based on object.Their object is led to by 3D information acquisitions Cross laser radar and realize a kind of 6DSLAM methods.The method obtains a cloud information first, then constructs coarse feature, as Face, wall etc..On this basis by detection of classifier object, by project objects to map.The information for finally giving is favourable In observation, augmented reality is conducive to realize.This method has been related to point cloud information, and the data volume for the treatment of is higher, while calculating Complexity is bigger than normal.

For the too wide in range problem of above-mentioned navigational semantic scope, effective solution is not yet proposed at present.

The content of the invention

A kind of air navigation aid and device are the embodiment of the invention provides, it is too wide in range at least to solve navigational semantic scope Technical problem.

A kind of one side according to embodiments of the present invention, there is provided air navigation aid, including：Obtain in predetermined space Starting point semantic region and target semantic region, wherein, the predetermined space is divided into multiple semantic regions, the multiple semanteme Region includes the starting point semantic region and the target semantic region, includes being used for from the multiple in each semantic region The feature of the semantic region is identified in semantic region；Based on semantic region planning from the starting point semantic region to the target The path of semantic region；Equipment is set to move to the target semantic region from the starting point semantic region according to the path.

Further, the equipment is made to move to the target semantic space from the starting point semantic region according to the path Domain includes：The image for obtaining is identified by the equipment obtain the semantic region corresponding to the image, wherein, it is described to set The standby semantic region identified is used for the semantic region where determining the equipment at present；According to language of the equipment where current Adopted region and the path make the equipment move to the target semantic region.

Further, wrap the semantic region for being identified obtaining to the image for obtaining corresponding to the image by the equipment Include：Obtain the characteristic point in described image；Characteristic point in described image is corresponding with each semantic region being pre-configured with Characteristic point in characteristic point storehouse is matched, wherein, be have recorded in the characteristic point storehouse each semantic region it is corresponding one or Multiple characteristic points；Determine the characteristic point in the characteristic point characteristic point storehouse corresponding with described each semantic region in described image Whether the quantity matched somebody with somebody meets predetermined condition；In the case where the predetermined condition is met, the corresponding semantic space of described image is determined Domain.

Further, the characteristic point in described image is obtained, including：Determine whether the quantity of characteristic point in described image belongs to In preset range；In the case where preset range is belonged to, the characteristic point in described image is obtained.

Further, it is determined that the quantity of Feature Points Matching in characteristic point in described image and the characteristic point storehouse whether Meeting predetermined condition includes：The spy in characteristic point characteristic point storehouse corresponding with described each semantic region in described image The quantity and predetermined threshold for levying Point matching are compared；Determine that the quantity of the matching meets the predetermined bar more than predetermined threshold Part.

Further, it is determined that the quantity of Feature Points Matching in characteristic point in described image and the characteristic point storehouse whether Meeting predetermined condition includes：The spy in characteristic point characteristic point storehouse corresponding with described each semantic region in described image The quantity of Point matching is levied, described each semantic region is ranked up according to the quantity order from more to less of the matching；Really Determine clooating sequence and meet predetermined condition in the first semantic region.

Further, wrapped from the starting point semantic region to the path of the target semantic region based on semantic region planning Include：Based on the semantic analysis to assignment instructions, the semantic relation in the assignment instructions is drawn；According to the semantic relation, really Fixed the starting point semantic region and the target semantic region, planning is from the starting point semantic region to the target semantic region Path.

Further, wrapped from the starting point semantic region to the path of the target semantic region based on semantic region planning Include：Semantic region in the predetermined space is mapped as the regular figure block on the corresponding map of the predetermined space；According to The corresponding regular figure slip gauge stroke in the semantic region path.

Further, drawing the path according to the corresponding regular figure slip gauge in the semantic region includes：According to Bayes The model of reasoning is planned from the starting point semantic region to the path of the target semantic region.

Further, the regular figure block be rectangle, and the equipment be only capable of in the rectangle along it is upper and lower, In the case that left and right four direction is moved, the model of the Bayesian inference is as follows：O represents the four direction of the rectangle,The upward direction movement of the semantic region N that the equipment can be in the predetermined space R is represented,Represent institute The equipment of stating can in downward direction moving in the semantic region N in the predetermined space R,Represent that the equipment can The left direction movement of the semantic region N in the predetermined space R,Represent that the equipment can be described predetermined The right direction movement of the semantic region N in the R of space.

Further, it is rectangle in the regular image block, and the equipment can turn every time in the rectangle In the case that dynamic direction is 90 °, determine that the equipment moves to the number of revolutions of target semantic region from certain semantic region, its In, the number of revolutions moves to a upper moving direction of a upper semantic region according to the equipment and the equipment is moved to and worked as The current moving direction of preceding semantic region determines that the Bayesian inference model includes as follows：Represent the equipment from upper One semantic region moves to the number of revolutions in current semantics region, the o_iRepresent a described upper movement side of the equipment To the o_jRepresent the described current moving direction of the equipment, the o_iWith the o_jFour, upper and lower, left and right are can use respectively Moving direction；If the upper moving direction o_iWith the current moving direction o_jIt is identical, then the number of revolutionsIf The upper moving direction o_iWith the current moving direction o_j90 ° are differed, then the number of revolutionsIf described upper one Moving direction o_iWith the current moving direction o_j180 ° are differed, then the number of revolutions

Further, transition probability is that the equipment after a upper moving direction of a semantic region, is moved again on moving to The probability of the current moving direction in current semantics region is moved, wherein, a upper semantic region and the current semantics region It is neighboring semantic region；Described in the transition probability moves to current semantics region according to the equipment from a upper semantic region Whole direction of rotation of number of revolutions and the equipment calculate, and the Bayesian inference model includes as follows：The mark The formula of transition probability is construed to：Wherein,Table Show that the equipment after a upper moving direction of a semantic region, moves to the current movement in current semantics region again on moving to The probability in direction, wherein,Represent the transition probability, R_MWith R_NIt is two neighboring semantic regions in the predetermined space； The computing formula of the transition probability is：Wherein, Ω_oRepresent the equipment whole rotation direction.

Further, according to Bayesian network recurrence relation and the transition probability, obtain every in same predetermined space The probability of paths, the Bayesian inference model includes as follows：Represent per paths Probability, wherein,According to the maximum probability for calculatingPlan the path.

Further, wrapped from the starting point semantic region to the path of the target semantic region based on semantic region planning Include：Planning includes predetermined access path from the starting point semantic region to the path of the target semantic region, wherein, it is described pre- Determine access path to pre-set, the starting point semantic region and the target semantic region are in different predetermined spaces.

Further, wrapped from the starting point semantic region to the path of the target semantic region based on semantic region planning Include：Semantic region in the predetermined space is mapped as the semantic mark on the corresponding map of the predetermined space, wherein, often The individual semantic region includes one or more semanteme marks, and the semanteme is designated the characteristic point of object in semantic region；Root According to corresponding one or more the described paths of semanteme mark planning in the semantic region.

Further, included according to corresponding one or more the described paths of semanteme mark planning in the semantic region：Root According to the planning of particle filter model from the starting point semantic region to the path of the target semantic region.

Further, can be in the case where semantic mark be gathered in being moved through, according to the predetermined sky in the equipment Between semantic mark on corresponding map, determine the neighboring semantic mark of the last time semantic mark of the equipment collection, it is determined that The equipment includes as follows by collecting the semantic probability for identifying after movement according to the particle filter model：N_L Represent the quantity of the neighboring semantic mark；N_PRepresent the quantity of particle in the particle filter model；The semantic markNeighboring semantic mark set beWherein, it is describedRepresent the semantic mark in the predetermined space K Know l, the Q^COnly one semantic mark in the predetermined space C is represented, it is describedRepresent the language in the predetermined space M Justice identifies m, describedRepresent the semantic mark r in the predetermined space R；q_tRepresent the current state of the equipment；O is represented The arbitrary semantic mark of collection；Represent the particle i of current state；Represent that the particle i of current state is allocated ArriveProbability；Expression collects semantic markProbability；Represent the i-th current power of particle Weight.

Further, mark transition probability is that the equipment once after semantic mark, is adopted again on collecting by mobile Collect the mark probability of current semantics mark, wherein, the semanteme is designated neighboring semantic mark；The mark transition probability root Determine that the particle filter model includes as follows according to the interval between two related to equipment movement semantic marks： The formula of the mark transition probability is construed to：Wherein,Table Show and collecting semantic markAfterwards, semantic mark is collected againProbability,Represent the mark transition probability, m It is the neighboring semantic mark with n, K and H is the neighboring semantic region；It is described mark transition probability formula be：Wherein, d represents the interval between two semantic marks related to equipment movement, σ_dRepresent It is spaced the standard deviation of d；Represent semantic markIdentified with semantemeBetween interval：Wherein,Represent in language Justice markIdentified with semantemeBetween it is adjacent；Represent semantic markIdentified with semantemeBetween also have one Individual semantic mark；Represent semantic markIdentified with semantemeBetween there is the predetermined of only one semantic mark Space；Represent that only one predetermined space of semantic mark is identified with semanticBetween interval；Wherein,Represent Only one predetermined space of semantic mark is identified with semanticIt is adjacent；Represent that only one semanteme is identified predetermined Space identifies with semanticBetween also have a semantic mark.

Another aspect according to embodiments of the present invention, additionally provides a kind of guider, including：Acquiring unit, for obtaining Starting point semantic region and target semantic region in predetermined space are taken, wherein, the predetermined space is divided into multiple semantemes Region, the multiple semantic region includes the starting point semantic region and the target semantic region, in each semantic region Including the feature for identifying the semantic region from the multiple semantic region；Planning unit, for based on semantic region Planning is from the starting point semantic region to the path of the target semantic region；Mobile unit, for making equipment according to the road Footpath moves to the target semantic region from the starting point semantic region.

Further, the mobile unit includes：Identification module, for being known to the image for obtaining by the equipment The semantic region corresponding to the image is not obtained, wherein, the semantic region that the equipment is identified is used to determine the equipment mesh The semantic region at preceding place；Mobile module, institute is made for the semantic region and the path according to the equipment where current The equipment of stating moves to the target semantic region.

Further, the identification module includes：First acquisition module, for obtaining the characteristic point in described image； With module, for by the feature in the characteristic point storehouse corresponding with each semantic region being pre-configured with of the characteristic point in described image Point is matched, wherein, corresponding one or more characteristic points in each semantic region are have recorded in the characteristic point storehouse；First is true Cover half block, for determining the characteristic point in the characteristic point storehouse corresponding with described each semantic region of the characteristic point in described image Whether the quantity matched somebody with somebody meets predetermined condition；Second determining module, in the case where the predetermined condition is met, it is determined that described The corresponding semantic region of image.

Further, the acquisition module includes：3rd determining module, the quantity for determining characteristic point in described image Whether preset range is belonged to；Second acquisition module, in the case where preset range is belonged to, obtaining the feature in described image Point.

Further, first determining module includes：Comparison module, the characteristic point in described image is every with described The quantity of the Feature Points Matching in the corresponding characteristic point storehouse in individual semantic region is compared with predetermined threshold；4th determining module, uses In it is determined that the quantity of the matching meets the predetermined condition more than predetermined threshold.

Further, first determining module includes：Order module, for the characteristic point in described image and institute State the quantity of the Feature Points Matching in the corresponding characteristic point storehouse in each semantic region, according to the matching quantity from more to less Order is ranked up to described each semantic region；5th determining module, for determining clooating sequence in the first semanteme Region meets predetermined condition.

Further, the planning unit includes：Analysis module, for based on the semantic analysis to assignment instructions, drawing Semantic relation in the assignment instructions；6th determining module, for according to the semantic relation, determining the starting point semantic space Domain and the target semantic region, plan from the starting point semantic region to the path of the target semantic region.

Further, the planning unit includes：First mapping block, for by the semantic region in the predetermined space It is mapped as the regular figure block on the corresponding map of the predetermined space；First planning submodule, for according to the semantic space The corresponding regular figure slip gauge in domain draws the path.

Further, the first planning submodule includes：First MBM, for the model according to Bayesian inference Planning is from the starting point semantic region to the path of the target semantic region.

Further, the planning unit includes：Predetermined module, for planning from the starting point semantic region to the mesh The path for marking semantic region includes predetermined access path, wherein, the predetermined access path pre-sets, the starting point language Adopted region and the target semantic region are in different predetermined spaces.

Further, the planning unit includes：Second mapping block, for by the semantic region in the predetermined space The semantic mark on the corresponding map of the predetermined space is mapped as, wherein, each described semantic region includes one or more Semanteme mark, the semantic mark includes the feature of the semantic region；Second planning submodule, for according to the semantic space Corresponding one or more the described paths of semanteme mark planning in domain.

Further, the second planning submodule includes：Second MBM, for being advised according to particle filter model Draw from the starting point semantic region to the path of the target semantic region.

In embodiments of the present invention, by obtaining the starting point semantic region in predetermined space and target semantic region, its In, the predetermined space is divided into multiple semantic regions, and the multiple semantic region includes the starting point semantic region and institute Target semantic region is stated, is included for identifying the semantic region from the multiple semantic region in each semantic region Feature；Based on semantic region planning from the starting point semantic region to the path of the target semantic region；Make equipment according to institute State path and move to the target semantic region from the starting point semantic region, and then it is too wide in range to solve navigational semantic scope Technical problem.

Brief description of the drawings

Accompanying drawing described herein is used for providing a further understanding of the present invention, constitutes the part of the application, this hair Bright schematic description and description does not constitute inappropriate limitation of the present invention for explaining the present invention.In the accompanying drawings：

Fig. 1 is a kind of schematic diagram of air navigation aid according to embodiments of the present invention；

Fig. 2 is the schematic diagram that a kind of optional semantic region according to embodiments of the present invention divides；

Fig. 3 is a kind of optional semantic region topological structure schematic diagram according to embodiments of the present invention；

Fig. 4 is the schematic diagram that a kind of optional semantic region according to embodiments of the present invention is mapped as graph block；

Fig. 5 is a kind of optional Bayesian network schematic diagram according to embodiments of the present invention；

Fig. 6 is a kind of optional particle filter model schematic according to embodiments of the present invention；

Fig. 7 is a kind of schematic diagram of guider according to embodiments of the present invention.

Specific embodiment

In order that those skilled in the art more fully understand the present invention program, below in conjunction with the embodiment of the present invention Accompanying drawing, is clearly and completely described to the technical scheme in the embodiment of the present invention, it is clear that described embodiment is only The embodiment of a part of the invention, rather than whole embodiments.Based on the embodiment in the present invention, ordinary skill people The every other embodiment that member is obtained under the premise of creative work is not made, should all belong to the model of present invention protection Enclose.

It should be noted that term " first ", " in description and claims of this specification and above-mentioned accompanying drawing Two " it is etc. for distinguishing similar object, without for describing specific order or precedence.It should be appreciated that so using Data can exchange in the appropriate case, so as to embodiments of the invention described herein can with except illustrating herein or Order beyond those of description is implemented.Additionally, term " comprising " and " having " and their any deformation, it is intended that cover Lid is non-exclusive to be included, for example, the process, method, system, product or the equipment that contain series of steps or unit are not necessarily limited to Those steps or unit clearly listed, but may include not list clearly or for these processes, method, product Or other intrinsic steps of equipment or unit.

According to embodiments of the present invention, there is provided a kind of air navigation aid and device embodiment, it is necessary to explanation, in accompanying drawing The step of flow is illustrated can perform in the such as one group computer system of computer executable instructions, and, although Logical order is shown in flow chart, but in some cases, can be performing shown different from order herein or retouch The step of stating.

Fig. 1 is a kind of schematic diagram of air navigation aid according to embodiments of the present invention, as shown in figure 1, the method is including as follows Step：

Step S102, obtains starting point semantic region and target semantic region in predetermined space, wherein, predetermined space quilt Multiple semantic regions are divided into, multiple semantic regions include starting point semantic region and target semantic region, in each semantic region Include the feature for identifying the semantic region from multiple semantic regions；

Step S104, based on semantic region planning from starting point semantic region to the path of target semantic region；

Step S106, makes equipment move to target semantic region from starting point semantic region according to path.

According to the above embodiment of the present invention, in predetermined space, according to the multiple features in the predetermined space, by predetermined sky Between be divided into multiple semantic regions, each semantic region is included for identifying the semantic region from multiple semantic regions Feature, also, the semantic region including equipment for dividing starting point semantic region and target semantic region, equipment is according to from starting point language Adopted region can move to target semantic space along path planning to the path planning of target semantic region from starting point semantic region Domain.Predetermined space is divided into multiple semantic regions according to feature, and according to the mobile road of the semantic region planning apparatus for dividing Footpath, can make equipment navigate to target semantic region from starting point semantic region exactly according to the path of planning, and then solve The too wide in range technical problem of navigational semantic scope.

In the above-described embodiments, can be that equipment is operated or mobile for dividing the predetermined space of semantic region Space, for the determination in the space, being manually set according to equipment user, or carried out according to the pre-set programs of equipment , can be operated for equipment or mobile space is divided into one or more predetermined space by identification, also, can also be right One or more of predetermined spaces carry out semantic region division.

In the above-described embodiments, semantic region division is carried out to predetermined space, can be according to equipment division set in advance Rule is realized, can also realized by being manually set division rule.As a kind of optional embodiment, in a certain predetermined sky Interior, the predetermined space can be divided into multiple semantic regions by equipment user by way of being manually set, wherein, language Dividing for adopted region can be divided according to the function of semantic region at this, can also be carried out according to the object in semantic region at this Divide, equipment can be identified by the function to semantic region at this, or object at this in semantic region is known Not, semantic region at this is determined.

Be there may be in predetermined space and be obscured by an object or do not allow the region that equipment is accessed, carried out to the part Can be blank semantic region by region division during semanteme is divided, the blank semantic region, will not by after division Accessed by equipment, i.e., equipment only passes through the semantic region.By the division carried out to the blank semantic region, can be by some The region for not allowing equipment to access isolates to equipment.

Alternatively, the equipment that activity is carried out in predetermined space can be robot device.

Fig. 2 is the schematic diagram that a kind of optional semantic region according to embodiments of the present invention divides, as shown in Fig. 2 equipment Moved in bedroom, hall and kitchen, each room is divided into multiple semantic regions, wherein, where each Roman number Shadow region represent a semantic region, the region of mark shade and Roman character is blank semantic region；Adjacent Magen David Line represent the boundary of each semantic region, each is used to represent that the Magen David English alphabet in semantic region boundary line enters rower Number.

The division of semantic region in predetermined space, can according in room some there is the thing of specific function or outward appearance Body, multiple semantic regions that can be named and can recognize that are divided into by room.For example, having refrigerator, micro-wave oven, gas-cooker etc. in kitchen Deng, there are bed, TV etc. in bedroom, each room can be divided into by multiple semantic regions according to the features described above in room.

Division to these semantic regions can also be divided according to the function in the region, for example, can draw kitchen It is divided into food holding area and machining area, dining room can be divided into region of having a meal, bedroom can also be divided into rest area Domain and sleep area etc..Division to these semantic regions can also be divided according to the outward appearance of object in the region, example Such as, kitchen can be divided into refrigerator region, micro-wave oven region, bedroom is divided into television area, sofa region and bed area Domain etc..During semantic region division is carried out to room, semantic region can be manually set according to the demand of user Division rule.

Equipment is based on the semantic region for completing to divide, and plans from starting point semantic region to the path of target semantic region, makes Equipment moves to target semantic region according to path from starting point semantic region.In order to ensure that equipment can accurately according to the road of planning Move in footpath, it is necessary to which the moment judged semantic region residing for equipment, be adjusted move mode.As a kind of optional reality Example is applied, equipment is moved to target semantic region from starting point semantic region according to path includes：By equipment to the image that obtains It is identified obtaining the semantic region corresponding to the image, wherein, the semantic region that equipment is identified is used to determine that equipment is current The semantic region at place；Semantic region and path according to equipment where current make equipment move to target semantic region.If It is standby to pass through image capture module, the image information of Real-time Collection equipment region, and determined to set according to the image information for obtaining The semantic region at standby place at present, and may determine that whether equipment moves according to the path of planning to target semantic region, from And equipment can be made to be moved in the semantic region route of regulation according to the route of planning.

In order to be able to enable in above-described embodiment really locking equipment at present where semantic region mode carried out by computer, Used as a kind of optional embodiment, the semantic region for being identified obtaining to the image for obtaining corresponding to the image by equipment is wrapped Include：Obtain the characteristic point in image；By the characteristic point storehouse corresponding with each semantic region being pre-configured with of the characteristic point in image In characteristic point matched, wherein, corresponding one or more characteristic points in each semantic region are have recorded in characteristic point storehouse；Really Whether the quantity for determining the matching of the characteristic point in the characteristic point storehouse corresponding with each semantic region of the characteristic point in image meets pre- Fixed condition；In the case where predetermined condition is met, the corresponding semantic region of image is determined.

Specifically, the image for equipment being obtained carries out feature point extraction, and the characteristic point that will extract and each semantic region Characteristic point in corresponding characteristic point storehouse is matched, and the characteristic point recorded according to matching result in gained image is semantic with each Whether the quantity of the matching of characteristic point, predetermined bar is met by the quantity of judging characteristic Point matching in the character pair point storehouse of region Part, the semantic region where carrying out determination equipment, wherein, each semantic region have recorded the semantic region in corresponding characteristic point storehouse One or more characteristic points.For said process, can be represented by formula：N_n,i=I_n∩Q_iI=1,2 ..., M, n= 1,2,3 ..., N } wherein, N_n,iRepresent n-th image I in image_nWith ith zone characteristic point storehouse Q_iCarry out matching what is obtained Matching characteristic point set.The program is matched by the characteristic point in image with the characteristic point in predetermined characteristic point storehouse, from And determine that the mode of the corresponding semantic region of image can be realized by calculating, and then equipment can be allow certainly according to the method Semantic region where dynamic identification equipment.

Alternatively, the characteristic point for matching is used in the corresponding characteristic point storehouse in each semantic region, can is the semantic region Whole characteristic points in corresponding characteristic point storehouse.

Equipment is where image is recognized, it is necessary to by the characteristic point in image and each semantic region during semantic region Characteristic point in corresponding characteristic point storehouse is matched, wherein, as a kind of alternative embodiment, the corresponding spy in each semantic region Levying a storehouse can be set up by way of obtaining image.The corresponding characteristic point storehouse in each semantic region is quite with equipment in the semanteme The whole characteristic points in image are extracted by obtaining image, you can set up corresponding characteristic point storehouse in region.

Equipment where image is recognized during semantic region, due to gained image be it is random obtain, wherein, some Image there may be the very few or excessive situation of characteristic point quantity in image.Determined according to characteristic point quantity very few in image Image where semantic region be inaccurate；According to semantic where the equipment that the excessive image of characteristic point quantity in image determines During region, because characteristic point quantity is more, matching process is slower.In order to solve the above problems this application provides one kind Optional embodiment, obtains the characteristic point in image, including：Determine whether the quantity of characteristic point in image belongs to preset range； In the case where preset range is belonged to, the characteristic point in image is obtained.The program, the characteristic point in image and preset range are entered Row matching, determines that feature is counted out and belongs to the image of preset range, obtains the characteristic point in the image.

As a kind of alternative embodiment, in the case of characteristic point negligible amounts in image, referred in such image Characteristic point negligible amounts, the characteristic point for referring to is easily changed, and the semantic region with where this feature point determines image Error can be very high, so abandoning matching to this kind of image.For the judgement of this kind of image, can by the characteristic point in image with it is pre- Determine scope to be matched, determine feature count out be less than the image of preset range abandon matching image.

As another alternative embodiment, in the case of characteristic point is a fairly large number of in the image, for example, one side has been pasted and has been entirely The wall of the wallpaper of rule distribution coloured speckle, in the case where such image is got, can detect in normal image Hundred times of feature is counted out, the foundation in this meeting interference characteristic point storehouse, also can seriously drag slow matching speed.For sentencing for this kind of image It is disconnected, the characteristic point in image can be matched with preset range, determine feature count out higher than preset range image i.e. It is the image for abandoning comparing.

During whether the characteristic point quantity for judging matching conforms to a predetermined condition, as a kind of alternative embodiment, really Determine the characteristic point in image includes with whether the quantity for matching of the characteristic point in characteristic point storehouse meets predetermined condition：According to image In characteristic point characteristic point storehouse corresponding with each semantic region in the quantity of matching of characteristic point compared with predetermined threshold； It is determined that the quantity of matching meets predetermined condition more than predetermined threshold.The embodiment, the characteristic point in image is semantic with each Characteristic point in the corresponding characteristic point storehouse in region is matched, and obtains the quantity of matching characteristic point, and in the number of Feature Points Matching Amount directly determines that the image is the image for meeting reservation condition more than in the case of threshold value.

During whether the quantity of judging characteristic Point matching conforms to a predetermined condition, as a kind of optional embodiment, Determine the characteristic point in image includes with whether the quantity for matching of the characteristic point in characteristic point storehouse meets predetermined condition：According to figure As in characteristic point characteristic point storehouse corresponding with each semantic region in Feature Points Matching quantity, according to match quantity by More to few order is ranked up to each semantic region；Determine that clooating sequence meets predetermined condition in the first semantic region. Clooating sequence the characteristic point during the first semantic region represents the corresponding characteristic point storehouse in the semantic region and image in spy The quantity for levying Point matching is most, it may be determined that the image closest to semantic region, such that it is able to where accurately determining the image Semantic region.

In order to more accurately determine semantic region where image by predetermined condition, as a kind of optional embodiment, base In semantic region, planning includes from starting point semantic region to the path of target semantic region：Based on the semanteme to assignment instructions point Analysis, draws the semantic relation in assignment instructions；According to semantic relation, starting point semantic region and target semantic region are determined, plan From starting point semantic region to the path of target semantic region.The assignment instructions that equipment sends to user carry out semantic analysis, obtain Assignment instructions include one or more semantic regions of target semantic region, and according to the relation between each semantic region, rule Draw from starting point semantic region to the path of target semantic region.

Fig. 3 is a kind of optional semantic region topological structure schematic diagram according to embodiments of the present invention, as shown in Figure 3.

Assignment instructions include one or more semantic regions, and the relation between each semantic region can be by topological structure Mark, is mapped as corresponding topological structure, as shown in figure 3, semantic region is by the schematic diagram that the semantic region shown in Fig. 2 divides A kind of network structure, is connected between each semantic region by topological structure, wherein, each circle represents a kind of semantic region, and Represented with greek numerals.Equipment is analyzed according to the assignment instructions for receiving to assignment instructions, determines to be closed between assignment instructions The topological structure of system, equipment is navigate on the position of needs according to the topological structure.For example, equipment receives assignment instructions being " taking out food from refrigerator to heat, before being sent to bed ", in this case, equipment is analyzed firstly the need of according to assignment instructions Wherein feature, determines semantic region where task, and planning apparatus are to the path of the semantic region.Above-mentioned is received in equipment , it is necessary to complete following operation, equipment is moved to the semantic region VII in kitchen after business instruction, food is taken from refrigerator and is made After with microwave stove heat, hall is entered by kitchen semantic region VI, VII, IV, II and I, then by some semantemes in hall Behind region, further through the semantic region I in bedroom, the semantic region II where reaching bed, before the food that will be heated is sent to bed.Root According to the process, you can path planning.In addition in above process, it is found that various spies may be included in single semantic region Levy, such as the semantic region VII in kitchen contains three kinds of features, respectively refrigerator, gas range and micro-wave oven, and robot is in kitchen In semantic region VII complete " taking out food from refrigerator " and " heating " two tasks.

During path planning, some semantic regions in predetermined space are that equipment has to pass through, such as, setting For in the case of needing to leave a certain room, it is necessary to by the semantic region where porch, reach the door of next predetermined space Semantic region where corridor, therefore in the case where the region of equipment movement includes multiple predetermined spaces, can first planning apparatus exist Path in each predetermined space (that is, each room), then each path is spliced, obtain complete path planning.By upper Embodiment is stated, path planning can be divided into the path planning spliced by the path planning in multiple individually predetermined spaces, its In, the predetermined space that path planning passes through is chosen, then determine the path planning in the predetermined space, it is possible to reduce path planning institute The semantic region quantity for referring to, so that the process of path planning is simpler quick.

Before planning is from starting point semantic region to the path of target semantic region, the semanteme that can be set up in predetermined space Region corresponding relation of that on map, so that equipment plans guidance path according to the rule of semantic region on map.Make It is a kind of alternative embodiment, the semantic region in predetermined space is mapped as the regular figure on the corresponding map of predetermined space Block；According to the corresponding regular figure block path planning in semantic region.Allow equipment according to the navigation of semantic region in semantic space Domain path planning on the then image block on corresponding map.It is mapped as on the corresponding map of predetermined space by by semantic region Regular figure block, equipment moves between semantic region, you can to think that equipment is moved between the regular figure block so that Starting point semantic region and target semantic region can be corresponded to equipment two regular figure blocks on map, and then can make to set The distance between standby two positions of regular image block that can be gone up according to the map and the two regular image blocks, cooking up can be with The path planning of help equipment accurate navigation.

Fig. 4 is the schematic diagram that a kind of optional semantic region according to embodiments of the present invention is mapped as graph block, such as Fig. 4 institutes Show, the schematic diagram that the semantic region shown in Fig. 2 divides is mapped as the regular figure block on corresponding map, each rule in Fig. 4 The corresponding semantic region of the Roman character in Roman alphabet diagram 2 in graph block, the nul in Fig. 4 represents the regular figure The corresponding blank semantic region of block.

Semantic region in predetermined space is mapped as the regular figure block on the corresponding map of predetermined space, i.e. will set The semantic region in standby place room is mapped as the regular figure block on the corresponding map in the room, can make equipment according to being capable of root According to the regular figure block on map, the moving direction of equipment is determined, in order to be navigated to equipment.

It is corresponding according to semantic region as a kind of optional embodiment during the mobile route of planning apparatus Regular figure block path planning includes：Model planning according to Bayesian inference is from starting point semantic region to target semantic region Path.Bayes method can quickly obtain the posterior probability of maximum, institute using Bayesian network under conditions of given input Can most meet constraints path can go out from Bayesian Network Inference.

Bayes method can be under conditions of given input, and the posteriority that maximum is quickly obtained using Bayesian network is general Rate, so can most meet constraints path from Bayesian Network Inference.

Fig. 5 is a kind of optional Bayesian network schematic diagram according to embodiments of the present invention, as shown in figure 5, according to given The Origin And Destination in path, dynamically sets up Bayesian network, it is determined that meeting the path of constraints.

It is rectangle in regular figure block, and equipment can at least along the movement of upper and lower, left and right four direction in rectangle In the case of, the model of Bayesian inference is as follows：O represents the four direction of rectangle,Expression equipment can be in predetermined space R In semantic region N upward direction movement,Expression equipment can be in predetermined space R semantic region N downwards To movement,Expression equipment can be in predetermined space R semantic region N left direction movement,Expression equipment can The right direction movement of the semantic region N in predetermined space R.

It is rectangle in regular image block, and equipment can be in the case where each rotation direction be 90 °, really in the rectangle Locking equipment moves to the number of revolutions of target semantic region from certain semantic region, wherein, number of revolutions is moved to according to equipment The current moving direction that a upper moving direction of a upper semantic region moves to current semantics region with equipment determines that Bayes pushes away Reason model includes as follows：Expression equipment moves to the number of revolutions in current semantics region, o from a upper semantic region_iExpression sets The standby upper moving direction for moving to a semantic region, o_jExpression equipment moves to the current movement side in current semantics region To o_iWith o_jFour, upper and lower, left and right moving direction is can use respectively；If upper moving direction o_iWith current moving direction o_jIt is identical, Then number of revolutionsIf upper moving direction o_iWith current moving direction o_j90 ° are differed, then number of revolutionsIf Upper moving direction o_iWith current moving direction o_j180 ° are differed, then number of revolutions

The mobile route of planning apparatus also needs to introduce transition probability this concept.Transition probability is equipment on moving to After a upper moving direction of one semantic region, the probability of the current moving direction in current semantics region is moved to again, wherein, upper one Semantic region and current semantics region are neighboring semantic regions；Transition probability is moved to currently according to equipment from a upper semantic region The number of revolutions of semantic region and whole direction of rotation of equipment calculate, and Bayesian inference model includes as follows：Mark turns The formula for moving probability is construed to：Wherein,Represent Equipment after a upper moving direction of a semantic region, moves to the current moving direction in current semantics region again on moving to Probability, wherein,Represent transition probability, R_MWith R_NIt is two neighboring semantic regions in predetermined space；The calculating of transition probability Formula is：Wherein, Ω_oExpression equipment whole rotation direction.

According to the recurrence relation and transition probability of Bayesian network, every paths in same predetermined space are obtained Probability, Bayesian inference model includes as follows：The probability per paths is represented, wherein, M,According to the maximum probability for calculatingPath planning.

In moving process, exist needs through some predetermined spaces, i.e., not carried out in the predetermined space for passing through equipment Any action in addition to related to movement.During equipment passes through the predetermined space, as a kind of alternative embodiment, it is based on Semantic region planning includes from starting point semantic region to the path of target semantic region：Planning is from starting point semantic region to target language The path in adopted region includes predetermined access path, wherein, predetermined access path pre-sets, starting point semantic region and target Semantic region is in different predetermined spaces.

Alternatively, needed through a feelings for multiple predetermined spaces from starting point semantic region to the path of target semantic region Under condition, the access path between multiple predetermined spaces is pre-set.Equipment, can be by during through some predetermined spaces According to the access path movement between the multiple predetermined spaces for pre-setting, equipment need not be done unrelated with movement in moving process Any work, it is ensured that equipment can be passed rapidly through in the predetermined space, and the access path for pre-setting can also It is shortest path of the equipment by space at this, equally can also accelerates equipment through the speed of the predetermined space.

During path planning, can also be according to the Identity Plan path in predetermined space, as a kind of optional reality Example is applied, is included from starting point semantic region to the path of target semantic region based on semantic region planning：Semanteme in predetermined space Region includes corresponding one or more the semanteme marks in semantic region, wherein, semanteme mark includes the feature of semantic region；According to language Corresponding one or more the semanteme mark path plannings in adopted region.

Alternatively, included according to corresponding one or more the semanteme mark path plannings in semantic region：According to particle filter Device model is planned from starting point semantic region to the path of target semantic region.

Fig. 6 is a kind of optional particle filter model schematic according to embodiments of the present invention, as shown in Figure 6.

The semantic mark of hexagonal star representation, small circle represents particle.Hexagonal star representation in major axis ellipse is truly collected Semantic mark, the semantic mark that the hexagonal star representation maximum probability in short axle ellipse is collected.Certain small circle and whole hexagonals The distance of star represents that this particle is assigned to the probability of all semantic marks, and probability is inversely proportional with distance.Model state is to adopt The probability of the semantic mark of collection, wherein, the probability of semanteme mark is the weight of semantic mark.

Neighboring semantic mark according to where equipment, and equipment semantic mark on collecting, determine that equipment is gathered To the probability of current semantics mark, included as follows according to particle filter model：N_LRepresent the quantity of semantic mark；N_PRepresent grain The quantity of son；Represent state setWherein,Represent semantic the mark l, Q of predetermined space K_CRepresent pre- Determine only one semantic mark in the C of space,Q_C；q_tState of the expression equipment in t；O represents that collection is arbitrary Semanteme mark；Represent states of the particle i in t；Represent that particle c is assigned in tProbability；Represent the semantic mark of collectionProbability；Represent weight of i-th particle in t.

After collecting upper one semantic mark according to equipment, the probability of the current semantics mark for collecting again determines equipment Mark transition probability, includes as follows according to particle filter model：Represent to work as and collect One semantic markAfter collect current semantics markProbability, wherein,Mark transition probability is represented, m and n is phase Adjacent semantic mark, K and H is neighboring semantic region；Identify transition probability formula be：Wherein, d tables Show the interval between two related semantic marks；σ_dRepresent the standard deviation of d；Represent and identified in semantemeWithIt Between do not have other semanteme mark；Represent semantic markWithBetween also have a semantic mark；Table Show two semantic marks between different predetermined spaces, and two predetermined spaces by only having a predetermined space for semantic mark It is connected；Represent that only one predetermined space of semantic mark is identified with semanticBetween interval；Represent semantic MarkPredetermined space with only one semantic mark is adjacent；Represent only one predetermined space of semantic mark WithBetween also have a semantic mark.

Assuming that each semanteme mark is independent acquisition, state vector is constituted by the probability of the semantic mark for collecting, Included as follows according to particle filter model：Expression state to Amount.

In the case where being collected without new semantic mark, state vector is constant, according to particle filter model Include as follows：Wherein, w_tThe vector of white noise composition is represented,

The observation model of each semanteme mark is obtained according to state vector, is included as follows according to particle filter model：

For each semanteme mark, observation model can be expressed as：

Wherein, v_tThe vector of white noise composition is represented, Represent that semantic mark m is adopted in predetermined space K Collection；Z_tExpression observes the probability of semantic expressiveness, and Z_tExpression collects the proportion of each semanteme mark.

According to the semanteme mark for truly collecting, the collected probability of all semantic marks is calculated by total probability formula, Included as follows according to particle filter model：It is total probability formula.

All particles in particle filter model are assigned to each semanteme mark, whole semantic regions according to random chance Gathered by equiprobability, included as follows according to particle filter model：Represent t₀Moment collects semanteme MarkProbability.

The weight for collecting each semanteme mark is represented that probability space description includes by probability：The power of each particle WeightBeing assigned to each semanteme mark can be expressed as： Wherein,w(cⁱ) weight of each particle is expressed as, weight can be empty by probability Between distance calculate：

According to the weight of each particle, the weight to each particle is standardized, it is ensured that all particles and be 1, according to Weight, calculates each particle and is assigned to the probability of a certain semantic mark, then updates the probability of all semantic marks, and according to complete The probabilistic programming path of the semantic mark in portion, includes as follows according to particle filter model：Represent particle power The standardization formula of weight；Particle realizes resampling according to weight；According to weight w (cⁱ), calculate each particle and be assigned to semanteme MarkProbabilityUpdate the probability of all semantic marksAnd according to the maximum of all semantic marks Probabilistic programming path, wherein all the new probability formula of semantic mark is：

Fig. 7 is a kind of schematic diagram of guider according to embodiments of the present invention, as shown in fig. 7, the device includes：Obtain Unit 71, for obtaining the starting point semantic region in predetermined space and target semantic region, wherein, predetermined space is divided into Multiple semantic regions, multiple semantic regions include starting point semantic region and target semantic region, include in each semantic region Feature for identifying the semantic region from multiple semantic regions；Planning unit 72, for based on semantic region planning from Path of the starting point semantic region to target semantic region；Mobile unit 73, for making equipment according to path from starting point semantic region Move to target semantic region.

Used as a kind of alternative embodiment, mobile unit includes：Identification module, for being carried out to the image for obtaining by equipment Identification obtains the semantic region corresponding to the image, wherein, the semantic region that equipment is identified is used to determine the current place of equipment Semantic region；Mobile module, makes equipment move to target language for the semantic region and path according to equipment where current Adopted region.

Used as a kind of alternative embodiment, identification module includes：First acquisition module, for obtaining the characteristic point in image； Comparing module, for by the characteristic point in the characteristic point storehouse corresponding with each semantic region being pre-configured with of the characteristic point in image Compare, wherein, corresponding one or more characteristic points in each semantic region are have recorded in characteristic point storehouse；First determines mould Block, the quantity of matching for determining the characteristic point in the characteristic point storehouse corresponding with each semantic region of the characteristic point in image is It is no to meet predetermined condition；Second determining module, in the case where predetermined condition is met, determining the corresponding semantic space of image Domain.

Used as a kind of alternative embodiment, acquisition module includes：3rd determining module, the number for determining characteristic point in image Whether amount belongs to preset range；Second acquisition module, in the case where preset range is belonged to, obtaining the feature in image Point.

Used as a kind of alternative embodiment, the first determining module includes：Matching module, characteristic point in image and each The quantity of the Feature Points Matching in the corresponding characteristic point storehouse in semantic region is compared with predetermined threshold；4th determining module, is used for It is determined that the quantity of matching meets predetermined condition more than predetermined threshold.

Used as a kind of alternative embodiment, the first determining module includes：Order module, for the characteristic point in image with The quantity of the Feature Points Matching in the corresponding characteristic point storehouse in each semantic region is right according to the quantity for matching order from more to less Each semantic region is ranked up；5th determining module, for determining that clooating sequence meets predetermined bar in the first semantic region Part.

Used as a kind of alternative embodiment, planning unit includes：Analysis module, for based on the semanteme to assignment instructions point Analysis, draws the semantic relation in assignment instructions；6th determining module, for according to semantic relation, determine starting point semantic region and Target semantic region, plans from starting point semantic region to the path of target semantic region.

Used as a kind of alternative embodiment, planning unit includes：First mapping block, for by the semantic space in predetermined space Domain mapping is the regular figure block on the corresponding map of predetermined space；First planning submodule, for according to semantic region correspondence Regular figure block path planning.

Used as a kind of alternative embodiment, the first planning submodule includes：First MBM, for according to Bayesian inference Model plan from starting point semantic region to the path of target semantic region.

Used as a kind of alternative embodiment, planning unit includes：Predetermined module, for planning from starting point semantic region to target The path of semantic region includes predetermined access path, wherein, predetermined access path pre-sets, starting point semantic region and mesh Mark semantic region is in different predetermined spaces.

Used as a kind of alternative embodiment, planning unit includes：Second mapping block, for by the semantic space in predetermined space Domain mapping is the semantic mark on the corresponding map of predetermined space, wherein, each semantic region includes one or more semanteme marks Know, semanteme mark includes the feature of semantic region；Second planning submodule, for according to semantic region it is corresponding one or more Semanteme mark path planning.

Used as a kind of alternative embodiment, the second planning submodule includes：Second MBM, for according to particle filter Model is planned from starting point semantic region to the path of target semantic region.

The embodiments of the present invention are for illustration only, and the quality of embodiment is not represented.

In the above embodiment of the present invention, the description to each embodiment all emphasizes particularly on different fields, and does not have in certain embodiment The part of detailed description, may refer to the associated description of other embodiment.

In several embodiments provided herein, it should be understood that disclosed technology contents, can be by other Mode is realized.Wherein, device embodiment described above is only schematical, such as division of described unit, Ke Yiwei A kind of division of logic function, can there is other dividing mode when actually realizing, such as multiple units or component can combine or Person is desirably integrated into another system, or some features can be ignored, or does not perform.Another, shown or discussed is mutual Between coupling or direct-coupling or communication connection can be the INDIRECT COUPLING or communication link of unit or module by some interfaces Connect, can be electrical or other forms.

The unit that is illustrated as separating component can be or may not be it is physically separate, it is aobvious as unit The part for showing can be or may not be physical location, you can with positioned at a place, or can also be distributed to multiple On unit.Some or all of unit therein can be according to the actual needs selected to realize the purpose of this embodiment scheme.

In addition, during each functional unit in each embodiment of the invention can be integrated in a processing unit, it is also possible to It is that unit is individually physically present, it is also possible to which two or more units are integrated in a unit.Above-mentioned integrated list Unit can both be realized in the form of hardware, it would however also be possible to employ the form of SFU software functional unit is realized.

If the integrated unit is to realize in the form of SFU software functional unit and as independent production marketing or use When, can store in a computer read/write memory medium.Based on such understanding, technical scheme is substantially The part for being contributed to prior art in other words or all or part of the technical scheme can be in the form of software products Embody, the computer software product is stored in a storage medium, including some instructions are used to so that a computer Equipment (can be personal computer, server or network equipment etc.) perform each embodiment methods described of the invention whole or Part steps.And foregoing storage medium includes：USB flash disk, read-only storage (ROM, Read-Only Memory), arbitrary access are deposited Reservoir (RAM, Random Access Memory), mobile hard disk, magnetic disc or CD etc. are various can be with store program codes Medium.

The above is only the preferred embodiment of the present invention, it is noted that for the ordinary skill people of the art For member, under the premise without departing from the principles of the invention, some improvements and modifications can also be made, these improvements and modifications also should It is considered as protection scope of the present invention.

Claims

1. a kind of air navigation aid, it is characterised in that including：

Starting point semantic region and target semantic region in predetermined space are obtained, wherein, the predetermined space is divided into many Individual semantic region, the multiple semantic region includes the starting point semantic region and the target semantic region, each semantic space Include the feature for identifying the semantic region from the multiple semantic region in domain；

Based on semantic region planning from the starting point semantic region to the path of the target semantic region；

Equipment is set to move to the target semantic region from the starting point semantic region according to the path.

2. method according to claim 1, it is characterised in that make the equipment semantic from the starting point according to the path Region moves to the target semantic region to be included：

The image for obtaining is identified by the equipment obtain the semantic region corresponding to the image, wherein, the equipment The semantic region identified is used for the semantic region where determining the equipment at present；

Semantic region and the path according to the equipment where current make the equipment move to the target semantic space Domain.

3. method according to claim 2, it is characterised in that be identified obtaining to the image for obtaining by the equipment Semantic region corresponding to the image includes：

Obtain the characteristic point in described image；

Characteristic point in characteristic point in described image characteristic point storehouse corresponding with each semantic region being pre-configured with is carried out Matching, wherein, corresponding one or more characteristic points in each semantic region are have recorded in the characteristic point storehouse；

Determine the matching of characteristic point in the characteristic point characteristic point storehouse corresponding with described each semantic region in described image Whether quantity meets predetermined condition；

In the case where the predetermined condition is met, the corresponding semantic region of described image is determined.

4. method according to claim 3, it is characterised in that obtain the characteristic point in described image, including：

Determine whether the quantity of characteristic point in described image belongs to preset range；

In the case where preset range is belonged to, the characteristic point in described image is obtained.

5. method according to claim 3, it is characterised in that determine the characteristic point in described image and the characteristic point storehouse In the quantity of Feature Points Matching whether meet predetermined condition and include：

The number of the Feature Points Matching in characteristic point characteristic point storehouse corresponding with described each semantic region in described image Amount is compared with predetermined threshold；

Determine that the quantity of the matching meets the predetermined condition more than predetermined threshold.

6. method according to claim 3, it is characterised in that determine the characteristic point in described image and the characteristic point storehouse In the quantity of Feature Points Matching whether meet predetermined condition and include：

The number of the Feature Points Matching in characteristic point characteristic point storehouse corresponding with described each semantic region in described image Amount, is ranked up according to the quantity order from more to less of the matching to described each semantic region；

Determine that clooating sequence meets predetermined condition in the first semantic region.

7. method according to claim 1, it is characterised in that based on semantic region planning from the starting point semantic region to The path of the target semantic region includes：

Based on the semantic analysis to assignment instructions, the semantic relation in the assignment instructions is drawn；

According to the semantic relation, the starting point semantic region and the target semantic region are determined, plan from the starting point language Path of the adopted region to the target semantic region.

8. method according to claim 1, it is characterised in that based on semantic region planning from the starting point semantic region to The path of the target semantic region includes：

Semantic region in the predetermined space is mapped as the regular figure block on the corresponding map of the predetermined space；

The path is drawn according to the corresponding regular figure slip gauge in the semantic region.

9. method according to claim 8, it is characterised in that drawn according to the corresponding regular figure slip gauge in the semantic region The path includes：

Model according to Bayesian inference is planned from the starting point semantic region to the path of the target semantic region.

10. method according to claim 9, it is characterised in that in the regular figure block be rectangle, and the equipment In the case of being only capable of being moved along upper and lower, left and right four direction in the rectangle, the model of the Bayesian inference is as follows：

O represents the four direction of the rectangle,Represent that the equipment can be in the semantic space in the predetermined space R The upward direction movement of domain N,Represent that the semantic region N's that the equipment can be in the predetermined space R is downward Direction is moved,The left direction movement of the semantic region N that the equipment can be in the predetermined space R is represented,Represent the right direction movement of the semantic region N that the equipment can be in the predetermined space R.

11. methods according to claim 10, it is characterised in that in the regular figure block be rectangle, and described set It is standby to determine that the equipment is moved from certain semantic region in the case where each rotation direction is 90 ° in the rectangle To the number of revolutions of target semantic region, wherein, the number of revolutions moves to the upper of a upper semantic region according to the equipment The current moving direction that one moving direction moves to current semantics region with the equipment determines that the Bayesian inference model is such as Under include：

Represent that the equipment moves to the number of revolutions in current semantics region, the o from a upper semantic region_iRepresent A described upper moving direction of the equipment, the o_jRepresent the described current moving direction of the equipment, the o_iWith the o_j Four, upper and lower, left and right moving direction is can use respectively；

If the upper moving direction o_iWith the current moving direction o_jIt is identical, then the number of revolutions

If the upper moving direction o_iWith the current moving direction o_j90 ° are differed, then the number of revolutions

If the upper moving direction o_iWith the current moving direction o_j180 ° are differed, then the number of revolutions

12. methods according to claim 11, it is characterised in that transition probability is equipment semanteme on moving to After a upper moving direction in region, the probability of the current moving direction in current semantics region is moved to again, wherein, a upper language Adopted region is neighboring semantic region with the current semantics region；The transition probability is according to the equipment from a upper semantic region Whole direction of rotation of the number of revolutions and the equipment that move to current semantics region calculate, and the Bayes pushes away Reason model includes as follows：

The formula of the transition probability is construed to：Wherein, Represent that the equipment is being moved on described after a upper moving direction of a semantic region, working as current semantics region is moved to again The probability of preceding moving direction, wherein,Represent the transition probability, R_MWith R_NIt is two adjacent languages in the predetermined space Adopted region；

The computing formula of the transition probability is：Wherein, Ω_oRepresent the equipment all rotation sides To.

13. methods according to claim 12, it is characterised in that recurrence relation and the transfer according to Bayesian network Probability, obtains the probability per paths in same predetermined space, and the Bayesian inference model includes as follows：

The probability per paths is represented, wherein,According to what is calculated Maximum probabilityPlan the path.

14. methods according to claim 1, it is characterised in that based on semantic region planning from the starting point semantic region Path to the target semantic region includes：

Planning includes predetermined access path from the starting point semantic region to the path of the target semantic region, wherein, it is described Predetermined access path pre-sets, and the starting point semantic region and the target semantic region are in different predetermined spaces.

15. methods according to claim 1, it is characterised in that based on semantic region planning from the starting point semantic region Path to the target semantic region includes：

Semantic region in the predetermined space is mapped as the semantic mark on the corresponding map of the predetermined space, wherein, Each described semantic region includes one or more semanteme marks, and the semanteme is designated the characteristic point of object in semantic region；

According to corresponding one or more the described paths of semanteme mark planning in the semantic region.

16. methods according to claim 15, it is characterised in that according to corresponding one or more languages in the semantic region The justice mark planning path includes：

According to the planning of particle filter model from the starting point semantic region to the path of the target semantic region.

17. methods according to claim 16, it is characterised in that semantic mark can be gathered in being moved through in the equipment In the case of knowledge, according to the semanteme mark on the corresponding map of the predetermined space, the last language of the equipment collection is determined The neighboring semantic mark of justice mark, determines the equipment by collecting the probability of the semantic mark after movement, according to described Particle filter model includes as follows：

N_LRepresent the quantity of the neighboring semantic mark；N_PRepresent the quantity of particle in the particle filter model；The semanteme MarkNeighboring semantic mark set beWherein, it is describedIn representing the predetermined space K Semanteme mark l, the Q^COnly one semantic mark in the predetermined space C is represented, it is describedRepresent the predetermined space M In semantic mark m, it is describedRepresent the semantic mark r in the predetermined space R；q_tRepresent the current state of the equipment； O represents the arbitrary semantic mark of collection；Represent the particle i of current state；Represent the particle i of current state It is assigned toProbability；Expression collects semantic markProbability；Represent i-th particle Present weight.

18. methods according to claim 17, it is characterised in that mark transition probability is the equipment one on collecting After secondary semantic mark, by the mobile mark probability for collecting current semantics mark again, wherein, the semanteme is designated adjacent language Justice mark；The mark transition probability is according to the interval determination between two related to equipment movement semantic marks, institute State particle filter model includes as follows：

The formula of the mark transition probability is construed to：Wherein, Expression is collecting semantic markAfterwards, semantic mark is collected againProbability,The mark transition probability is represented, M and n is the neighboring semantic mark, and K and H is the neighboring semantic region；

It is described mark transition probability formula be：Wherein, d represents related to equipment movement Interval between two semantic marks, σ_dRepresent the standard deviation of interval d；

Represent semantic markIdentified with semantemeBetween interval：Wherein,

Represent and identified in semantemeIdentified with semantemeBetween it is adjacent；

Represent semantic markIdentified with semantemeBetween also have a semantic mark；

Represent semantic markIdentified with semantemeBetween there is only one predetermined space of semantic mark；

Represent that only one predetermined space of semantic mark is identified with semanticBetween interval；Wherein,

Represent that only one predetermined space of semantic mark is identified with semanticIt is adjacent；

Represent that only one predetermined space of semantic mark is identified with semanticBetween also have a semantic mark.

19. methods according to claim 18, it is characterised in that semantic mark is gathered in moving process by the equipment The probability of knowledge constitutes the state vector of the probability, is included as follows according to the particle filter model：

Represent the state vector.

20. methods according to claim 19, it is characterised in that in the case where being collected without new semantic mark, The state vector is constant, the state transition model of all particle filters, as follows according to the particle filter model Including：

Wherein, w_tThe vector of white noise composition is represented,

21. methods according to claim 20, it is characterised in that according to the state transition model of all particle filters, The observation model of each semanteme mark in the predetermined space is obtained, is included as follows according to the particle filter model：

For each semanteme mark, observation model can be expressed as：

Wherein, v_tThe vector of white noise composition is represented,

Represent that semantic mark m is collected in predetermined space K；

Z_tRepresent the probability that semantic mark is observed by observation model, and Z_tExpression collects the proportion of each semanteme mark.

22. methods according to claim 21, it is characterised in that according to the sight of each semanteme mark in the predetermined space Model is surveyed, the total probability formula in the observation model is extracted, is included as follows according to the particle filter model：

It is total probability formula.

23. methods according to claim 22, it is characterised in that according to the total probability formula and true collect Semanteme mark, calculating all collected probability of semantic mark includes, by all particles in particle filter model according to Random chance is assigned to each semanteme mark, and the population and whole number of ions being assigned to are identified according to described each semanteme, obtains Go out the probability for collecting each semanteme mark, included as follows according to the particle filter model：

Expression collects semantic markProbability.

24. methods according to claim 23, it is characterised in that collect the weight of each semanteme mark by probability Represent, probability space description includes：

The weight of each particleBeing assigned to each semanteme mark can be expressed as：

Wherein,w(cⁱ) be expressed as often The weight of individual particle, the weight can be calculated by the distance of probability space：

25. methods according to claim 24, it is characterised in that according to the weight of each particle, to the power of each particle It is standardized again, it is ensured that all particles and be 1, according to weight, calculates the probability that each particle is assigned to a certain semantic mark, Update the probability of all semantic marks again, and according to the probabilistic programming path of all semantic marks, according to the particle filter Device model includes as follows：

Represent the standardization formula of particle weights；

Particle realizes resampling according to weight；

According to weight w (cⁱ), calculate each particle and be assigned to semantic markProbability

Update the probability of all semantic marksAnd the path is planned according to the maximum probability of all semantic marks, wherein entirely The new probability formula of the semantic mark in portion is：

A kind of 26. guiders, it is characterised in that including：

Acquiring unit, for obtaining the starting point semantic region in predetermined space and target semantic region, wherein, the predetermined sky Between be divided into multiple semantic regions, the multiple semantic region includes the starting point semantic region and the target semantic space Domain, includes the feature for identifying the semantic region from the multiple semantic region in each semantic region；

Planning unit, for being planned from the starting point semantic region to the path of the target semantic region based on semantic region；

Mobile unit, for making equipment move to the target semantic region from the starting point semantic region according to the path.